‘ L; f' llilliilllllillfllilllilllllWlillillillili 4 1 5 E H43 Differentials in Supermarket Drawing Power and per Capita Sales by Store Complex and Store Size by Bernard Joseph La Londe Under the Sponsorship of Programs in Food Marketing Management Department of Marketing and Transportation Administration Michigan State University 1961 q Research Report in Mass Marketing Management No. 2 IIIIIIIIIIILIIIIIIIIIII IIIIIIIIIIIIIIIIIIIIIIIIIII 0053639 ”may w" Michigan State University PLACE IN RETURN BOX to remove thle checkout from your record. TOA ID FINES retmnonorbetoreddedue. DATE DUE DATE DUE DATE DUE l I—I____I__I ‘ Well-enleer-I DIFFERENTIALS IN SUPERMARKET DRAWING POWER AND PER CAPITA SALES BY STORE COMPLEX AND STORE SIZE BY Bernard Joseph La Londe AN ABSTRACT of a Thesis submitted to the School for Advanced Graduate Studies of Michigan State University in partial " fulfillment of the requirement for the degree of DOCTOR OF PHILOSOPHY Department of Marketing and Transportation Administration 1961 2 Bernard Joseph La Londe ABSTRACT Since World war II sizable population shifts into the sub- urbs, along with the advent of large scale retailing, have caused corresponding shifts in the structure of retail distri- bution.- One of the major problems facing the post war retailer is that of optimal location to serve a shifting market. The multi- unit retailer must solve the additional problem of expanding the store network so that optimal market coverage is attained in any given market area. The topic of this research is in the general area of re- tail store location. The specific problem deals with the influ- ence of store size and store complex upon the drawing power and per capita sales of the supermarket. The study was structured so that store size and store com- plex were independent variables and drawing power and per capita sales were dependent variables. The two independent variables selected for study were store size and store complex. Store size was chosen since it repre- sents the closest approximation of product offering within the retail store short of an actual physical inventory. The store size variable was operationally defined as the total square TEE-lllllllllI-IIDI‘. Bernard Joseph La Londe footage of selling area within the store. The store complex variable reflected the influence of the product offering at the retail cluster. It was operationally defined in terms of 1) number and type of stores surrounding the survey store, 2) traffic arteries, and 3) population and population density. The guiding hypothesis for the research, then, was that the individual consumer is influenced in food purchasing behavior by product offering within the store and product offering at the retail cluster. Specific hypotheses were constructed to test the relationship between the independent and dependent variables. The dependent variables of drawing power and per capita sales were designed to provide insights for both optimum store network expansion and individual store development. That is, the drawing power variable indicates the appropriate place- ment of network units for optimum market coverage. The per capita sales measurement proVides a framework for analyzing the sales potential of a single unit in the store network. A total of fifteen supermarkets were selected to provide empirical data for the research. In order to hold price and promotional factors relatively constant, the fifteen stores were selected from the same chain organization and metropoli- tan area. Customers of the individual survey stores were _.I-IIII|III I I I I I I I I I I I I I I I.“ luau"... 4 Bernard Joseph La Londe asked a series of questions, including a request for their home addresses. These addresses were plotted on a line map of the area and thus provided the basic empirical data for the study. The number of interviews for each survey store were established by using a sales based quota. The quota proce- dure resulted in a total of 5,300 usable customer interviews for the fifteen survey stores. The completed customer spotting maps and 1960 Census Tract population data were used in calculating drawing power and per capita sales totals. The data were analyzed statis- tically using analysis of Variance and Correlation procedures. Appropriate tables and summary analysis of the data and the implication of the data are also presented. On the basis of the research the following general conclu- sions can be drawn: (1) Store complex is an important influ- ence in determining the drawing power and per capita sales of the supermarket. (2) Store size is not an important variable in determining the drawing power and per capita sales of the supermarket. (3) There exist distinct and significant patterns of drawing power and per capita sales which can be isolated and quantitatively analyzed as a basis for future location decision. FIIII!IIIIIIIIIIIIII DIFFERENTIALS IN SUPERMARKET DRAWING POWER AND PER CAPITA SALES BY STORE COMPLEX AND STORE SIZE BY Bernard Joseph La Londe A THESIS Submitted to the School of Advanced Graduate Studies of Michigan State University in Partial Fulfillment of the Requirements for the Degree of DOCTOR OF PHILOSOPHY Department of Marketing and Transportation Administration 1961 III—IIIIIII-IIUIDIIHHE 252254 I/lo C 1 Copyright by Bernard Joseph La Londe 1961 el-llll-ll iii ACKNOWLEDGEMENTS A research work is built upon the ideas and experiences of both the writer and those who have influenced and guided his ideas and experiences. The author would like to single out the following individuals and organizations for their significant contributions to the formulation and execution of the research problem. Dr. Arthur E. Warner, Graduate School of Business Administra- tion, under whose direct guidance the research was com- pleted, for his patience, understanding and inspirational guidance. Dr. Frank H. Mossman, Department of Marketing and Transporta- tion Administration, who aided the author in the original formulation of the problem and provided a constant source of guidance and information. Dr. Edward M. Barnet, Graduate School of Business Adminis- tration, who was instrumental in providing a realistic industry orientation for the research problem. Dr. Thomas A. Staudt, Head, Department of Marketing and Trans- portation Administration, who served as chairman of the candidate's guidance committee and encouraged this research. iv Dr. Saul B. Cohen, Department of geography, Boston University, who freely shared the fruits of his research in the problem area. The individuals of the cooperating supermarket chain, with- out whose cooperation the research could not have been conducted. The General Electric Company, Michigan State University and the programs in Food Marketing Management whose financial support enabled the writer to devote his full resources to the research problem. Barbara and Lisa Renee for their understanding and cooperation. The author expresses his deep gratitude and sincere thanks to the above individuals and organizations, as well as all of the individuals whose direct and indirect assistance have made the research possible. It is with a measure of confidence that the author can state that the experience gleaned from the research will be reflected and magnified both in his own future contributions and the con- tributions of his students. TABLE OF CONTENTS CHAPTER I - INTRODUCTION . . . . . . . . . . . . . . . . Background of the Problem . . . . . . . . . . . . . The Supermarket Industry and the Location Decision. Recent Trends Affecting Supermarket Location . The Scope of the Problem. . . . . . . . . . . . . . Problem Statement. . . . . . . . . . . . . . . General Hypotheses. . . . . . . . . . . . . . . . . Method of Investigation . . . . . . . . . . . . . . Terms and Definitions . . . . . . . . . . . . . . . Limitations . . . . . . . . . . . . . . . . . . . . Some Possible Contributions of the Study. . . . . . Organization of the Study . . . . . . . . . . . . . CHAPTER II - CONSUMER SPACE PREFERENCES AND THE RETAIL, TRADING AREA. 0 e e e e e o e e e e e e e 0 Introduction. . . . . . . . . . . . . . . . . . . . Some Perspectives for Consumer Space Preferences. . Some Economic Contributions . . . . . . . . . . . . Some Consumer Utility Contributions . . . . . . . . Some Marketing Contributions. . . . . . . . . . . . Definitions and Criteria for Delineation of the Retail Trading Area . . . . . . . . . . . . . . . . Contributions of the Literature to the Research . . CPIAPTER III - RESEARCH DESIGNe e e e e e e e e e e e e e IntrOduction. O O O O I 0 C O O O I O O I O O O I I Page 10 13 14 15 19 24 25 28 31 31 35 41 46 51 54 59 62 62 Survey Stores . . . . . . . . . . . . . . . . . . . Store Classification . . . . . . . . . . . . . Survey Store Classification for this Research. Evaluation of Selected Survey Stores in Rela— tion to Criteria . . . . . . . . . . . . . Customer Survey Procedures. . . . . . . . . . . . . Customer Interviewing. . . . . . . . . . . . . Map Spotting . . . . . . . . . . . . . . . . . Measurement Procedures. . . . . . . . . . . . . . . Drawing Power Measurement. . . . . . . . . . . Per Capita Sales Measurement . . . . . . . . . CHAPTER IV - PRESENTATION OF FINDINGS. . . . . . . . . . IntrOduction. O O O O O O O O O O O O O O O O O O 0 Drawing Power Measurement . . . . . . . . . . . . . Introduction . . . . . . . . . . . . . . . . . Drawing Power and Store Complex. . . . . . . . Drawing Power and Store Size . . . . . . . . . Statistical Significance of Drawing Power Measurement. . . . . . . . . . . . . . . . . . Per Capita Sales Measurement. . . . . . . . . . . . Introduction . . . . . . . . . . . . . . . . . Per Capita Sales and Store Complex . . . . . . Per Capita Sales and Store Size. . . . . . . . Statistical Significance of Per Capita Sales Measurement. . . . . . . . . . . . . . . . . . Other Findings . . . . . . . . . . . . . . vi Page 63 63 66 72 73 73 75 75 75 77 82 82 85 85 87 88 95 98 98 99 101 106 110 vii Page Introduction . . . . . . . . . . . . . . . . . . 110 The Influence of Population Density on Drawing Power and Per Capita Sales . . . . . . . . . . . 111 The Influence of Date of Opening on Drawing Power and Per Capita Sales . . . . . . . . . . . 112 The Influence of Competition on Drawing Power and Per Capita Sales . . . . . . . . . . . . . . 112 Statistical Significance of Findings . . . . . . 113 CHAPTER V - SUMMARY AND CONCLUSIONS. . .I. . . . . . . . . 120 Introduction. . . . . . . . . . . . . . . . . . . . . 120 Evaluation of Hypothesis. . . . . . . . . . . . . . . 120 Store Complex Hypothesis . . . . . . . . . . . . 120 Store Size Hypothesis. . . . . . . . . . . . . . 123 Conclusions . . . . . . . . . . . . . . . . . . . . . 125 Statistical Findings . . . . . . . . . . . . . . 125 Analytical Conclusions . . . . . . . . . . . . . 128 Store Complex and Drawing Power . . . . . . 128 Store Size and Drawing Power. . . . . . . . 130 Store Type and Per Capita Sales . . . . . . 131 Store Size and Per Capita Sales . . . . . . 136 Implications of Findings for Location Policy and Strategy. 0 O O O O O O O O O O O O O O 9 O O O O O O 137 Store Complex and Location Policy. . . . . . . . 139 Store Size and Location'Policy . . . . . . . . . 144 The Dynamics of Location Policy and Strategy. . . . . 145 . - . . .V . w . . . . . _ . . . .4 Suggested Areas for Further Research. . . . . . . . . APPENDICES C O O O O O O O O I O I 0 O O O O O I O O O O O A. B. Delineating Retail Trading Areas in Urban Areas. Instructions for Customer Spotting . . . . . . . Forms, WOrksheets and Measurement Devices used in Research Procedure. . . . . . . . . . . . . . Statistical Procedures . . . . . . . . . . . . . Empirical Data - Survey Store Population Estimates by Quadrant and Distance Interval. . . Empirical Data — Survey Store Customers by Distance Traveled to Survey Store and Quadrant . Empirical Data - Location of Competitive Supermarkets by Survey Store.. . . . . . . . . . viii Page 146 149 150 172 178 188 195 198 209 10. ll. 12. l3. 14. 15. 16. LIST 9;; TABLES Classification of Retail Location Types . . . . . . . Unplanned Survey Store Data . . . . . . . . . . . . . Planned Survey Store Data...... . . . . . . . . . . . Types of Retail Units in Survey Planned Shopping centers 6 I e I O o e e e e o o e o e e e o e e o e 0 Number of Interviews and Date of Survey by Survey Store 0 O O O O O I O I O O O O C O I O O O O O O O I Customers by Distance Interval by Store Type. . . Drawing Power at 70 and 90 Per Cent Customer Levels by Store Type O O O O O O O O O O O O O O O O I O O 0 Drawing Power at 70 and 90 Per Center Customer Levels by Survey Store . . . . . . . . . . ._. . . . . . . . Comparison of Mean and Median Values of Drawing Power at 70 and 90 Per Cent Customer Levels by Survey Store Drawing Power at 70 and 90 Per Cent Customer Levels by Rank Order of Store Size . . . . . . . . . . . . . Drawing Power at 70 and 90 Per Cent Customer Levels by Store Type and Store Size. . . . . . . . . . . . . Percentage of Customers by Store Type at One-half, One and One-fourth and Two Mile Distance Intervals. . Per Capita Sales by Store Type at One-half, One and One-fourth and Two Mile Distance Intervals. . . . Per Capita Sales by Survey Store at One-half, One and One-fourth and Two Mile Distance Intervals. . . . Per Capita Sales at One-half, One and One-fourth and Two Mile Distance Intervals by Rank Order of Store Size. . . . . . . . . . . . . . . . . . . . . . Per Capita Sales at One-half, One and One-fourth and Two Mile Distance Intervals by Survey Store and Store Size. 0 C I I O O O O O O O O O O O O O O 0 ix Page 65-66 69 7O 71 84 86 9O 91 92 93 94 101 102 103 104 105 , ,, T. L x | 1.— 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. Population Density, Per Capita Sales and Drawing Power by Survey Store . . . . . . . . . . . . . . . Population Density, Per Capita Sales and Drawing Power by Store Type . . . . . . . . . . . . . . . . Per Capita Sales, Drawing Power and Date of Opening by Survey Store . . . . . . . . . . . . . . . . . . Per Capita Sales, Drawing Power and Date of Opening by Store Type C O C O O O O I O I O O C O O C C O 0 Number of Competitive Supermarkets by Distance Interval and Survey Store . . . . . . . . . . . . . Number of Competitive Supermarkets, Per Capita Sales and Drawing Power at One-half Mile Distance Interval by Survey Store. . . . . . . . . . . . . . Range of Drawing Power - General Classification by Store Type . . . . . . . . . . . . . . . . . . . Range of Drawing Power - General Classification by Survey Store . . . . . . . . . . . . . . . . . . Range of Per Capita Sales at One and One—fourth Mile Distance Interval - General Classification by Store Type . . . . . . . . . . . . . . . . . . . Range of Per Capita Sales at One and One—fourth Mile Distance Interval - General Classification by Survey Store. . . . . . . . . . . . . . . . . . . . Range of Per Capita Sales at Two Mile Distance Interval - General Classification by Store Type . . Range of Per Capita Sales at Two Mile Distance Interval - General Classification by Survey Store . Procedure for calculating Estimated weekly Sales for Proposed Site . . . . . . . . . . . . . . . . . Outline of Factors Influencing Retail Trading Areas Page 114 115 116 117 118 119 129 129 133 133 134 135 142 164 ll'll'llll.[‘re‘[.[TE-l—rr!lrl 1o. 11. 12. 13. LIST pg ILLUSTRATIONS Trading Area Under Perfect Competition. . . . Losch's Demand Cone . . . . . . . . . . . . . Total Travel Products (Troxel). . . . . . . . Total Loss of At-Home Satisfactions (Troxel). Net Travel Products (Troxel). . . . . . . . . Marginal Net Products (Troxel). . . . . . . . Analysis of Variance -- Store Complex and Drawing Power at 70 Per Cent Customer Drawing Power Level Analysis of Variance -- Store Complex and Drawing Power at 90 Per Cent Customer Drawing Power Level Analysis of Variance -— Significance of Per Capita Sales Distance Intervals. . . . . . . . . . . Analysis of Variance -- Store Complex and Per Capita Sales at One-half Mile Distance Interval. . . Analysis of Variance -- Store Complex and Per Capita Q Sales at One and One-fourth Mile Distance Interval. Analysis of Variance -- Store Complex and Per Capita Sales at Two Mile Distance Interval . . . . . Hypothetical Distribution Network . . . . . . Xi 43 44 47 47 48 48 95 96 106 107 108 109 140 flllllllll'llllllllsl CHAPTER I INTRODUCTION Background of the Problem_ During decade 1950-1959, the supermarket1 grew to a position of dominance in the food distribution field. In 1952, the 16,540 stores classified as supermarkets accounted for approximately 43 per cent of total grocery sales.2 By 1959, the number of stores had climbedto 32,000 with almost a 70 per cent share of total grocery sales.3 The elements of the sizable growth of supermarket share of market and sales volume have originated largely from four sources. First, the chief source of growth was business captured by the supermarket from the smaller and usually less efficient retailer. 1The term "supermarket" is a commonly used trade term denot- ing a large departmentalized retail store dealing primarily in food. The Super Market Institute states the following: "The original definition of a Super Market was coined by Super Market Merchandisingin 1936, and widely accepted everywhere. It de- _ scribed aSuper Market then as a retail establishment with a self- service grocery department, and meat, dairy and produce depart- ments--doing a combined volume of at least $250,000 a year. The minimum sales figure was revised to $500,000 recently, since the food price index has risen well over 100 per cent, making it logical to double the minimum, at least." ("The True Look of Super Market Industry, 1959." Super Market Merchandising, May, 1960 p. 1 -reprint-.) The formulation'is adopted with the $500,000 min- imum sales volume for definitional purposes in this research. It should be noted that this minimum annual sales volume was recently raised to $1,000,000 by the Super Market Institute. 2"Facts in Grocery Distribution," Progressive Grocer, April, 1960, p. F7. 3Ibid., p. F7. The supermarket distinguished itself as one of the leading innovators in internal efficiency of operations during the Post-WOrld War II period. Less efficient food retailers who would not, or could not, follow the lead of the more progres- sive supermarkets found themselves surrendering sales volume in ever increasing amounts.4 A second source of growth was due to population increases and new family formations. The rapidly increasing population of the United States provides a constantly growing base for expanded food sales. A third growth source, reflecting a constantly increasing standard of living, is found in the changing food preparation and buying habits of the American consumer. A recent report of the United States Department of Labor states: Many more foods than formerly are purchased either partially or completely prepared. Fruits and vege- tables are canned or frozen. Mixes, particularly those for baking, make cooking far simpler. Even fresh vegetables are more nearly ready to serve; spinach, for example, is washed, stemmed, and neat— ly packaged. Poultry is cleaned and dressed. More and more frequently hams are being sold ready-cooked. Coffee is already ground and packaged (with the customer's option, however, in many stores, of grind- ing it on the premises). In addition, foods ready- prepared by a variety of establishments, notably frozen food manufacturers, delicatessen stores, 4 Staff Report to the Federal Trade Commission, Economic Inquiry into Food Marketing: Part I Concentration and Inte- gration in Retailing (Washington: U.S. Government Printing Office, 1960), pp. 50-73. 1 | . xv \ 1 I I 1 I I . . , I I 1 I! v a .o . . ,1 - 7 UV fr , 1 f 1 I bakeries, and dairies, make it possible to dine at home with virtually no cooking. The increased cost that the consumer is willing to pay for partially or completely processed food provides added sales volume for the modern supermarket. A final source of growth for the supermarket industry was the addition of non-food lines to the product offering with estimates indicating that on an average from seven to ten per cent of total store sales are from non-food products.6 However, inthe next decade these four sources of growth will not offer equal opportunity for market expansion. Many industry authorities believe that the segment of growth taken from the small food retailer has reached practical limits and that the supermarket share of total food sales will stabilize at about seventy-five per cent of total sales.7 Thus, one of the prime sources of growth will disappear in the next decade. In contrast, population growth and increasing levels of quality food consumption, as well as non-food products, will probably continue to contribute to the supermarket's growth during the next decade. 5 . . . U.S. Department of Labor, How American Buying Habits Change (washington: U.S. Government Printing Office, 1959), p. 105. 6"The Importance of General Merchandise" (The Dillon Study: Part 5), Progressiye Grocer, August, 1960, pp. 57-69. 7 a I 0 ' ll ’ "Facts in Grocery Distribution. op. c1t., also: "The True Look of the Super Market Industry," Super Market Merchandising, May, 1960, p. 77. The Supermarket Industry and the Location Decision The problem of available location sites for supermarkets is rapidly becoming important in the supermarket industry. The day of widespread availability of prime sites during the early 1950's has disappeared. The location aspects of supermarket operation have re- ceived relatively little attention during the past decade. There are probably several reasons that can be advanced for the lack of interest. ‘ First, in the period of rapid growth, research efforts of the supermarket industry were focused mainly on internal opera— ting efficiency.9 In a period of rapid and profitable growth, business enterprise typically focuses its efforts on the core of the profit opportunity and either ignores completely or rele- gates to a position of secondary importance the peripheral efficiency considerations. 8The Supermarket Institute estimates that in 1940 there were 5,659 families per supermarket; by 1950, the total had dropped to 3,014 families per supermarket and, by 1959, had dropped further to 2,310 families per supermarket ("The True Look of the Supermarket Industry, 1959," op. cit.). 9"Over the period 1951-1956, when retail food store prices were relatively constant, sales per full-time store employee equivalent increased from $29,700 to $43,000. This is an in- crease of 46 per cent in sales per cent in sales per employee, a record which, it is believed, cannot be matched by an other kind of major retailing institution." (Theodore H. Beckman, Harold H. Maynard, and William R. Davidson, Principles of Marketing, 6th ed. [New York: Ronald Press, 1957], p. 225.) 5 The second reason is the orientation of the function respon- sible for acquiring and developing supermarket locations. Tradi- tionally, in the:supermarket industry, this function is performed by the real estate department. The highly complex legal problem of lease arrangements and agreements often requires a specialized legal staff for review of lease contracts and other legal matters. The prime function of the real estate department in most super- market chains is currently to develop new locations. This func- tion has expanded in importance due to the complexity of larger stores and the more competitive bidding for available sites. The net result of the increasing complexity in store development has been that the real estate manager has become a market analyst, construction superintendent, legal counsel, and property manager. The multiplication of functions without additional staff has caused, in many instances, a superficial treatment or "rule of‘ thumb" treatment of the added functions due to time and back- ground limitation.10 A third reason for the lack of emphasis on location analysis stems from a unique attitude toward competition on the part of supermarket chains. In considering market coverage, the super- market chain is more likely to consider the placement of its own stores in relation to a new site than it is to consider the O . . . ”Chains Reveal Rules-of-Thumb for ChOOSIng Store Locations," Chain Store Age, January, 1960, pp. E33-E38. location of competitive stores. This attitude is probably based on the belief that a store will capture a certain share of potential business in any market area. That is, the belief that a store will achieve "X" per cent of total available sales in its trading area regardless of the amount of compe- tition in the area. As a result, the problem becomes one of location analysis in relation to the chain's present distribu- tion and store network, and the amount of potential business available becomes a secondary consideration. There is undoubt- edly a valid basis for this consideration of "sister" stores.11 However, considering the fact that there is a limited amount of food dollars in any given area, more intensive coverage from all chains or independents operating in the area will presumably result in a smaller share of the total food dollar for each chain. A synthesis of the problem-solving analysis in supermarket location is briefly summarized as follows: Question 1. Where are the available sites suitable for supermarket sites? Question 2. What is the relationship of the proposed site to our present store and distribution structure? l . . 1 The term Sister store used here 18 a frequently used trade term referring to another store of the same chain. The term will be used in similar context throughout the dissertation. IHUIIIII-II. TO-CIFTS & EXCHANGES IV— ‘*“ ‘ “" V ”W. FOR POSSIBLE SELECTION ADD TO COLLECTION veg" / DESTINATIONfi (ZS No . of Added E COPIES NEEDED ' REMARKS: Aw.“ LOC I 41% I; W“ Q d DATE: ?:/‘ 3’0 . . . . 12 SIGNATU —{;Q{En:‘cting an increase in product offering. iof supermarkets led to increased land and 12 nat is the relationship of the proposed site 3 the competitive structure of the market area nder consideration? mat is the potential sales volume for the roposed site? hat are the possibilities of future sales rowth in the market area of the proposed site? an the site be developed within an econom- - cally justified framework? thoroughness and sophistication employed in ess described above varies widely in the super- ecting Supermarket Location st decade two trends have developed in the _ try which require a re-evaluation of super- nalysis procedures. The first is the increasing fered by the typical supermarket, which in recent considerable increase in the average size of This Staff Report to the Federal Trade Commission, op. cit., pp. 68-70. Question 3. What is the relationship of the proposed site to the competitive structure of the market area under consideration? Question 4. What is the potential sales volume for the proposed site? Question 5. What are the possibilities of future sales growth in the market area of the proposed site? Question 6. Can the site be developed within an econom- ically justified framework? The degree of thoroughness and sophistication employed in the decision process described above varies widely in the super- market industry. Recent Trends AffectingSupermarket Location During the past decade two trends have developed in the supermarket industry which require a re-evaluation of super- market location analysis procedures. The first is the increasing range of goods offered by the typical supermarket, which in recent years has shown a considerable increase in the average size of store, thus reflecting an increase in product offering.12 This increasing size of supermarkets led to increased land and 12Staff Report to the Federal Trade Commission, op. cit., pp. 68-70. IIIUII-.-_-._1_.__._.‘1l.l facility costs, which further points up the importance of proper site selection.13 A second trend, not as easily quantified as the first, is the ever increasing range of location types within a city. The competitor in today's market place must consider location alternatives ranging from a store in the central business district to a store located in a sparsely settled suburban area. The move to the suburbs and the increasing amounts of large scale competition have forced the supermarket developer to investigate site development in areas that would not have been considered in 1950. Both trends have also contributed to a stratification of stores by function and type of market area served.14 Thus, supermarkets built in 1960 are much larger than those built in 1950 and vary more widely in the types of retail complexes in which they are located. 13According to the Super Market Institute, store equip- ment and fixtures for a "typical" supermarket of 20,500 square feet cost approximately $250,000 in 1958. Including land and other costs, a total investment of between $400,000 to $500,000 was required: From: "Facts about New Super Markets Opened in ' 1958." Super Market Institute, Chicago, 1959, p. 10. Also: ”What Will It Cost To Open a New Supermarket in 1961?" Chain Store Age, December, 1960, pp. 38-39. The above article es- timates that it will cost $640,500 to open a modern 16,000 square feet supermarket (11,000 sq. ft. of selling area). 14Bart Jacob Epstein, "The Quincy Food Market: A Study in Marketing Geography" (unpublished doctoral dissertation, Depart- ment of Geography, Clark University, 1956), pp. 92-93. 9 A classification of current location alternatives for the supermarket industry include the following: 1. A store can be built in an area already serviced by one or more supermarkets and meet competition directly. This is becoming an increasingly common situation and has fostered some problems of effective and profitable merchandising. 2. Another choice is to build a store on the outlying fringes of the community and wait for the community to "grow-into" the store. This type of decision raises many questions, some of which are: 1) What size store to build, and 2) What is the rate of growth of the area? Some supermarket chains have pursued this policy of location strategy to the point where they have become shopping center developers. 3. A third location alternative is to seek location where land costs would normally prohibit a supermarket location. An unusual number of apparently successful downtown loca- tions offering no parking facilities and multi-floor shopping have appeared recently.15 4. A fourth alternative, although not formally a location problem, is acquisition of an already existing supermarket or chain of supermarkets. From a locational point of view, acquisition is a method of obtaining a site that is already occupied. Chain acquisition has become such a common practice that the government recently investigated the problem.16 Also William Applebaum and Saul B. Cohen,"The Dynamics of Store Trading Areas and Market Equilibrium," The Annals of the American Association of Geogrgphers, Vol. 51 (March, 1961). PP. 75-77. 15 . . , . "Retailers Create New Life, Profit Downtown," Food Topics, February 9, 1959, p. 2; also: "In the Shadow of the Empire State," Food Merchandising, May, 1960, p. 62. 6 . . . . . Economic Inquiry into Food Marketing," 0 . Cit. See es- pecially Chapters 4 and 5. 10 The Scope of the Problem The increasing size of the supermarket and the increasing diversity of location has caused a variety of new problems to develop in the traditional approach to supermarket location analysis. The traditional approach usually revolved around the single variable of distance. That is, in measuring the poten- tial business for a proposed site, a circle is drawn around such a site. The families residing within the circle are then enumerated and their food expenditures estimated. Family enumeration is followed by the application of a fixed percen- tage figure against the total volume of available business in the area based on historical market share, and a potential store volume is obtained. The basic error in the traditional procedure appears to be the use of distance as an independent variable. The distance a consumer will travel to get to any point is a function or dependent variable rather than independent variable. For example, a customer desiring to purchase his food requirements would probably be influenced by at least the following factors:17 1. Distance to alternative sources of supply. 17The distance the consumer would be willing to travel would probably depend upon the above factors plus many factors which are more subjective and less easily determinable. ("Factors in a Purchase Decision," Convenience Goods Purchasing: NeededpResearch [Ann Arbor: The Foundation on Human Behavior, 1957], p. 2.) ll 2. Range of products at alternative sources of supply. 3. Prices and price structure at alternative sources of supply. The guiding hypothesis of the research is that the distance a customer travels to purchase convenience goods is dependent upon product offering at the retail site. In most metropolitan areas there are probably more than a dozen supply sources for convenience goods within the daily travel sphere of the normal consumer. Given alternative sources the normal urban consumer is presented with several alternative sources for fulfilling his purchase objectives. The hypothesis holds then, that the total product offering at the retail site will influence the customer's choice of alternatives in fulfilling his convenience goods purchasing objectives. The concept of product offering as it is commonly used in marketing literature can present a very wide range of meaning.18 When used here, "product offering" refers to the number and variety of available goods and services at the retail site. Used in the above context, product offering can be viewed from two perspectives. The first perspective is the range of goods and services (product offering) provided within the retail 8Eugene J. Kelley, "The Importance of Convenience in Con- sumer Purchasing," Journal of Marketing, July, 1958, pp. 23-38. 12 unit. In the case of the supermarket, the measurement device used to reflect the range of goods and services is the square foot of selling area within the individual retail unit. The square footage measurement reflects in approximate proportion the number and variety offered and probably represents the best measurement short of an actual physical inventory. The second perspective for viewing product offering is provided by the retail stores making up the shopping area in which the individual store is located. With the rapid growth of the shopping center movement since world war II, the consumer is receiving ever-increasing exposure to the idea of "one-stop" shopping.19 Empirical evidence suggests that the consumer is willing to travel further to patronize a shopping center than for single purpose trips.20 It appears that the attraction of any individual store within a shopping area is enhanced by the fact that other stores providing a range of different products or services surround it. Thus, it is postulated that in the case of a shopping center or retail cluster the combination of stores possesses an attraction 9 . . 1 Chain Store Age estimates that there are currently 4,500 shopping centers in operation in the U.S. with an additional 1,000 scheduled to be opened in 1961. ("Centers Open on Target in '60," Chain Store Age, January, 1961, p. E26.) 20William L. Garrison, et a1., Studies of Highwaprevelop- ment and Geographic Change (Seattle: University of Washington Press, 1959), Chapter 11. .I I 4 w . 13 to the consumer that is greater than any of the stores, taken individually, due to a large total product offering. These two perspectives for viewing product offering pro- vide the independent variables for the study. Store size, measured in square feet of selling area, provides the variable ‘designed to measure the influence of product offering within the individual store. The product offering created by multi- ple retail units clustered at a geographical point or focus provides the second independent variable for this research. The term store complex is used to identify product offering when it refers to a combination or cluster of stores offering similar and dissimilar products. Problem Statement The objective of the study is to investigate empirically the influence of store size and store complex upon customer attraction and per capita sales. Answers are sought for the following questions: 1. What relationship does store size have to the distance traveled by the customer for food purchasing purposes? 2. What relationship does store complex have to the distance traveled by the consumer for food purchasing purposes? 3. What influence does store size have on the per capita sales of a supermarket? I ._II l4 4. What influence does store complex have on the per capita sales of a supermarket? The research can be termed a pilot inquiry into the problem areas as outlined above. On the basis of the fifteen observations in the study, certain generalizations are pre- sented, and guide lines for further research are established. General Hypotheses The major hypothesis is: The variables of store size and store complex are significant variables in influencing the consumer's decision on the distance he will travel to fulfill his food purchasing objectives. The major hypothesis, then, posits that of the range of variables influencing purchase behavior, the two selected are significant in determining the nature of space preferences for the food buying objectives of the consumer.21 The specific hypotheses relating to store complex are: 1. As the product offering at a retail complex increases (as measured by number and types of different stores), the drawing power of the supermarket increases. lDavid L. Huff proposes an interesting model for the treatment of the full range of variables influencing the con- sumer purchasing decision. ("A Topological Model of Consumer Space Preference," Occasional Paper No. 11, Bureau of Bugipegg Research, University of Washington, Seattle, December, 1959.) I'll-IIIIII-IIIIIIIIE 2. The The The permits 15 A small town relatively isolated from any other city demonstrates drawing power patterns Similar to the medium sized shopping center (community). AS the product offering at a retail complex increases (as measured by number and types of different stores), per capita sales of the supermarket increase. specific hypotheses relating to store Size are: As the size of store (as measured by square feet of selling area) increases, drawing power of the super- market increases. As the size of store (as measured by square feet of selling area) increases, per capita sales of the supermarket increase. variables of store size and store complex are regarded as independent variables with the drawing power and per capita sales variables regarded as dependent variables. The research design is structured so that the hypotheses are verified or disproven on the basis of the relationship of the dependent variable to the independent variable. Method of Investigation22 specific problem is to develop a methodology that the measurement of drawing power and per capita sales 22A complete outline of methodology can be found in Chapter "Research Design." 10 I I. ' I _ . , ,_ 16 over a range of store sizes and store complexes. The survey stores selected for study are fifteen in number and range in size from 4,000 to 16,800 square feet of selling area. The same survey stores are used to investigate both the variables of store size and store complex. In a large urban area, the variety of store complex sit- uations probably range from the isolated retail unit23 to the central business district. The entire range of store complex situations are not con- sidered but, rather, a selected portion of the most common types of stores are analyzed. The relevant range is from the Urban Strip store to the store located in the regional shop- ping center. Selected types within the range are singled out for analysis. For purposes of verification or rejection of the hypotheses, these stores are ranged on the following scale: Store Type No. of Survey Stores 1. Urban Strip 3 2. Urban Cluster 3 3. Small Town 3 4. Neighborhood Shopping Center 2 5. Community Shopping Center 2 6. Regional Shopping Center 2 15 23 . . , , . The isolated retail unit With no other retail stores around it is usually termed a "free-standing location." It is used in this context in this study. :III-IIl-IIII-III:H1HIIII 17 The above ranking is on an ordinal rather than a cardinal basis. That is, it indicates that the Urban Strip type of store is lower on a continuum of store complex than any of the other types, but makes no estimate on how much lower. Similarly, the Urban Cluster store is lower than any of the four stores above it and higher than the Urban Strip type on the store complex continuum. As outlined above, these stores are ranked on the basis of the product offering provided by the store complex. Hence, the logic of the above scaling would imply that product offering increases over the range from the Urban Strip type to the Re- gional Shopping Center type. The majority of the survey stores are located in the metropolitan central city. However, for purposes of compari- son and representative store complex structure, a number of the survey stores are located in outlaying areas. The study included outlets of one large regional super- market chain. In using one chain for survey purposes the price and non-price aspects of the product offering of the individual store are held constant, with the exception of individual differences in the store manager”s ability. If these factors are held constant, it is probable that more reliable and comparable data will result. 18 The field investigation phase of this research consists of one interview per one hundred dollars in sales per week in each of the survey stores. The interviews were conducted on a Friday and Saturday, which, besides being the largest volume days, are probably the most typical shopping days of the week. The interview replies were recorded on a card designed for this purpose and the customer's address recorded by the interviewer.24 The next step in the research procedure is a plot of each customer's home address on a map of the store area. The dis- tance of the customer's home from the supermarket was measured and recorded. From this data, average drawing power (mean average distance traveled) data were calculated and related to the in- dependent variables of store size and store complex. In order to calculate per capita sales, population estimates were made of a two-mile area surrounding the survey site. Both 1960 Census Tract and Enumeration District data were used to calculate population for this area. The last phase of this research consists of evaluating the data in light of the hypothesis and presentation of findings. There is a commonly used methodology for this step in the research design. See: Bart J. Epstein, "Evaluation of an Established Planned Shopping Center," Economic Gepgraphy, January, 1961, pp. 12-21; or: William Applebaum and Richard F. Spears, "How To Measure a Trading Area," Chain Store Age, January, 1951. _ 19 Terms and Definitions Supermarket - A retail establishment with a self-service grocery department and meat, dairy and produce department doing a com- bined volume of at least $500,000 per year. Survey Store - The term survey store refers to a store selected for study in this research. There are fifteen such stores divided into six classifications of store types. Eganned Site - A planned site refers to a shopping center which is developed and built according to a prepared plan which pro- vides for a balanced number of retail outlet and consumer con- veniences. Unplanned Site - An unplanned site is a location that is developed in an area with no prior planning or control for the purposes of balance in number and types of retail outlets and customer facilities. In an unplanned area site avail- ability over a given time period provides the basis for the range of goods available. Store Complex - Store complex refers to the complex of stores surrounding the survey store° When used in references to shopping centers it connotes the complete range of stores within the individual shopping center. When used in reference to unplanned sites, it refers to the retail stores located ijfliHHHHMMMMIITIIII 20 within 1/3 mile of the survey site (walking distance). When used in reference to a small town, it refers to all of the retail stores comprising the business district of the town. Store Type - Store type refers to a particular classification of stores possessing certain measurable location character- istics. The primary characteristic delineating one store type from another is the retail complex in which it is located. Small Town Store - A small town store is a supermarket located in a town of less than 5,000 population.25 ‘grban Cluster Store - An Urban Cluster store is a supermarket located within walking distance (1/3 mile) of a major business intersection where shopping goods are available for purchase.26 urban Strip Store - An Urban Strip store is located on a major traffic artery in an urban area. The urban Strip site is surrounded by other convenience type retail stores.27 Shopping Centgp - The term "shopping center" used in this research indicated a center of the controlled or planned variety possessing the following characteristics: See page 67 for selection criteria. 6 2 See page 67 for selection criteria. 7See page 67 for selection criteria. .. r 1 .1 u 21 1) "Land on which the center is situated is owned by a single owner. 2) An integrated assortment of different retail outlets offering a balanced representation of goods and services is featured.29 3) ”Planning is done in advance of construction. The completed shopping center is designed as an integrated, harmonious unit."30 4) Off street parking is provided for customers of the center. Usually the parking ratio exceeds 3:1 except in the case of very small centers. Regionai_§hopping Center - A Regional Shopping Center possesses all of the characteristics indicated in the definition of shop- ping centers. In addition to these requirements, there are in excess of fifty separate retail units within the center which are dominated by a complete full size department store. Community Shopping Center - A Community Shopping Center posses- ses all of the characteristics listed under the definition of shopping center. This type of center contains between 16-49 2 . 8Eugene J. Kelley, Shopping Centers: Locating Controlled Regional Centers (Saugatuck: The Eno Foundation, 1956), p. 4. 291bid., p. 4. 30Ibid., p. 5. -_-_-1-‘-. 22 retail units and is dominated by a large branch department store or departmentalized specialty store. Neighporhood Shopping Center — A Neighborhood Shopping Center possesses all of the characteristics of the planned shopping center. It is designed to accommodate approximately 15 stores offering a balanced assortment of convenience goods and services and dominated by a supermarket.31 Store Size - Store size refers to the total number of square feet of selling area within the supermarket. Drawing Power - Drawing power is defined as the average main distance traveled by seventy and ninety per cent of the stores' customers. Drawing power is calculated by both seventy and ninety per cent in order to provide a relationship of concen- tration and dispersion. Per Capita Sales - Per capita sales is defined as the dollar amount of sales per person (not family) per week within a given geographic area. 31 . . . . There is no ba51c agreement on the exact criteria and proper nomenclature for the planned shopping center. Additional insights can be obtained by referring to: Max 8. Wehrly, and J. Ross McKeever, Eds., The Community Builders Handbook (Washing— ton: Urban Land Institute, 1954), p. 122; Paul E. Smith, Shop- ping Centers: Planning and Management (New York: National Retail Dry Goods Ass'n., 1956), pp. 17—18; Gordon H. Steadman, "The Rise of Shopping Centers," Journal of Retailing, XXXI, No. 1 (Spring, 1955), pp. 14-15; Eugene J. Kelley, o . cit., pp. 4—8. 23 Distance Interval32- The term distance interval identifies a specific distance zone in relation to the survey store. The three distance zones used are the 1/2 mile, 1 1/4 mile and 2 mile distance zones. The term denotes the geographic distance extending from the last distance interval line to the one specified. Example: The term 2 mile distance interval indi- cates that distance between 1 1/4 mile and 2 miles from the survey store. The term 1 1/4 mile distance interval indica— tes the distance between 1/2 mile and 1 1/4 mile from the sur— vey store. The term 1/2 mile distance interval indicates the area between the survey store site and 1/2 mile. Quadrant33- The area surrounding the survey store is divided into four equal quadrants for purposes of analysis. A verticle line is drawn directly north and south. A horizontal line in- tersecting at the survey store site is drawn north and south, dividing the area into four equal segments. These segments are labled quadrants 1, 2, 3, 4 and have the following charac- teristics: Quadrant No. Direction Degrees 1 Northeast 270:— 360: 2 Southeast 1810- 270 3 Southwest 910- 180: 4 Northwest - 90 2 O O O O O A graphic presentation of this procedure is presented in Appendix C—4. 3 . . . . . A graphic presentation of this procedure is presented in Appendix C-2, and C-4. 24 Customer Spottinngap - A customer spotting map is a street map on which the residences of the survey store's customers has been identified. Limitations The limitations of the study are as follows: The research assumes equal consumer time-distance mobility in all directions. The variable of site accessibility is not explicity integrated into the research. In the selec- tion of the sample, care was taken to select stores where time-distance mobility appeared to approach equality. This factor, to some degree, is self-adjusting in that increasing population density generally results in decreasing time- distance mobility and increased amounts of competition. The research is limited to the supermarket industry, and research results pertain only to this industry. It can provide insights into the entire range of convenience- goods locations problems, but implications for other areas cannot be scientifically justified on the basis of this study. The research is concerned only with the spatial relation- ships of the customer to the retail site. On the part of the customer, there are social and psychological consider- ations in his shopping behavior that can affect the extent .- Iii-la“ I. I I I I. I. I 25 of the trading area. On the part of store management, there are considerations of special promotions, store image, etc. that can affect the size of the trading area at any given time. 4. The research design, as presented, is essentially static in nature. That is, it is conducted at one point in time under given market conditions. While this represents a weakness to some extent, in most urban areas, the market place is subject to gradual rather than violent altera- tions in its basic structure. Some Possible Contributiong of the Study In January, 1961, at the mid-year meeting of the Super- market Institute, Mr. Curt Kornblau, Director of Research, made the following statement: Nearly two out of every three new supermarkets (62%) are doing less business than expected. . . The difference between actual and estimated sales is quite substantial in many cases, ranging from 54 per cent below forecast to 49 per cent above. Supermarkets with sales better than expected averaged 13 per cent above the estimate, and supers with sales less than expected averaged 20 per cent below the estimate. All the new supermarkets combined averaged sales 10 per cent less than predicted.34 4 . . . . "A New DimenSion: Economics and Marketing Geography," Food Topicp, March, 1961, p. 7. 26 There are probably a variety of reasons for these faulty sales forecasts. Undoubtedly one of the major causes is the heavy reliance placed upon intuition and rules of thumb for site selection in the supermarket industry.35 This evidence suggests that the decision to develop the size and number of supermarkets currently being built by some chains is being made on the basis of promotional differentiation rather than economic justification.36 There is generally a premium rental or price on more developed and desirable locations. Currently there is some question in the supermarket indus- try regarding size and number of stores.37 Annual surveys by Supermarket Merchandising indicate that the average square footage (including basements of new supermarkets opened) has increased from 8,000-9,000 in 1949 to 17,000 in 1958.38 The question involved in the 50-per cent increase in store size may be restated as: Is store size being decided on an economi- cally justified basis or is it being used in individual situa- tions to overpower a competitor with bigness? 3SIbido I pp. 8-100 36Bob R. HOldren has produced a very interesting study of market structure with implications for economy of scale using the supermarket case as a decision model. (The Structure of a RetaiigMarket [Englewood Cliffs: Prentice-Hall, 1960].) 7 . . . . InterView Wlth Dr. Saul B. Cohen, Boston UniverSity, January 18, 1961. 38 . . L . . Economic Inquiry into Food Marketing, op. Cit., pp. 56-58. _ _. _- _ _‘ I _ II I I _w .1 27 The research, by examining the relationship between store size and customer attraction and per capita sales, should pro— duce some insights into the scale of store decision. Presumably, many market opportunities are ignored because they will not support the volume of the large-scale supermarket currently being built. By varying the scale of store to correspond to the scale of market opportunity a more efficient individual operation results and a more healthy industry situation envolves. The second variable under study, that of store complex, should provide some insights into the economics of location in various types of store complex situations. There is gen- erally a premium rental or price on more developed and desir- able locations. The research, by assessing the relative drawing power and per capita sales of different types of retail clusters, should allow a more informed judgement to be made on the question of the value or rent differentials or property values in relation to anticipated revenues. A third contribution of this study could be the intro- duction of a practical methodology leading to greater effi- ciency of operation for the supermarket industry. Some super- market chains are very aware of the value and uses of location analysis. But, unfortunately, these chains are so few as to represent exceptions to the normal state of development for location analysis in the supermarket industry. The research 1 If] I I II-I:-I I I. I I M I I I 28 could function as a pilot study for supermarket chains to duplicate and expand to fit their own particular needs and prOblems. A fourth contribution of the research could be a new perspective for evaluating growth potential. With the market becoming increasingly saturated in the more densely populated areas, supermarket developers have, in recent years, been faced with the problem of evaluating prospective growth areas for potential market development. This research would provide some indication or guideline for anticipating the level of market opportunity when the marketplace has reached advanced stages of maturity. A final contribution of this study is the integration of a number of theoretical concepts and practical application developed in other disciplines into the field of marketing. With few exceptions, the pioneering work in retail location analysis has been done in disciplines outside the field of marketing. Organipation oiithe Study The introductory chapter, Chapter I, presents the back- ground and rationale for the research. It seeks to set in proper perspective the research problem in relation to the supermarket industry and in relation to the more general field of marketing theory and practice. 29 The objective of Chapter II is threefold: l) to present some of the perspectives employed in the theoretical and em- pirical analysis of consumer space preferences and the trading area, 2) to review some of the operational criteria used to delineate the perimeters of the retail trading area, and 3) to review, in the process of achieving the first two objectives, literature relevant to the area of the research. In Chapter III, the research design is presented. The primary objective of Chapter III is to construct the design in such a manner that any supermarket with a similar research problem can utilize the design as a methodological base. The concepts and terminology used are operationally defined, and the precise methodology and survey methods are outlined in Chapter III. It provides the operational basis for the em- pirical portion of the research. The presentation of findings are found in Chapter IV. The results of the completed research design are tabulated and presented using appropriate tables, charts, and diagrams. The presentation provides the factual basis for the evaluation of hypotheses in the following chapter. The analysis and conclusions of the research are presented in Chapter V. .The research findings are evaluated in terms of the hypotheses presented in Chapter I. Conclusions are pre- sented in Chapter I. Conclusions are presented on the 30 reliability of the hypotheses, the research design, and the findings of this research. Implications for further research and extentions of the study are offered at the conclusion of Chapter V. 31 CHAPTER II CONSUMER SPACE PREFERENCES AND THE RETAIL TRADING AREA Introduction In recent years much attention has been focused on the changing structure of the metropolitan area. These changes have been presented mainly in terms of shifting population concentra- tions. However, the impact of such population shifts in turn initiates a whole series of secondary adjustments designed to meet the needs of the changing population concentrations.1 One of the most significant adjustments occurs in the retail structure of the community. The retail structure exists to serve the population of the area. Since it serves a definate need, it shifts in response to a change in that need. Many studies have indicated that the retail structure of a community arranges itself functionally so as to best serve the needs of the communi- ty and systematic patterns of functional relationships have been empirically established for the retail structure of the community.2 1'. . ' ‘ . . William M. Dobriner, The Suburban Community, G. P. Putnam's Sons (New Ybrk), 1958, and Research Monograph No. 2, Metropgli- tanization of the United States, Urban Land Institute (Washington, D.C.), 1959. 2Richard U. Ratcliff, "The Problem of Retail Site Location," University of Michigan (Ann Arbor), Michigan Business Studies, V01. IX, No. l, 1939. Also: Malcolm J. Proudfoot, The Major Out- lying Business Centers of Chicago, University of Chicago, 1938. 32 The decentralization of the urban population within the metropolitan area has caused a similar decentralization in the retail structure of the community.3 Probably the out- standing symptom of decentralization has been the rapid growth of the suburban shopping center. Just as significant, though not as obvious, has been the growth of the retail structure along suburban streets and highways.4 The decentralization of the retail structure has prompted an increased interest in location and location analysis by the forward thinking retailer. The retailer, in adjusting to the pressure toward larger scale retailing and the move to the suburbs of great numbers of his customers, has been faced with a great variety of problems.5 Net the least of these problems has been what type of stores to build and where to build new stores. The suburban market, the suburban customer, and the suburban retail struc- ture was a new and unfamiliar arena of competition for many merchants with traditional roots in the downtown commercial districts.6 3Edward F. Staniford, Business Decentralization in Metro- politan Log Angeles (Los Angeles: Bureau of Government Research, university of California, 1960). 4 E. B. weiss, Highwgy Retailing--The Next Great Retail Revoiption, (New Yerk: Doyle-Dane-Bernbach, Inc., 1958). 5 . . . . John W. Wingate and Arnold Corbin, Changing Patterns in Retailing (Chicago: R.D. Irwin, Inc., 1956), particularly Parts I, II, and X. 6 C.T. Jonnassen, Downtown vs. the Suburbs (Columbus: Ohio State University Press, 1955). See Appendix. 33 During the post-war decentralization of retail business, most firms at one time or another probably faced the following questions: 1. Where can I best locate to serve potential customers without weakening present market position?7 2. What type and size store should be built consistent with merchandising policy and long-range planning for market development?8 3. Within the framework of the preceding two questions, from where will our customer come? Will the downtown location lose volume to the suburban store? The above questions are probably more easily resolved for retailers of shopping and speciality goods9 than for those in convenience or service goods. The retail institutions selling, shopping or specialty goods presumably would make fewer decisions and would be relatively more limited in alternatives due to the nature of these types of retailing. 7A schematic diagram of the J. L. Hudson Co. (Detroit) master decentralization plan is presented, and this question is dis- cussed in: Victor Gruen and Larry Smith, Shopping Towns U.S.A. (New Ybrk: Reinhold Publishing, 1960), pp. 35-37. 8Ibid., p. 37. 9 . . . . The terms for convenience, shopping, and speCiality goods, are used here in the traditional marketing context (Marketing Definitiong: A Glossary of Marketing_Termsf Chicago: American Marketing Association, 1960). 34 The retailers of convenience goods are faced with the problem of locating so that they are convenient to their market. They are usually serving small segments of the mar- ket and hence must develop discriminating techniques for market delineation. These techniques assume importance as the metropolitan area becomes saturated with convenience stores, leaving smaller and smaller segments of the market open to market development. The convenience goods retailer is faced with a twofold problem in evaluating a proposed site. He must first evaluate the site on the basis of its profitability as an individual site. He must determine if enough potential sales volume exists at the site to justify its development. A second question is posed by the proposed site's relationship to his present market structure. Both of these problems probably can be resolved by an analysis of the trading area of the proposed site.10 Unfor- tunately, it is more difficult to delineate the trading area for a proposed site than an existing site. The process of delineating the trading area is, in its simplest form, an attempt to establish perimeters for con- sumer space preference. These perimeters can be viewed from See Appendix A for some commonly used approaches to the delineation of a trading area. .I I '1' I. I I“ IN I Ir I 4 I IL I‘ 35 various perspectives, and the objective criteria for establishing them can consist of one or a combination of factors influencing both consumer travel patterns and consumer travel patterns converted into potential store profitability (e.g., per capita sales). The degree of thoroughness and the level of analysis can vary from the simple one-variable distance measurement to a field interview sample of the projected trading area. The primary purpose of Chapter II is to present some of the perspectives employed in the theoretical and empirical analysis of consumer space preferences and the trading area. A second purpose is to review some of the operational criteria used to delineate the perimeters of the retail trading area. The third purpose is to review, in the process of achieving the first two objectives, some literature relevant to the area of this research. Some Perspectives for Consumer §pace Preferences The consumer is spatially situated11 and must satisfy his economic wants within an imperfect market. This imperfection derives from two sources. The first source is geographical in nature. The consumer does not face an equal time-distance l . . 1The term "spatially Situated" refers to the fact that the consumer is located at some point in the marketplace and must overcome the "friction of space" to move in any direction from the point of location. I I I I It Ilr In I I I! I I I I I I I II: 36 movement rate in every direction. The urban area is crossed by rivers, limited access freeways, industrial developments, etc., all of which influence and limit the consumer's spatial mobility. A second source of imperfection exists in the consumer's perception of the marketplace. The consumer usually does not possess perfect knowledge of the market and the product offerings of the market place. His knowledge probably ranges from complete certainty to complete uncertainty over the range of goods that he would normally purchase. None the less, the consumer is faced with satisfying his economic wants within a relatively restricted framework. He must satisfy these economic wants over time, which is a limited commodity with many alternative uses.13 Hence, it can be concluded that the consumer would probably arrange his shopping patterns in such a manner that he will receive an.op- timum amount of utility for the time and effort expended. While this is a very convenient generalization and probably true as a generalization, several definitional problems arise. The paramount problem with the above conclusion seems to be the concept of utility. It has been empirically demonstrated that shopping pattern behavior and satisfactions are a function 2 . . . l MeIVin L. Greenhut,"Space and Economic Theory," Regional Science Association, Papers and Proceedings, Vol. V (Chicago, 1959). pp. 268-72. 3Frank B. Curran and Joseph T. Stegmaier, "Travel Pat- terns in 50 Cities," Bulletin 203, Highway Research Board (Washington: National Academy of Sciences--National Research Council, 1959). pp. 111-19. 37 of factors in addition to distance.14 Thus, it follows that maximum or optimum utility cannot be measured exclusively in terms of distance. There is also evidence that the consumer views different types of goods in a different perspective as regards the psychological context of shopping behavior.15 A study in Houston, Columbus, and Seattle also substantiated this fact, indicating that an average of 98.9 per cent of the food purchased during a given period of time was purchased at the suburban shopping center, while only 32.2 per cent of the clothing and 41.9 per cent of the furniture were purchased in the surburban shopping center by the same families.16 A study conducted in the washington, D.C. metropolitan area indicated , , 17 Similar results. 14C. T. Jonassen, "Shopper Attitudes" Special Report ll-A, Highwangesearch Board, (washington: National Academy of Sciences --National Research Council, 1955), pp. 18-24. 5William H. Form and Gregory P. Stone, "The Local Commun- ity Clothing Market: A Study of the Social and Social Psychoe logical Contexts of Shopping," Technical Bulletin 262, Michigan State University, Agricultural Experiment Station, East Lansing, Michigan, 1957. l6Jonassen, "Shopper Attitudes" Special Report ll-A, op, cit., p. 11. 17 . Alan M. Voorhees, Gordon B. Sharpe, and J. T. Stegmaier, Shopping Habits and TraveliPatterns, Technical Bulletin No. 24 (Washington: Urban Land Institute, 1955), pp. 20-24. Also, Gordon B. Sharpe, "Travel to Commercial Centers of the washing- ton, D.C. Metropolitan Area," Bulletin 79, Highway Research Board, (washington: National Academy of Sciences-- National Re- search Council, 1953). 38 The objective of the retailer should be to locate in such a manner so as to offer consumer utility to the extent that, at a minimum, a threshold level of sales will be attained.18 The retailer, then, must strike a balance between profitable operation and consumer convenience.19 One added burden for the retailer is that he is required to make the location decision in what might be termed a competitive vacuum. At best, he can only guess what competitive actions will be taken to counteract his location decision. A competitor by appropriate location strategy can disturb the balance between profitable operation and consumer convenience. This disturb- ance is often caused by the uninformed competitor who enters a market without regard for the potential business necessary to profitably support more than one competitor. In the case of the inimformed competitor situations might well emerge where both competitors units result in unprofitable operations 18The "threshold" level of sales refers to that level which will provide the minimum amount of sales volume for survival over the long run. 19Robert D. Lundy explores the question in relation to gasoline service stations and concluded that "if the criter- ion of adequacy is to maximize the profits of individual sta- tion operators, anything more than approximately 2,000 stations would have been too many in 1946. . . On the other hand, four million or even five million stations might have been necessary to provide gasoline plus maximum convenience for customers." ("How Many Service Stations Are 'Too Many'," Reavis Cox and wroe Alderson, eds., Theory in Marketing, (Chicago: R. D. Irwin, Inc., 1950), p. 333. ._T_ I! In... till!“ , n ‘ ,JLw , n ‘ -I (,1 . , , rill! 7 39 regardless of the thoroughness of the original retailer in choosing his location.20 In addition to the pragmatic considerations, one of the conceptual problems in operationally defining a trading area is the diverse meanings of the trading area concept to dif- ferent disciplines. The economist often considers the trading area as a "per- fect" selling zone with its boundaries determined by plant location and and transport costs. The individuals within the trading area are not subject to promotional or psychologi- cal pressures, but react in the best traditions of the "eco- nomic man" in their purchase behavior. The geographer is likely to view the trading area as a complex geographical phenomenon in which the geographical imperfections of the area determine the direction and degree of flow of the inhabitants. Fortunately, this is not completely true in every case; a branch of geography termed marketing geography is making some contributions in integrating the geographic perspectives of the trading area concept with the practical problem of store traffic.21 This particular problem was raised on several occasions during personal interviews with supermarket chain executives. It might be paraphrased as: The best laid plans of mice and men . . can be upset by a reckless and uninformed competitor. 21William Applebaum and Saul B. Cohen, "The Dynamics of Store Trading Area and Marketing Equilibrium," The Annals of the American Agsociation of Geographers, VOl. 51, No. 1, (March, 1961), pp. 73- 101. 40 A third, somewhat blurred perspective is presented in the marketing literature. This blurred perspective probably results from the common practice of accepting the consumer as "spatially- given" and concentrating on other aspects of demand creation. In most basic marketing texts, the question of consumer space preferences is handled by a presentation of Reilly's Law of Retail Gravitation22 with some of the proposed alterations and implications of the gravitational approach.23 In marketing literature, the individual space preferences of the consumer appear to be implicitly, rather than explicitly, regarded as a function of the product offering. To some degree, increasing suburbanization of the consumer and resulting retail decentralization have focused new attention on the usefulness of the trading area concept. However, much of the renewed in— terest has been directed toward research in shopping center evaluation and development with little emphasis of individual store trading area.24 2The original formulation of Reilly's Law of Retail Gravitation can be found in William J. Reilly, "Method for the Study of Retail Relationships," Bureau of Business Re- search, Research Monograph No. 4, University of Texas Bul- letin No. 2944 (Austin: University of Texas Press, 1929). p. 16 23 . . . E. Jerome McCarthy, Basig Marketing: A Managerial Approach (Chicago: R. D. Irwin, Inc., 1960), pp. 378-81. See: Victor Gruen and Larry Smith, Shopping Towns, U.S.A.: The Planningpof Shopping Centers (New YOrk: Rein- hold Publishing Corporation--Progressive Architecture 41 In summary, it appears that no coherent direction in the study of spatial aspects of consumer behavior has been developed in the field of marketing since 1929. Advances have been made in several disciplines and by several individual retail firms or groups of firms in a greater understanding of the spatial aspects of consumer behavior. The following section outlines in more detail the various approaches and perspectives for re- tail spatial behavior and retail trading area delineation. Some Economic Contributions Micro-economic theory has been oriented primarily to the economics of the firm and, as such, does not contain a great deal of relevance to the question of consumer space preferences. Hewever, in recent years, certain developments in the field of economics have contributed conceptual insights, both directly and indirectly, to the area of consumer space preferences. In economic literature, the trading area is generally con- sidered as a sub-problem or derivative of the general location problem.25 The location problems are generally proposed 24Library, 1960); Eugene J. Kelley, Shopping Centers: Locating Controlled Regional Centers (Saugatuck: The Eno Founda- tion for Highway Traffic Control, 1956);Pau1 E. Smith, Shopping Centers: Planning and Management (New York: National Retail Dry Goods Association, 1956): Geoffrey Baker and Bruno Funaro, Shopping Centers: Design and Operation (New Yerk: Reinhold Pub- lishing Corporation--Progressive Architecture Library, 1951). 5 . walter Isard, Location and Space Economy (New Yerk: Wiley, 1956). — -I I. Fr ll... -I—.. IL Illif I I I I I I I I I I, lllnll 42 with the following assumptions: 1) location of supply points and location of receiving points are given, 2)transportation costs are fixed, and 3)the amount supplied and the amount demanded are given. The problem then becomes one of solving for the least cost location within a given area or region.26 A formulation of the trading area problem presented by Fisher allows for variation in costs and economics of scale at the supply- ing site.27 As such, the above cited examples do not represent models of consumer space preferences but, rather, seek to find optimum location and location relationships under given conditions. The consumer is treated as spatially given and his behavior deter- mined exclusively by product price and proximity to the supply site. Given the following assumptions, a modified model of perfect competition can be constructed. 1. Identical stores with identical product lines (products handled by no other retail outlets). 2. The purchasers have identical tastes and rates of con- sumption. 26Marcello L. Vidale, "A Graphical Solution of the Trans- portation Problem," Operationngegearch, Vol. IV, 1956, pp. 193- 203: or Paul A. Samuelson, "Spatial Price Equilibrium and Linear Programming," Amerlgan Economic Review, 1952, 42, pp. 283-303. walter D. Fisher, "Economic Aggregation as a Minimum Dis- tance Problem," Econometrica, Vol. 25, No. 3, p. 363. 43 3. Equal population and income distribution. 4. Transportation costs (supply) are zero. 5. Identical time-distance travel mobility in all directions. 6. Identical retail operating cost structure. An example of the area served by a store in this system is given in Figure 1. FIGURE I TRADING AREA UNDER PERFECT COMPETITION28 28The algebraic proof for this hexagonal shape of the market can be found in:‘August Losch, The Economics of Loca- tion, Trans., 2nd RevglEd., Wblfgang F. Stolper.(New Haven: Yale University Press, 1954), Ch. X, pp. 109-23. -mummmlllllmnmummmnunfll 44 The economic concept of the "perfect" market is related to the marketing classification of convenience, shopping, and speci- ality goods. In Figure l, a series of hexagonal nests are pre- sented. The entire area within the largest hexagon represents the market for specialty goods, and, in an urban area, point A is the central business district. The adjoining "nests" of hexagons could represent shopping goods centers, and point B is either a secondary business district or a community shopping center. Each of the individual hexagons (C) represent con- venience goods trading areas with the supply point being the center point of the hexagon. If the above assumptions hold. Figure 1 represents a perfect market from both a marketing_and' economic point of view. Losch introduces another concept that hasysome direct theo- retical bearing on the trading area and its use in marketing strategy. Losch termed the diagram, illustrated in Figure 2, a demand cone.29 The diagram is used by Losch to demonstrate FIGURE 2 LOSCH'S DEMAND CONE 29Ibid., p. 106. I I-4I-I I I 45 demand intensity for the product (in the case cited, beer) as a function of distance and supply cost. Point OP is the center of the cone and that point at which the brewery is located. At this point, the users pay the established price for a barrel of beer. As the beer is shipped in the direction of R, the buyers must pay established price plus transport cost, which in the case cited, causes the demand intensity (amount purchased) to diminish. Thus, Losch introduces the example of price elasticity of demand in its relation to the trading area. Hewever, if the brewery instituted a one-price policy .throughout its entire market area, the "demand cone" would probably still serve to illustrate the relative degrees of demand intensity. This would be due to the influence of two factors. First, unless it is assumed that the brewery in ques- tion is in monopolistic position, the consumer would have al- ternative sources of supply. If the alternative sources of supply were located in close proximity to the consumer, at some point, he would probably be indifferent to Which brewery he patronized. A second, and probably more important, factor is the "friction of space" that the consumer must overcome. Regardless of the price, a consumer would probably not travel over a certain distance to purchase a product that was not a necessity. Viewed in another perspective, the utility of the product to the consumer would probably diminish as the friction of space to be overcome increases. 46 In summary, then, it seems that Losch's concept of the "demand cone" would hold regardless of the pricing policies of the producer. Some Consumer Utility Contributions The consumer utility perspective for consumer space pre- ferences is oriented to the consumer rather than the firm and considers the factors that generate utility and disutility in consumer travel patterns. An interesting theoretical formulation of this approach is presented by Troxel.30 Troxel begins by making the assump- tion that the home is the "primary place of demand expression."31 With the home as the reference point, the family becomes the basic aggregate unit for travel planning. Elapsed time is selected as the unit of measurement on the following basis: Time is a common quantitative relation between the home and non-home alternatives. Using some amount of time to travel beyond the home position, a person or family for- goes something that otherwise is obtainable at home in the same period.3 Working with these assumptions, Troxel turns to a graphi- cal presentation of his argument. 4 0 Emery Troxel, Economics of Transport, (Rinehart: New York, 1955), Chapter VII, pp. 144-68. 31Ibid., p. 145. 32Ibid., p. 150. ILIIIIVIIIIIIIII I I, I I I I I I I I I I I I 47 FIGURE 3 TOTAL TRAVEL PRODUCTS H m >m m u H 0 E423 H'g At Available Speed m H +:m O E Elapsed Hours of Time As Figure 3 indicates, the total travel product increases as time away from home increases. It should be noted, however, that the time away fnom home increases at a decreasing rate as the hours of elapsed time increase. FIGURE 4 TOTAL LOSS OF AT-HOME SATISFACTIONS g (52* Total Loss of At—Home Satisfactions ,//Elapsed Hours of Time Next, Troxel (Figure 4) considers the losses of satis- faction and products caused by travel. Time spent traveling could be spent at home or in other pursuits yielding satis- factions or alternative products. The portion below the Y 48 FIGURE 5 FIGURE 6 NET TRAVEL PRODUCTS MARGINAL NET PRODUCTS Net Travel Products m 4.) U 5 "O o H m l... NTP g m g: <-MNP H m C.‘ -H O1 E i l T A l , Elapsed Hours of Time Elapsed Hours of Time axis recognizes the fact that for some people there is a net loss at being home too much. If Figures 3 and 4 are combined, a net travel product can be calculated, as in Figure 5. As indicated in Figure 6, the traveler reaches a point (T1) where additions to travel result in losses of NTP. From the slope of the NTP curve (first derivative) a marginal net product curve can be derived. At point A on the curve, it would be necessary to pay the traveler to travel further in order to offset the losses in product satisfaction. Point A corresponds to point T in (Figure 6). l 49 While Troxel's analysis is based upon logic rather than empirical data, it provides an interesting framework for the analysis of consumer space preferences. Another theoretical contribution has been made by Ide and Baumol, who have apparently developed their ideas in connection with the program of Alderson and Sessions, management consultants. Progress on the theoretical developments has been reported in at least two places.3 The Ide and Baumol formulation for determining consumer space preferences begins with a consideration of the probability of a customer finding what he desires to purchase at any given location. An assumption of constant price and promotion is made at the outset. The probability that the consumer will find some mix of items necessary to make his shopping trip suc- cessful is established as P (N), where N is the number of vari- eties sold by the retailer. Thus, a complete probability of success would be P (N) a 1 and complete probability of failure P (N) =0. Next, a consideration of shopping costs is introduced. These costs are divided into three segments. The entire cost of shopping is represented by: cdD + cn VIN + ci.. The first segment of cost (cdD) represents the cost of the consumer in 33 W. J. Baumol and E. A. Ide, "Variety in Retailing," Management Science, October, 1956, Vol. III, No. 1, pp. 93-101: and "Basic Report on Consumer Behavior," Alderson and Sessions: Philadelphia (mimeographed). 50 getting to the store. These costs are assumed to be proportional to distance (D); hence, cd is a constant. The second segment of cost represents the (cost) difficulty in finding the items after the customer enters the store. If the store carries a great variety of items, the customer must walk further to find the one that he seeks. For a one-story store, it is assumed that the distance walked may be expected to increase proportional to the square root of the number of items ( W/N—Q) carried by the store. For a two-story building, the authors propose a cube root measurement. The third item of cost represents lost opportunity costs (ci) of other shopping opportunities foregone by consumer. It is possible that this segment of cost (and perhaps cn) could be negative for those who enjoy shopping. Thus, the three segments of costs are assumed to be: 1) cost of getting to the store, 2) actual costs of shopping, and 3) oppor- tunity costs. The formulas for the probability of satisfaction and cost are then combined to form a demand function for the consumer. The consumer will shop where his demand function for the formula is positive. f (N,D) = wp (N) = v (cdc + cn VN + Cl) In the above formula, w and v are subjective weights assigned by the consumer when evaluates each segment of the formula. It will be noted that the minimum number of items necessary to induce a customer to patronize a given site varies with distance. ”lwl;m.lwll.nlwil.1.li I I l I I I I I I I I I 51 A third consumer utility perspective is presented by Huff in a recent paper. In the author's words, the study was under- taken to: "1) identify elements that affect consumer travel making decisions, 2) investigate the connections and relations among these elements, and 3) examine the relative degree of interdependence of each of these elements."34 Using the concepts and terminology of the field of social psychology and communications, Huff attempts to build a model indicating basic interactions of the various factors that in- fluence the consumer's space preferences. The model inputs are analyzed using graph theory and matrix algebra. The advantages of Huff's formulation are its flexible framework for empirical analysis. While, as the author states. in its present form, it is not a predictive device, it might well possess the necessary elements for prediction after fur- ther empirical validation.35 34 . . Dav1d L. Huff, A Topological Model of Consumer Space Preferences, Occasional Paper No. 11, Bureau of Business Re- search, University of washington, Seattle, 1959. 5Several other presentations of similar types of analysis can be found in: Duane F. Marble, "Transport Inputs at Urban Residential Sites," The Regional Science Association Papers and Proceedings, Vol. V, 1959, pp. 253-66; and William L. Garrison, et. al.,op. cit., pp. 181-97. 52 Some Marketing Contributions Two distinct orientations appear in marketing literature regarding location and consumer movement. The first, differen- tial advantage,36 does not formally concern consumer movement but, rather, is concerned with location strategy on the premise of a "spatially given" consumer. It reflects the common bias of marketing literature in assuming a given market and concen- trating on optimum policies and strategy to develop the "given" market. In effect, it abstracts from the problem of the ulti- mate retail outlet and concentrates on the problems of the firm. Probably one of the most concise summaries of this approach is found in Chamberlin where he states: The availability of a commodity at one location rather than another being of consequence to purchasers, we may regard these goods as differentiated spatially and may apply the term "spatially monopoly" to that control over supply which is a seller's by virtue of his location. A retail trader has complete and absolute control over the supple of his "product" when this is taken to include the advantages, to buyers, of his particular location. Other things being equal those who find his place of business most convenient to their homes, their habitual shopping towns, their goings and comings from business or from any other pursuit, will trade with him in preference to accepting more or less imperfect substitutes in the form of identical goods at more distant places. . .37 6 3 The term 'differential advantage' was originally used by J. M. Clark in his development of the concept of workable compe- tition. (Studies in the Economics of Overhead Costs, [Chicago: University of Chicago Press, 1923.1). 7 3 Edward H. Chamberlin, The Theory of Monopolistic Competi- tion, 7th ed., (Cambridge: Harvard University Press, 1956), pp. 62‘63 o 53 The premise that location represents a differential ad— vantage to the firm is more fully developed within a marketing 38 context by wroe Alderson. A second distinct marketing orientation to consumer move- ment originated with "Reilly's Law of Retail Gravitation” in 39 . 4O . 1929. The term 'comparative advantage. probably best fits Reilly's and his successor's approach to consumer movement. In its original form, "Reilly's Law" stated that: Under normal conditions, two cities draw retail trade from a smaller intermediate city or town in direct pro- portion to some power of the population of these two larger cities and in an inverse proportion to some power of the distance of each of the cities from the smaller intermediate city.41 The "law" is applicable in predicting relative trading areas for shopping goods purchases. The two variables involved in the original formulation were population and distance. The 8W'roe Alderson, Marketing Behavior and Executive Action, (Homewood, Illinois: R. D. Irwin, Inc., 1957), pp. 101-29, 337-50. 39Reilly, op. cit. 40The principle of comparative advantage is, "Generally speaking, each area tends to produce those products for which it has the greatest ratio of advantages or the least ratio of disadvantages as compared with other areas." (Raleigh Barlowe, Land Resource Economics, [New YOrk: Prentice Hall, 1958], p. 224.) For purposes of this research, this definition can be reworded to read: Generally speaking, each buyer will tend to direct his patronage toward those locations which have the greatest ratio of advantage or the least ratio of disadvantage as compared with other areas in fulfilling his purchase objectives. 41Reilly, op. cit., p. 16. 54 variables have been largely replaced or reworked by other contributors to the gravitational approach. Some authors have sought to extend the range of applica- tion by substituting additional variables and rearranging the weighting procedure.43 Some of the recent contributions are discussed in greater detail in Appendix A. One methodological consideration that all of the gravita- tional formulations possess is the consideration of two points and the breaking distance of the trading area between these two points. By assessing the impact of a selected range of variables, the comparative advantage between these two points can presumably be predicted with a "reasonable" degree of certainty. Definitionswand Criteria for Delineation of the Retail TradinggArea The American Marketing Association defines a trading area as: A district whose size is usually determined by the boundaries within which it is economical in terms of volume and cost for a marketing unit or group to sell and deliver a good or service.44 42 Paul D. Converse, "New Laws of Retail Gravitation," Journal of Marketing, October, 1949, pp. 382-83. 43Harry J. Casey, Jr., "The Broadening Perspective of Marketing," American Marketing Association (Chicago, 1956), p. 82. See Appendix A for a more detailed consideration of this type ' 44 Marketing Definition: A Glossary of Marketing Terms (American Marketing Association, Chicago, 1960), p. 22. 55 The above definition is not operational in the sense that it sets no objective criteria for evaluating what are the limits of the trading area in any actual situation. It is significant to note that this definition also contains the bias of the firm in that it does not consider what is "economical" to the consumer but, rather, what is "economical" to the firm. Cohen and Applebaum offer the following review of operational criteria for the delineation of the trading area:45 A. Drawing Power 1. The trading area is "that area from which the community receives approximately 90 per cent of its total retail patronage."46 2. ”The area of influence from which a shopping center could expect to derive as much as 85 per cent of its total sales volume is defined as the trade area of the center."47 B. Per Capita Sales 1. The trading area is "that area which will provide (a general merchandise store) a minimum annual per capita sales of one dollar."48 45 . Applebaum and Cohen, op. c1t., p. 3. 46 Isodore V. Fine, Retail Trade Analysis, University of Wis- consin, Bureau of Business Research and Service, Madison, 1954, p. 11. 47Victor Bruen and Larry Smith, Shopping Towns U.S.A., (New York: Reinhold Publishing Company, 1960, p. 278.) 4 8Howard L. Green, Montgomery Ward Co., Correspondence with William Applebaum and Saul B. Cohen, April, 1959. 56 C. Time 1. "Generally speaking. . . a large majority of customers are willing to travel 12-15 minutes, and a maximum of . . 4 25 minutes to reach a regional shopping center." 9 D. Population 1. "2,000 families would spend $2,000,000 for food annually, or enough to support a large supermarket--if all families . "50 were to patronize one store. All of the above operational criteria have certain weak- nesses when applies to all retail units or groups. The weak- nesses are discussed at some length in the cited article. The authors conclude with the following statement: As a broad definition, the authors suggest that the trading area is the area from which a store gets its business with- in a given span of time. This does not exclude the reality of overlap. It also emphasizes the "area" in trading area. People must come to a store from a specific area. If other stores offer equal attractions, then the trading area of a given store will be related to the store's distance and convenience of access from the origin and destination of the potential customers."51 A further refinement of the trading area concept integrating the factor of demand intensity into the analysis and measurement of trading areas is presented in the following statement: 49Gruen and Smith, op. cit., p. 33. 50 Max S. wehrly and J. Ross McKeever, eds., The Community Builders' Handbook, Community Builders' Council of the Urban Land Institute, washington, D.C., 1954, p. 134. 51 Applebaum and Cohen, op. cit., p. 6. 57 The term primary trading area. . .refers to that portion of the trading area which provides the greatest density of customers in relation to population. This is based upon findings that are tentative and subject to further valida- tion. In metropolitan areas, generally, the primary trad- ing area for supermarkets provides 45 to 75 per cent of the customers, and a ratio of cusomers to population that is at least twice the customer population ratio of adjoin- ing portions of the trading area. The secondary trading area includes 20 to 40 per cent of the customers and the fringe trading area includes 5 to 15 per cent.52 It should be noted that the above categories, while offered as experimental or tentative guide lines, are not operational in that too great a degree of overlap exists among the categories. Relevant to the problem of operation criteria for trading area analysis, a recent study contained the following statement: A Chain which was open to consideration of a new site must weight two objectives: optimal site selection and optimal net work expansion. The two objectives would not always lead to the same decision. From one viewpoint the site would be evaluated as an independent opportunity in which the sole criteria were objective measures of the potential business for a supermarket at that spot. From the other viewpoint, the chain is attempting to complete the network of stores with which it covers the market as a whole.53 In considering the two objectives of optimal site selec- tion and optimal network expansion, the two objective measure- ments of drawing power and per capita sales provide a useful analytical framework for trading area analysis. 52 Saul B. Cohen and William Applebaum, "Evaluating Store Sites and Determining Store Rents," Economic Geography, Vol. XXXVI, No. 1, January, 1960, p. 11. 53 . . . . . . The Structure of Retail Competition in the Philadelphia Market," Wharton School of Finance and Commerce, University of Pennsylvania, December 31, 1960, p. 51 (mimeographed second draft). 58 The factor of drawing power indicates the geographical nature of market converage and, hence, is relevant to the prob- lem of optimal network expansion. On the other hand, per capita sales provide a measurement of the quality of any individual site in terms of sales penetration. The two measurements, taken together, provide a framework for decision in site loca- tion and development planning. The problem of application of the drawing power and per capita sales measurements is by no means a simple matter. Probably no two sites are exactly similar, and some judgement must be employed in effectively using these objective measure- ments. Distinct location profiles could be developed on the' basis of past experience if relevant variables could be iso- lated and integrated into the analysis of a proposed site. An example is the influence of population density or competi- tion on the drawing power and per capita sales of a store in a certain size community. It is entirely possible that, if enough observations were made, a reliable statistical rela- tionship could be formulated.54 4This type of analysis (multiple regression procedure) has been formulated in an attempt to explain the number of stores (supermarkets) that a given chain possesses in par- ticular areas on the basis of certain critical independent variables. Ibid., p. 44. 59 Contributions of the Literature to the Research In the formulation of a research problem the literature usually contributes both conceptual and methodological insights into the research problem. In this research, several different disciplines contributed both to the conceptual framework and methodOlogical basis for the study. The hypotheses are constructed using the concept of pro- duct offering at the retail store. The concept of product of- fering from the firm perspective is largely developed in market- ing and economic literature. The concept of product offering from the consumer's perspective is developed in the treatment of the principles of consumer utility. The combination of these disciplines provides the conceptual framework for the development of the concept of product offering as it is used in this research. Empirical support for the influence of product offering upon consumer travel patterns is found in the literature of economic geography and highway research and development. The methodological basis for the research are also drawn from a variety of disciplines and areas of interest. Elements of the store classification procedure are found in supermarket trade literature, real estate and appraisal literature, and economic geography publications. The survey procedure, with some modification, is adopted from store development practice 60 both in the supermarket and shopping center location and analysis literature. The measurement procedures employed in the research are developed specifically for purposes of the research and are not found in the literature. Since WOrld War II the amount of literature dealing with retail location has increased. However, the content of the lit- erature has been largely unchanged since the turn of the century. The literature presents largely guidelines and checklists for establishing a retail store. Very little empirical evidence is presented to substantiate the sweeping generalizations made by many location analyses. The guidelines and checklists are pre- sented and, if accepted, must be accepted on the basis of the expert‘ess or experience of the writer. Some original contributions are being made in the area of economic geography. The discipline of economic geography is attempting to integrate the geographical phenomena surrounding the retail site with the performance of the store in its trading area. While this appears to provide insights beyond the tradi- tional approach to retail location analysis, it still suffers from what might be termed a "clinical bias". That is, certain locations are analyzed and after the analysis conclusions are drawn on the basis of the data collected. Generalizations are made after the data is collected, rather than the generaliza- tions providing the guiding hypothesis for the research. Another 61 weakness of the "clinical approach" is that the "diagnosis" depends upon the skill of the "clinician." If the economic wastes of poor location decisions are to be avoided, retail location analysis must be elevated to the level of rigorous analytical research. The experience of previous location decisions must be quantified and used as a basis or reference point for future location decisions. 62 CHAPTER III 'RESEARCH DESIGN Introduction The purpose of this chapter is to structure a research design so that the relationship between store size and store complex as independent variables and drawing power and per capita sales as dependent variable can be investigated. As a pilot inquiry into the relationship, Chapter Three also has as its purpose a clear and complete presentation of methodology so that the research findings of the study can be either confirmed or invalidated. The survey stores were selected and the data collected in a large midwestern city. The survey stores were retail units of a chain ranking as one of the largest in the United States measured in terms of both sales and number of stores. At the request of the cooperating chain, the chain or city will not be identified. This, however, in no way adversely influences or affects the generalizations or findings. This research design is divided into three sections. The first section discusses survey store classification, establishes criteria for classification and discusses the selected survey stores in light of these criteria. The second section dis- cusses the in-store interviewing procedure for obtaining the basic data through customer interviews. The third section 63 discusses the methods employed in the analysis of the data. Included in this section are descriptions of the basic pro- cedures used to calculate population density, customer attrac- tion and per capita sales. Survey Stores Store Classification Various classification of supermarkets have been presented in supermarket trade literature. Typically the classifications found in the trade literature are used for identification pur- poses rather than functional classification. A selected number of classifications are presented in this section in order to provide an industry perspective for the functional store classifications used here. 1) New Small Shopping Center (combined selling area of less than 100,000 square feet) 2) New Large Shopping Center (combined selling area of over 100,000 square feet) 3) Neighborhood 4) Highway 5) Old Large Business Center1 Another classification is presented by Bart J. Epstein and was developed in conjunction with his study of the Quincy, l"Operational Facts About New Supermarkets Opened in 1960", Progressive Grocer, April, 1961. PP. 111-112. 64 Massachusetts food market. He presents the following nomencla- ture and definition: 1) 2) 3) 4) 5) Downtown stores: which are situated in or at the edge of the central business district, forming an integral part of the retailing complex of that area. Highway stores: which are isolated from other food stores and located so that almost all business is car- borne. Stores located in major outlaying shopping centers: where the large markets are part of a cluster of stores and serve a well defined, restricted area. Stores located in Minor outlaying shopping centers: where the large markets in a small neighborhood cluster has a relatively small trading area. Service stores: which provide a special service that the customer desires. The trading area of these stores depend, to a great extent, upon the distribution of consumers who want this special service. A third classification of retail types is presented by William Applebaum and Saul Cohen in a recent publication (Table 1). It is noted that Table 1 presents a classification 2Bart J. Epstein,"The Quincy Food Market: A Study in Market- ing GeograPhY9,unpublished Doctoral Dissertation, Clark Univer- sity, 1956, pp. 92-93. 65 of general retail location types rather than a classification referring directly to the supermarket. All of these lists for store classification presented here suffer the serious short-coming acknowledged in the Cohen and Applebaum classification. That is, they are not quantified classifications hence precision or definition is lacking. The lack of a quantified classification prevents a comparable analy- sis of distinctive characteristics of location types from being made over time and in different geographical areas. TABLE 1 CLASSIFICATION OF RETAIL LOCATION TYPES* A. "Unplanned Business Districts 1. Centrai_Busines§ District 2. CBD String,Stores, adjoin the CBD 3. Secondaiy Business District, serves portions of a cen- tral, city or suburb.. 4. SecondaryStringStores, adjoin secondary business dis- tricts. 5. Neighborhood Stores, occur in small clusters or in iso- lation. 6. Outlaying Highway Stores, occur in strings or in isola- tion. *Authors NOte: "This is not a quantified classification: there- fore, it lacks precision of definition, however, it is a classi- fication which can be readily applied. For an approach to a qualified, functional classification, see Shopping Center Report (preliminary report), Cleveland Regional Planning Commission, Cleveland, Ohio, 1958, p. 26." 66 B. “gianned Shopping Centers 1. CBD Planned Shopping Center, arise through urban re- newal. 2. Regional Planned Shopping Center, in strong competi- tion with the CBD. 3. Community Planned Shopping Center, in competition mainly with secondary business districts, or with the CBD in smaller cities. 4. Neighborhood Planned Shopping Center, frequently called neighborhood "strip." 5. Outlaying Planned Shopping Center, draws, in part, upon the passing parade of highway traffic."3 Survey-Store Classification fOr This ReSearch The two general criteria fOr store classification are: 1) Planned and unplanned retail areas 2) The types of stores surrounding the survey site. More detailed criteria are used to set up each of the individual classifications. These criteria are presented below and the sur- vey stores are evaluated in terms of these criteria. It should be noted that the store types presented do not represent all of the possible types of supermarket locations within the urban and suburban area. But rather the six types 3William Applebaum and Saul Cohen, "The Dynamics of Store Trading Areas and Market Equilibrium," The Annals of the Ameri- can Association of Geographers, vo1. 51, No. 1 (March, 1961) pp. 74-77. 4"Surrounding" is defined as 1/3 mile in the case of un- planned centers and the entire planned center in the case of shopping centers. 67 represent distinctive store types as measured by the estab— lished criteria. A. Urban Strip Store The criteria for selection of the Urban Strip Store as follows: 1) Located in an unplanned business development. 2) Located in proximity to retail stores selling con- venience type merchandise. At least ten stores of this classification with 1/3 mile of the survey store. 3) Located on major traffic artery in urban area. 4) Located in an area where population density is at least 7,500 people per square mile. B. gipap.giuster Store The criteria for selection of the gip§p_gip§ter Store are as follows: 1) Located in an unplanned business development. 2) Located in proximity (1/3 mile) to at least one large departmentalized store selling shopping goods. 3) Located on a major traffic artery near intersection with another major traffic artery. The survey store is located within one-mile of this intersection. A traffic light controls traffic at this intersection. 4) Located in an area where population density is at least 7,500 people per square mile. C. Small Town Store The criteria for selection of the Small Town Store are as follows: 1) 2) 3) 4) 68 Located in planned shopping development. Located in center with over fifty retail units. Located in center dominated by a large department store. Off street parking provided for at least 5,000 automobiles. E. Community Shopping Center The criteria for selection of the Community Shopping Center 1) 2) 3) are as follows: Located in planned shopping development. Located in center with over sixteen and less than fifty retail units. Located in center dominated by medium sized depart- ment store or large specialized variety store. F. Neighborhood Shopping Center The criteria for selection of the Neighborhood Shop- ping Center are as follows: 1) 2) 3) Located in planned shopping development. Located in center with over seven and less than sixteen retail units. Located in center dominated by a supermarket and/ or a medium sized variety store. 69 TABLE 2 UNPLANNED SURVEY STORE DATA Population 3:3: Mile Urban Strip Urban Strip-l 10/44 12/60 6,945 Urban Strip-2 6/50 12/60 9,703 Urban Strip—3 1/58 10/59 15,505 Urban Cluster Urban Cluster—l 3/52 7/60 8,829 Urban Cluster-2 9/59 9/60 15,403 Urban Cluster—3 8/59 7/60 14,455 Small Town Small Town-l 1/41 1/61 3,5002 Small Town-2 8/51 12/60 4,0002 Small Town-3 5/57 12/60 2,0002 lPopulation per square mile within a two-mile radius of the survey store. Approximate population based upon final 1960 census data. — I :Ii1l: Ili I .l l l: (I. I I I: ,l: -l. -l tlv '3 (I, I 7O .wuoum mw>usm mnu mo msflpmu wHHE 030 m GHQDHB mafia mumsqm Hem coaumasmom N .wnoum >m>u5m mo maficwmo mo camp on mummwu Umcmmo manna eam.a omh umxumsummsm m oo\ma nm\m muumucoo mcflmmosm woosnonnmflmz omn.m ooo.a umxumauwmsm ma ow\ma mm\a Alumucmo msflmmonm pOOSHonsmHmz Housmo UGWMQOSm vooauonnoflmz muoum ucwfiuummma maa.m omH.H ssflcmz as oo\oa mm\H Nunmucmo maflmmonm maficseeoo muoum quEuHmmwQ mom.m ooo.a ssfinmz ma ooxm mm\m Hummucmo maummonm muncsseoo kucmu UCHQQOSm SDHGDEEOU muoum ooo.m ucmEuummwQ mem.m um>o magma om um>o oo\m >m\e muumuamo mcflmmonm Hmcoummm muoum ooo.m pamEuummmo Aoa.a um>o mmumu om um>o oo\~H vm\m Hummucmo mcwmmonm Hmcoflmwm Hmucmo unfiamonm Hmcofimwm 0H8: m mmumswm Hm mmumam muoum monoum Umhw>usm Umcmmo Hmucwo mo make muflmsmn mcflxumm unmcwfion mo Hwnasz muma mama cowumasmom H MDm QMZZdAm m mqmfifi 71 TABLE 4 TYPES OF RETAIL UNITS IN SURVEY PLANNED SHOPPING CENTERS Community Neighborhood Type of Store Shopping Center Shopping Center 2 l 2 1 Dept. Store - Large - Medium 1 , l - Small 1 1 Variety Store - Large 1 l - Medium 1 - Small 1 Supermarket 2 l l 1 Drug Store 1 l 1 Men's Clothing 1 2 l 1 Women's Clothing 2 2 l 3 Children's Clothing 1 1 Shoes 2 l 3 Restaurant la Delicatessen 1 Bakery 1 Finance Company 1 Jewelry 1 1 Yard Goods and Draperies l 1 Paint Store Paint 1 ll 13 9 l4 a Large bakery and restaurant combined. 73 traffic intersection specified in criteria while Urban Cluster — three is located about one-third of a mile from a major inter— section. Smail Town Survey Stores - The survey stores in the small town category fulfill all of the requirements of the criteria. Regionai Shopping Center - The two survey stores selected in the Community Shopping Center category fulfill all of the criteria requirements with the exception of number two. Both Community Shopping Center - one and Community Shopping Center - two have less than the minimum number of stores required for the Community Shopping Center. Hewever, the size and nature of the dominent stores in these centers clearly place these centers in the Community Shopping Center category. Ngighborhood Shopping Cents; - All of the criteria requirements are satisfied by the two survey stores selected in the Neighborhood Shopping Center category. Customer Survey Procedures Customer Interviewing The customer interviewing portion of the study took place between October, 1959 and January, 1961. H6wever, most of the interviewing (11 out of 15) took place during the last six months of 1960. 74 Instructions to the interviewers for the actual procedure are presented in Appendix B. The number of interviews per store was set at one interview per one hundred dollars in sales per week. The weekly sales figure was calculated as the average of weekly sales for a period of four weeks prior to the inter- view date. For example, if weekly sales averaged $25,000 for a period of four weeks prior to the interview date a quota of 250 interviews would be set for the interviewer. In order to obtain 250 usable interviews (addresses) a seven per cent in- crease over intended quota should be established. This is to insure enough usable interviews considering misstated addresses or addresses that cannot be found on a map of the city. When, after deducting unusable interviews, there remain too many interviews for the quota, the excess interviews are discarded using a table of random numbers. The customer interviewing procedure adopted for the collec- tion of data is a frequently used technique for location re- search in the supermarket industry. Prior testing established the validity of the technique for accurately determining the trading area of the supermarket.5 5 Bart J. Epstein, “Evaluation of an Established Planned Shopping Center" op. cit., pp. 12-21 and William Applebaum and Richard F. Spears, op. cit. 1. , __ _ . . 75 Map Spotting The map spotting procedure is a mechanical operation and a procedural presentation for this step is found in Appendisz- It consists of plotting the customer's home addresses on a map of the survey area. There are several possible sources of error in an analysis using comparative map data. First, if the quadrant type of analysis is used as it is used in this research, caution must be exercised so that true north is accurately established and consistantly used in all comparative analysis. Another possible source of error is found in using maps of different mileage scale. In the analysis of the data collected, three different map scales were used. Prior to performing any analysis or drawing any circles, the map scale was checked in order to elimi- nate errors in analysis or time consuming recalculations of the data. Measurement Procedures Drawing Power Measurement Drawing power was defined as the mean average distance traveled by a fixed percentage of the customers. The question was posed as to what percentage to use in the calculation of drawing power. Experience indicates that it is reasonable to assume that some of the shoppers interviewed are what might be termed "accidental shoppers" in that they are not regular patrons 76 of the store.6 Ninety per cent of the closest customers to the site are used to calculate the basic drawing power measurement. In order to show a relative measure of concentration, the main average distances were also calculated at the seventy per cent customer level. The procedure for calculation was as follows: Using the customer spotting map as a base, a transparent overlay was placed over the survey site (Appendix C-2). The overlay had a series of concentric circles drawn upon it which are one-eighth of a mile apart scaled to the map. It also had a vertical and horizontal axis dividing the overlay into four equal segments or quadrants. The overlay was placed on the map so that the intersection of the horizontal and vertical axis was placed directly over the survey store. The vertical axis was aligned so that it pointed directly north and south. The customer spottings were then recorded by distance interval and quadrant using a worksheet designed for the pur- pose (Appendix C-3). The quadrants were totaled in the last column resulting in a customer count for each one-eighth of a mile distance interval. The main average distance was then calculated using the following formula: 6"Profile of Supermarket Customers" Part 8, Super Value Study, August, 1958, p. 8112-13. 77 - FM M1 N Where - M the mean distance F = number of customers falling within distance interval .2 I ' mid-point of distance interval (e.g. .0675 for 1/8 of) (a mile interval: .1875 for) (1/4 mile interval, etc.) N Number of customers The above calculation was made for all survey stores using data for both the closest seventy and ninety per cent of the customers. PepiCapita Saigg Meagurement Per capita sales within an area is an important measure- ment device for analyzing sales penetration in any given market segment. It introduces into analysis the factors of population and population density not considered in the drawing power‘ measurement. In order to calculate per capita sales in an area, three types of data must be available. First, the segment of the market must be clearly delineated for analysis so that both population totals and sales may be gathered using a comparable reference point. Secondly, accurate population data must be available to serve as the basis for calculation of per capita sales. Third, sales figures must be available for the market segment in question. 78 The entire analysis of per capita sales must be accom- plished over a relatively short time span so that sales and population refer to the same point in time. Shifting popula- tion concentrations and sales patterns cause distortions in final calculation if too great a time span exists between the population estimate and the sales estimate. The customer interviewing procedure was based upon a quota of one interview per one hundred dollars in sales per week. When the customers addresses are spotted on a map, both a distribution of customers and a distribution of sales results. Thus, if in a one-half square mile area ten customer spottings are made, it can also be said that there is $1,000 per week sales in that area. If the population in the area is 1,000 people, then it can be further stated that per capita sales in the area is one dollar per week or fifty-two dollars annually. The requirements of distinct market parameters was re- solved so that data gathered to provide drawing power infor- mation could also be utilized in the per capita sales portion of the analysis. That is, market segments were laid out accord- ing to concentric circles drawn from the survey site. The first circle was drawn one-half mile from the store; the second, one and one-quarter miles from the store; and the third, at a dis- tance of two miles from the store. The decision on which dis- tance measurements to use was made after a visual analysis of the 80 spotting maps and a preliminary analysis of the drawing power results. It is believed that these distance increments repre- sent distinct market areas for the survey stores as measured by drawing power and per capita sales. The quadrants used in the per capita sales measurement were established to coincide with those quadrants used in the drawing power analysis. A graphic presentation of this approach can be found in Appendix C-4. After the two steps outlined about were completed, the task of accurate population estimation remained. Fortunately, 1960 census data by census tract and enumeration district were available for the survey area. The availability of the data also eliminated the problem of corresponding sales and population estimates. As indicated in Tables 2 and 3, all of the surveys were made within a period of fifteen months. Thirteen of the fifteen survey stores were surveyed in 1960, the year for which census data was most applicable. Appendix C-5 contains a copy of the worksheet used for population estimates. The general procedure established called for estimation by census tract within the central city and es- timation by enumeration district outside the central city. The Census Tract Map was drawn on to a street map of the area and the circles and quadrants were then also drawn on the city street map. 81 The next step in the procedure was to list population and housing counts by enumeration district or census tract. After the listing was completed, the percentage of each census tract and enumeration district following within a distance in- terval and quadrant were estimated. The total number of in- habitants in the census tract or enumeration district were then multiplied by the percentage estimate. The results were summed and population by distance interval and quadrant was then available for further analysis. The following formula was used to calculate per capita sales: Where - Sa = per capita sales in Area A. Ca = customer spotting in Area A. P = population of Area A. For example, if a population estimate of 5,500 were made for a survey store at the distance interval of 1/2 mile and there were 65 survey store customers spotted within the 1/2 mile distance interval, the calculation of per capita sales would be as follows: : 65 x 100 = 6,500 = a 5,500 5,500 ”1'08 A total per capita sales of $1.08 per person within the 1/2 mile distance interval would result for the hypothetical survey store. 82 CHAPTER IV PRESENTATION OF FINDINGS Introductigp The purpose of this chapter is to present the empirical findings of the study. The chapter is divided into four gen- eral parts designed to present the findings in a meaningful and useful context. The first part presents the purposes, objectives and basic framework for the presentation of findings. The second part, entitled "Drawing Power Measurements," presents findings rela- rive to the dependent variable of drawing power and its rela- tionship to store complex and store size. The third section presents data relative to the dependent variable of per capita sales and its relationship tolstore complex and store size. The fourth section presents research findings which are not directly relevant to the research problem, but were discovered in the analysis of the data and believed to be significant. In order to insure clarity and continuity of presentation, most of the raw data have been appended and can be found in Appendices E-G. The procedure for the analysis of variance can be found in Appendix D. UtiliZing the quota sample set up in Chapter III, a total of 5,300 customer interviews were taken. The classification of survey store customer interviews is presented in Table 5. 83 At the close of each section the statistical significance of the results are presented. The data are first analyzed using Analysis of Variance and the significance of inter-group compari- sons are established using Multiple Range Tests. Where appro- priate, Correlation procedures were used to analyze the data. The basic formula and identities for these procedures are pre- sented in Appendix D. The results of the statistical analyses are presented in terms of levels and significance» The conventions adopted for presentation are the standard conventions for reporting the re- sults of statistical procedures. That is, where results are reported to be significant at the 95 per cent level, the nota- tion indicated that there are five chances out of 100 that the data are the result of chance occurances. In the case of a 90 per cent level of significance this notation indicated that there are but ten chances in 100 that the data are due to chance occurrences. Probability tables with the appropriate degrees of freedom were used to establish the levels of significance for the data. The 95 per cent level of significance is the minimum level of significance accepted as reliable for the statistical section of the research findings. The 95 per cent level is tradition— ally the acceptable minimum criterian for research significance. TABLE 5 NUMBER OF INTERVIEWS AND DATE OF SURVEY BY SURVEY DATE 84 Store T e Number of Date of yp Interviews Survey Urban Strip-l 165 12/60 Urban Strip—2 242 12/60 Urban Strip-3 479 10/59 Total 886 - Urban Cluster-l 368 7/60 Urban Cluster-2 487 9/60 Urban Cluster-3 564 7/60 Total 1419 - Small Town-l 106 1/61 Small Town-2 233 12/60 Small Town-3 188 12/60 Total 527 - Neighborhood Shopping Center-l 211 12/60 Neighborhood Shopping Center-2 345 12/60 Total 556 - Community Shopping Center-1 262 10/60 Community Shopping Center-2 526 3/60 Total 788 - Regional Shopping Center-l 576 12/60 Regional Shopping Center—2 548 3/60 Total 1124 - TOTAL 5300 - 85 DrawingoPower Measurement Intpoductipp_ The definition formulated in the previous chapter for drawing power was the mean average distance traveled to the survey store by seventy and ninety per cent of the store's customers. Prior to a consideration of drawing power using the above definition, a presentation of customer profiles at the 100 per cent level of patronage provide a basic framework for further analysis. In Table 6, profiles of customers by dis- tance interval are presented by store type. Several dis- tinctive patterns can be immediately seen in Table 6. The amount of customers drawn from under one-half mile declines steadily ranging from a high of 51.4 per cent for the Urban Strip type to a low of 2.4 per cent for the Regional Shopping Center types. The data presented by column is continous with the exception of the Small Town type where there is some de- viation by distance interval. In the intermediate distance interval ranges (1-1/2 miles to 2 1/2 miles) the planned centers show relatively strong customer draw while the Urban Strip and Urban Cluster types taper off sharply beyond the one mile dis- tance interval. The only store type drawing customers in sig- nificant amounts over 3 1/2 miles are the Regional Shopping Center and Small Town types. TABLE 6 CUSTOMERS BY DISTANCE INTERVAL BY STORE TYPE 86 Unplanned Planned Distance Neigh- Com— Region- Interval Urban Urban Small borhood munity a1 (Miles) Strip Cluster Town Shopping Shopping’ Shopping Center Center Center 1/2 51.4% 46.7% 34.6%» 20.1% 16.8% 2.4% 1 30.7 26.8 16.3 29.4 27.3 9.0 1-1/2 6.1 12.6 9.1 19.0 17.6 14.7 2 2.9 5.8 5.8 14.5 12.4 11.2 2-1/2 2.6 2.9 7.7 5.6 7.7 10.8 3 1.5 1.3 4.4 3.1 5.3 9.9 3-1/2 1.3 0.8 5.0 1.9 4.3 8.4 Over 3_1/2 3.5 3.1 17.1 6.4 8.6 33.6 Total 100.0 100.0 100.0 100.0 100.0 100.0 87 Drawing Power and Store Complex The empirical findings for the relationship between drawing power and store complex are presented in a series of three tables. All tables are presented at the seventy and ninety per cent levels of customer drawing power in order that the relative concentration and dispersion of customers can be analyzed. In Table 7 a summaryqtable of drawing power findings by store type is presented. A range of from .38 miles to 1.95 miles appears in a continous series ranging from the urban Strip type to the Regional Shopping Center type at the seventy per cent level of customer drawing power, at the ninety per cent level of customer drawing power the range shifts upward to .52 miles for Urban Strip type to 2.53 miles for the Regional Shopping Center type. The data are also presented in a continous series with the exception of the Small Town type at the ninety per cent customer drawing power level. In Table 8 a presentation of drawing power at the seventy and ninety per cent levels by survey store is presented. The purpose of Table 8 is to demonstrate the degree of consistency within each classification of store types. A comparison of mean1 and median values at the seventy and ninety per cent level is presented in Table 9. The comparison The average mean values are used in the calculation of drawing power at the seventy and ninety per cent levels. 88 of the mean and median values provides an approximate measure of the symmetry of distribution for the data. Where the mean, median and mode are identical, a perfectly symmetrical distri- bution is indicated. Where the median exceeds the mean gener- ally the skewness will be positive. Where the mean exceeds the median generally the skewness will be negative.2 At the seventy per cent level the distribution for data in Table 9 indicate a relatively symmetrical distribution with the exception of the Small Town type stores. However, when the analysis is extended to the ninety per cent level, the data show a definite pattern of negative skewness. The negative skewness is particularly evident in the case of the Small Town type stores. Drawing Power and Store Sigg_ The research findings for the relationship between drawing power and store site are presented in a series of two tables. Several problems are presented in an analysis of store size and its relationship to drawing power. The first problem concerns the interrelationship between store size and store type. It is a reasonable assumption that a store located in a Regional Shopping Center is larger than a store located in a small town. 2 . . . . The Pearsonian coeff1c1ent of skewness is calculated by: 3 (mean - median) SK = . . standard deViation 89 If this assumption holds, store size can be influenced by store type. If there exists a relationship between store size and store type, then the analysis of these two factors can best be made with the store type classification rather than over the entire range of store types. It is the purpose of Tablell to provide some insights into the relationship of store type and store size. A second problem concerns the time span over which stores are constructed. The average size of stores has been steadily increasing. The period during which a store was built deter- mines to a large extent, its' size. Along with its' size,'the period in which it is built also determines to some extent the nature of its' competitive environment. Stores can be expected to decline in productivity as they approach obsolescense. Thus, to remove the influence of these factors an analysis of stores all built in the same year is required. This problem is pre- sented at greater length under "other findings" below. In Table 10 the stores are ranked by store size in the pre- sentation of drawing power at the seventy and ninety per cent levels. In the table no systematic pattern for the relationship of drawing power to store size emerges. However, when the stores are arrayed by store type (Table 11), there is a relationship within the store type classification for certain of the store types. 90 TABLE 7 DRAWING POWER AT 70 AND 90 PERCENT CUSTOMER LEVELS BY STORE TYPE Drawing Power Store Type 70 Percent 90 Percent Urban Strip .38 Miles .52 Miles Urban Cluster .43 .62 Small Town .79 1.38 Neighborhood Shopping Center .79 1.05 Community Shopping Center .87 1.22 Regional Shopping Center 1.95 2.53 TABLE 8 91 DRAWING POWER AT 70 AND 90 PERCENT CUSTOMER LEVELS BY SURVEY STORE Drawing Power Survey Store 70 Percent 90 Percent Urban Strip-l .32 mi. .39 mi. Urban Strip—2 .37 .57 Urban Strip-3 .41 .55 Urban Cluster-1 .43 .62 Urban Cluster-2 .49 .67 Urban Cluster-3 .39 .56 Small Town—1 .54 1.22 Small Town—2 .84 1.44 Small Town-3 .87 1.41 Neighborhood Shopping Center-l .88 1.24 Neighborhood Shopping Center-2 .74 .97 Community Shopping Center-1 .98 1.34 Community Shopping Center-2 .82 1.17 Regional Shopping Center-l 2.42 3.15 Regional Shopping Center-2 1.57 1.86 9 2 TABLE 9 COMPARISON OF MEAN AND MEDIAN VALUES OF DRAWING POWER AT 70 AND 90 PERCENT CUSTOMER ‘ LEVELS BY SURVEY STORE —— L 1 Drawing Power (Miles) Survey Store 70 Percent 90 Percent Mean Median Mean Median Urban Strip-1 .32 .32. .39 .39 Urban Strip-2 .37 .38 .57 .44 Urban Strip—3 .41 .41 .55 .49 Urban Cluster-l .43 .41 .62 .50 Urban Cluster-2 .49 .45 .67 .55 Urban Cluster-3 .39 .37 .56 .44 Small Town-l .54 .41 1.22 .50 Small Town-2 .84 .65 1.44 .88 Small Town-3 .87 .52 1.41 .78 Neighborhood Shopping Center-l .88 .87 1.24 1.03 Neighborhood Shopping Center-2 .74 .70' .97 .88 Community Shopping Center-1 .98 .97 1.34 1.21 Community Shopping Center-2 .82 .79 1.17 .95 Regional Shopping Center—l 2.42 2.43 3.15 2.94 Regional Shopping Center-2 1.57 1.51 1.86 1.73 93 TABLE 10 DRAWING POWER AT 70 AND 90 PERCENT CUSTOMER LEVELS BY RANK ORDER OF STORE SIZE Size Drawing Power Store Type of . Store1 70% (miles)90% Small Town-1 4,000 .54 1.22 Urban Strip-l 4,698 .32 .39 Small Town-3 5,455 .87 1.41 Small Town-2 5,673 .84 1.44 Urban Strip-2 9,030 .37 .57 Neighborhood Shopping Center-2 10,004 .74 .97 Neighborhood Shopping Center-1 10,094 .88 1.24 Urban Cluster-l 10,320 .43 .62 Urban Strip-3 10,464 .41 .55 Community Shopping Center-1 10,629 .98 1.34 Urban Cluster-2 10,780 .49 .67 Urban Cluster-3 11,368 .39 .56 Regional Shopping Center—2 11,582 1.57 - 1.86 Community Shopping Center-2 13,468 .82 1.17 Regional Shopping Center~1 16,800 2.42 3.15 Store size = square feet of selling area within the store. DRAWING POWER AT 70 AND 90 PERCENT CUSTOMER TABLE 11 LEVELS BY STORE TYPE AND STORE SIZE 94 Size Drawing Power2 Store Type of Store1 70% 90% Urban Strip-1 4,698 .32 .39 Urban Strip-2 9,030 .37 .57 Urban Strip-3 10,464 .41 .55 Urban Cluster-1 10,320 .43 .62 Urban Cluster-2 10,780 .49 .67 Urban Cluster-3 11,368 .39 .56 Small Town-l 4,000 .54 1.22 Small Town-2 5,673 .84 1.44 Small Town-3 5,455 .87 1.41 Neighborhood Shopping Center-1 10,004 .88 1.24 Neighborhood Shopping Center-2 10,094 .74 .97 Community Shopping Center-1 13,468 .98 1.34 Community Shopping Center-2 10,629 .82 1.17 Regional Shopping Center-1 16,800 2.42 3.15 Regional Shopping Center-2 11,582 1.57 1.86 Store size = square feet of selling area within the store. 21n miles. 95 Statistical Significance of Drawing Power Measurement 1. Relationship between drawing power (seventy per cent customer level) and store complex. The relationship between drawing power and store complex was analyzed both at the seventy and ninety per cent customer levels using Analysis of Variances. The results of this analysis using the seventy per cent drawing power measurement are presented below. the 99 per cent level. Figure 7 ANALYSIS OF VARIANCE - STORE COMPLEX AND DRAWING POWER AT 70 PER CENT CUSTOMER DRAWING POWER LEVEL (Column = Store Complex) The results were highly significant at Source of Variation DF SS MSQ, F Total 14 4.2943 Between Columns 5 3.8347 .7669 15.0166* 'Within 9 .4596 .0511 (*Reference Point - 95% 3.48;99% 6.06) . 4 . . The Multiple Range Tables were then used to examine the dif- ferences between store complex and drawing power. On 3See Appendix D-l for formula and procedure for of Variance (one-way classification). 4 See Appendix D-3 for formula and procedure for Range Test. the basis of Analysis Multiple 96 the Analysis of Variance, the following differences were es- tablished at the 95 per cent level. A. The Regional Shopping Center type is significantly different in drawing power from all other store types. B. The urban Strip Store type is significantly dif- ferent in drawing power from the Community and Regional Shopping Center types. 2. Relationship between drawing power (90% customer level) and store complex. Using Analysis of Variance to analyze this relationship, the following results were obtained: Figure 8 ANALYSIS OF VARIANCE - STORE COMPLEX AND DRAWING POWER AT 90 PER CENT CUSTOMER DRAWING POWER LEVEL (Column = Store Complex) Source of , Variatipn DF SS QMsg 1 F Total 14 6.8706 Between Columns 5 5.9336 1.1867 11.3996* Within 9 .9370 .1041 (*Reference Point - 95%13.48;99%.6.06) These results are significant at the 99 per cent level. Using the Multiple Range Tables to establish the differences within the categories the following results were obtained at the 95 per cent level of significance. 97 A. The Urban Strip type is significantly different in drawing power from the Small Town type. ' B. The Urban Strip type is significantly different in drawing power from the Community Shopping Center type. C. The Urban Strip type is significantly different in drawing power from the Regional Shopping Center type. D. The urban Cluster type is significantly different in drawing power from the Small Town type. E. The Urban Cluster type is significantly different in drawing power from the Community Shopping Center type. F. The Urban Cluster type is significantly different in drawing power from the Regional Shopping Center type. 3. Relationship between drawing power (70%.customer drawing power level) and store size. The store size and drawing power variables were correlated for the 70 per cent level of customer drawing power using data from all fifteen survey stores. A coefficient of correlation of +.575 was obtained which is significant at the 95 per cent level of significance. 4. Relationship between drawing power (90%.custOmer drawing power level) and store size. The store size and drawing power variable were correlated for the 90 per cent level of customer drawing power using data from all fifteen survey stores. The resulting coefficient of correlation was +.444 which is significant at the 90 per cent level of significance, but not at the 95 per cent level. 5The significance of the coefficient of correlation is es- tablished by using Table 11 "Percentile Values of r for n Degrees of Freedom When p equals 0." (Helen M. walker and Joseph Lev., Statigtical Inference. New YOrk: Henry HOlt & Company, 1953, p.470). 98 Per Capita Saies Measurements Introduction Per capita sales was defined as the dollar amount of sales per person per week within a given geographic area. This measure- ment provides a yardstick of productivity for a retail store. Presumably, this measurement would also be responsive to the amount and location of competition. The problem of the influence of competition is discussed at greater length under "Other Find- ings" below. The population estimates necessary for the per capita sales figure were made at three distance intervals: 1/2 mile, 1 1/4 miles and 2 miles. These distance intervals were chosen because they are most meaningful considering the clustering of the drawing power data. The statistical significance of the three distance intervals used for per capita sales measurements were established using a two way classification of Analyses of vari- ance in conjunction with the Multiple Range Tests. The results are presented below in "Statistical Significance of the Per Capita Sales Measurement." The Analysis of Variance, Multiple Range Tests and Corre- lation procedures were also used to establish the statistical significance of the relationship between the independent vari- able of per capita sales. Since Census TraCt and Enumeration District Data are not available for the outlaying areas of the Small Town type stores, they are not included in any of the 99 statistical measures of relationship beyond the 1/2 mile dis- tance interval. The population estimations for these outlaying areas were made using the 1960 census (Township) data and the most recent maps of the area. The resulting population es- timates are probably as accurate as the remaining data, how- ever, since different estimation procedures were used, these three survey stores were excluded from statistical analysis of the data. In order to provide a framework for per capita sales figures, Table 12 indicates the percentage of customers coming from each distance interval up to two miles. Except for the Small Town and Regional Shopping Center types, approximately 75 per cent or more of the customers came from within two miles of the store. A special caution must be taken in inter- pretation of per capita sales for the Regional Shopping Center type where over 65 per cent of the customers are located farther than two miles from the store. As a result, the per capita sales data are only applicable to 35 per cent of the customers of the Regional Shopping Center type. Pg; Capita Saiggoand Store Compigg The research findings establishing the relationship be- tween per capita sales and store complex are presented in two tables (Tables 13-14). Both these tables are constructed so as to show interval and cumulative per capita sales by distance in- terval. 100 In Table 13 a summary table of findings by store type is presented. The Small Town type stands out prominently due to its high per capita sales up to and including the two mile distance interval. Another interesting pattern is also shown by the per capita sales of the Urban Cluster and Urban Strip in the 1-1/4 and 2 mile distance intervals. Per capita sales decline rapidly for the Urban Strip and Urban Cluster typed contrasted to the Neighborhood Shopping Center and Community Shopping Center types in the larger distance intervals. Some overlap between the Neighborhood Shopping Center and Community Shopping Center types is indicated in Table 14. Hew- ever, with this exception the data shows a systematic distribu- tion with no overlap in any of the other categories at any distance interval. Per Capita Saies and Store Size The research findings relating to the relationship be- tween per capita sales and store size are presented in Tables 15-16. In Table 15 the data are grouped by store size in three categories. There is no systematic relationship between store size and per capita sales indicated in Table 15 with almost complete overlap of data occuring at every distance interval. The same lack of pattern is shown in Table 16 where the data are arranged by store type and presented for each survey 101 TABLE 12 PERCENTAGE OF CUSTOMERS BY STORE TYPE AT 1/2, 1-1/4 AND 2 MILE DISTANCE INTERVALS Store Type Distance Interval 1/2 Mile 1-1/4 Mile 2 Mile Interval Cumulative Cumulative Urban Strip Urban Cluster Small Town Neighborhood Shopping Center Community Shopping Center Regional Shopping Center 52.0%: 85.8% 91.5% 46.7 80.3 91.9 34.6 56.1 65.8 20.1 58.2 83.0 16.8 54.9 74.0 2.4 19.8 37.3 102 Hm. ma. mm. pm. me. uwucmo mCHdOOCm Hmcoamwm «4. ON. on. me. mm.H kucmo mcflomonm maficsasoo we. mm. am. an. oa.a umucmo mnemmosm coonuonnmflmz mm.m vo.H om.m em.~ o¢.~ c309 HHmSm SN. mo. Om. mm. mm.a umumsao amen: om. w «0. m as. 8 am. 8 mm.aw mfluum zany: :ww. 1mm. :me h“. :53: we“: ~\H man: m man: ¢\Hna muoum >m>usm moamm MDHQMU mom .unlllllnln, .hunInnunnununnnuunuunnnnununnnnnnnnunnnnuuuunnununununnunnnnunuunnnuuu mq¢>mmezH mozmemHo mqu m ozm «\Hua ma mnm<8 .N\H Bfi mm»? mmOBm Wm mmflfiu Hm> m>wu am> IOHDEDO IHODGH IMHDEDO InmuCH Hm>HODCH OH OHS: m mafia ¢\HIH H.z N\H muoum >0>usm moamm muammu Hmm mq<>mmezH mozmom mm mmn O>..n# HM> IMHSEDU IHTDCH IMHDEDU IHTUGH Hm>HmucH muoum OHS: ~\H mo muoum >m>usm OHS: m OHM: «\Hua mNHm mwamm muflmmo Mom MNHm mmOBm m0 mmflmo M2¢m Mm .mQ<>mm92H mUZ0u Hm> w>00 . Hm> HM>HTHGH IMHDESU IHTHGH ImHoESU, IHOHGH m a \ muoum OHH N m H \ I H.2 N H MO OHOum mm>u3m .2 0.2 0 0 0 0000 mmamm 000000 000 MNHm HMOBW QZ< mmoem Mm>mDm Mm .mfl¢>KMBZH MUZNBmHQ HQHZ N .v\HIH N\H Ed mmgdm dfiHmdu mmm 0H mamfiB 106 store. There is little relationship within the store type classification between store size and per capita sales. Statisticaigsignificance of Per Capita4§ales Mea§Urement 1. Statistical significance of distance intervals of 1/2,- 1 1/4 and 2 miles. In order to establish the relavence of the distance in- tervals used, the Analysis of Variance and Multiple Range' Tests were applied against the per capita sales data using a two Way classification problem procedure.6 Figure 9 ANALYSIS OF VARIANCE - SIGNIFICANCE OF PER CAPITA SALES DISTANCE INTERVALS (Row = Distance Intervals) (Column = Store Complex) Source of E A .Variation DF S§_g Mpg»; F Total 14 6.1725 Row 2 4.6504 2.3252 21.1960* Columns 4 .6447 .1612 l.4695** Error 8 .8774 .1097 (*Reference Point: 95%-4.46;99% 8.65) (**Reference Point: 95%-3.84;99% 7.01) The Analysis of Variance6 procedure established the sig- nificance of the distance intervals choosen for per capita sales See Appendix D-2 for formula and procedure for AnalySis of Variance (Two-way Classification) 107 analysis at the 99 per cent level of significance. The Multiple Range Test was then applied to the result and the following con- clusions were established: A. The 1/2 mile distance interval is significantly different (95%-level) from the 1-1/4 and 2 mile distance interval in per capita sales. B. The 1-1/4 mile distance interval is significantly different (95% level) from the 2 mile distance interval in per capita sales. 2. Relationship between per capita sales at the 1/2 mile distance interval and store complex. Using the Analysis of Variance and the Multiple Range Tests, the following results were obtained: Figure 10 ANALYSIS OF VARIANCE - STORE COMPLEX AND PER CAPITA SALES AT 1/2 MILE DISTANCE INTERVAL (Column - Store Complex) Source of Variation DF SS qu F Total 14 10.9294 Between Columns 5 6.6928 1.3386 2.844* Within 9 4.2366 .4707 (*Reference loint: 90% 2.61; 95% 3.48;99% 6.06) The results were significant at the 90%-level. The,Multip1e Range Test established the following relationship at the 95 per cent level. A. The Regional Shopping Center type is significantly different than the Small Town type in per capita sales at the 1/2 mile distance interval. 108 3. Relationship between per capita sales at the 1-1/4 mile distance interval and store complex. The Analysis of Variance disclosed the following sig— nificance at the 1-1/4 mile distance interval. Figure 11 ANALYSIS OF VARIANCE - STORE COMPLEX AND PER CAPITA SALES AT 1-1/4 MILE DISTANCE INTERVAL (Column = Store Complex) Source of Variation DF SS qu, F Total 11 .5508 Between Columns 4 .4642 .1161 9.3629* Within 7 .0866 .0124 (*Reference Point: 95% 4.12; 99% 7.85) The resultant F is significant at the 99 per cent level of sig- nificance. Th significant (9 interval: e Multiple Range Tests established the following 5%-1evel) differences at the 1-1/4 mile distance The Urban Strip type is significantly different in per capita sales from the Community Shopping Center type at the 1-1/4 mile distance interval. The Urban Strip type is significantly different in per capita sales from the Neighborhood Shopping Center type at the 1-1/4 mile distance interval. The Urban Cluster type is significantly different in per capita sales from the Community Shopping Center type at the 1-1/4 mile distance interval. The Urban Cluster type is significantly different in per capita sales from the Neighborhood Shopping Center type at the 1-1/4 mile distance interval. 109 E. The Regional Shopping Center is significantly different in per capita sales from the Community Shopping Center type at the 1-1/4 mile distance interval. F. The Regional Shopping Center is significantly different in per capita sales from the Neighborhood Shopping Center type at the 1-1/4 mile distance interval. 4. Relationship between per capita sales at the two mile distance interval and store complex. The Analysis of Variance disclosed the following significance at the two mile distance interval. Figure 12 ANALYSIS OF VARIANCE - STORE COMPLEX AND PER CAPITA SALES AT 2 MILE DISTANCE INTERVAL (Column = Store Complex) Source of Variatign DF SS Mpg, F Total 11 .2115 Between Columns 4 .1571 .0393 5.0385* Within 7 .0544 .0078 A *Reference Point: 95%-4.12;99%>7.85) The resulting F is significant at the 95 per cent level of significance. The Multiple Range Tests indicated the following significant (95% level) variations among the store types: A. The urban Strip type is significantly different in per capita sales from the Community Shopping Center type at the 2 mile distance interval. B. The urban Cluster type is significantly different in per capita sales from the Community Shopping Center type at the 2 mile distance interval. 110 5. Relationship between per capita sales (1/2 mile distance interval) and store size. A test for correlation was made between the store size and per capita sales variables at the 1/2 mile distance interval using data from all fifteen survey stores. The procedure re- sulted in a Coefficient of Correlation of-0540 which is signi- ficant at the 95 per cent level of significance. 6. Relationship between per capita sales (1-1/4 mile dis- tance interval) and store size. A test for correlation was made between the store size and per capita sales variables using data from twelve stores.7 A Coefficient of Correlation of +.192 was obtained which is not significant at the 90 or 95 per cent level of significance. 7. Relationship between per capita sales (2 mile distance interval) and store size. A test for correlation was made between the store size and per capita sales variables using data from twelve survey stores. A Coefficient of Correlation of +.l66 was Obtained which is not significant at the 90 or 95 per cent level of significance. Other Findings Intioduction This selection offers comments and insights into two general types of propositions. In the literature of supermarket location 7Excluding Small Town type stores (3) 111 analysis several generalizations are found concerning the rela- tionship between density and drawing power. One such general- ization states: "The denser the population, the larger the size of the trading area, but the greater the per cent of sales that come from Close-in. Dense population attract more and larger supermarkets to one focal point, and the effect of a number of supermarkets located side by side is a bigger trading area for each store.8 Data are presented that provide insights into the validity of this type of proposition. A second type of proposition that is investigated in this section is that prompted by discovery of systematic patterns of data. The proposition is not directly relavent to the research problem but it is a significant periphreal area in the problems of store location. The Influence of Population Density on Drawing Power and Per Capita Sales In Table 17 the relationship between population density, per capita sales and drawing power is presented. The data are arrayed according to population density at the 1-1/4 mile distance interval are divided into density categories. When the data are arrayed in this manner, there is a definite connection between population density and per capita sales. With the exception 8 . . . William Applebaum and Saul B. Cohen, "Store Location Stra- tegy in a Changing Market" Proceedings of the 1961 Midyear Con- ference (Chicago: Supermarket Institute, 19611 p. 8 (reprint). 112 of the Regional Shopping Centers an inverse relationship occurs when population density is compared to drawing power at the 70 and 90 per cent level. However, the relationship must be interpreted with the awareness that there is also a relationship between store type and population density. In Table 18 the relationship is demon- strated by grouping the data in their respective store type classifications. IanHgngg of Date of Qpening Upon Diawing,Powep and Per Capita Sales In Table 19 the fifteen survey stores are ranked by date of opening. There appears to be little relationship between,‘ date of opening and per capita sales or drawing power in this table. HOwever, when the data are ClaSSified by store type (Table 20), there is a relationship within certain of the store type groupings. Influence of Competition Upon Drawing Power ‘ and PeioCapita Saigg In Table 21 the number of competitive supermarkets is pre- sented by distance interval. The exact spatial relationship of these supermarkets to the survey store is presented by degree and distance in Appendix G. The influence of competition at the 1/2 mile distance interval is shown in Table 22. 113 Statistical Significance of Findings Rank Order Correlations were made for the following rela- tionships: Beiationship 35y. j..Population Density and Drawing Power (70%) - .482 Population Density and Drawing Power (90%) - 579 Population Density and Per Capita Sales 1/2 mile distance interval - .271 1-1/4 mile distance interval - .512 1-1/4 mile cumulative - .406 2 mile distance interval — .610 2 mile cumulative - .400 114 .Hm>00020 0020000U 000E 0\010 pm 00000 000000 000 0>00005550 .00000 mm>usm mo mafipmu 0005 0\HIH £00003 0008 mumswm 00m 0000005000” 00. 00. 00. . 0n0.00 0-00000 00000 n0. 00. 00. 000.00 0-0000000 00000 00. 00. 00. 000.00 0-0000000 00000 00. 00. 00. 0n~.00 0-0000000 00000 00.0 n0.0 00. 000.0 0-000000 00000000 00000000 n0. n0. 00. 000.0 0-00000 00000 00. N0. 00. 000.0 0-00000 00000 00.0 00.0 .00. 000.0 0-000000 00000000 00000000 n0.0 00. 00. n00.n 0-000000 00000000 000000000 n0. 0n. 00. n00.0 0-000000 00000000 000000000002 00.0 00. 00. 000.0 0-000000 00000000 000000000 00.0 00. 00. n00.0 0-000000 00000000 000000000002 00.0 00. 00.0 000.0 0-0300 00000 00.0 00. 0n.0 000 0-0300 00000 00.0 n0. 00.00 0n0 0-0300 00000 Rom 30005 $00 mmamm .0000ch 00300 000300Q NmuHQmo 000 02000009000 00000 hm>nsm EMOBm Vm>mDm Mm MHZOm 02Hkflm0 QZ¢ mMQ¢m ¢BHm¢U mmm .NBHmeQ ZOHBNADmOm NH mqm<8 115 .00>H00c0 00000006 0>0000sEoo 000E v\0u0 00 N .00000 >0>000 00 000600 0008 ¢\0I0 000003 0008 000000 000 00000050000 mm.m mm.0 mm. m00.0 000000 mc0mmonm 00000000 00.0 00. 00. 000.0 000000 00000000 000000000 00.0 00. 00. 000.0 000000 00000000 000000000002 mm.0 on. om.~ 000 0309 0008m 00. 00. 00. 000.00 0000000 00000 00. . 00. 00. 0 000.00 00000 00000 000 000002V 000 000mm m00mc0a 00%9 0HO0m H0300 0:0300Q ~000mmu H00 00000000000 mmMB MMOBm Mm mmzom UZHZEmQ QZfi mm0nsm may mo msficmu mafia oBu m GHAMHB mmamm muflmmo Hmma no. mv. mm. mm\m muumumsao away: mm. mm. mm. mm\m muumumsao swans em.H mm. mm. mm\H Huumucmo mcflmmosm nuancesou 5H.H mm. om. mm\m maumucwo mcflmmosm muflcsesou mm. ‘H¢. mm. mm\a mumfiuum away: «N.H mm. as. nm\m Huumucmo maflmmonm woosuonsmflmz om.a nm.a ¢m. nm\n muumucmo mcfimmonm Hmcofimmm H¢.H um. no.v nm\m muczoe HHmEm no. «b. He. \{mmxw, muumuamo mcflmmozm coonuonnmflmz mH.m Nv.m 5H. ¢m\m Hummucmo mcflmmonm Hmaoflmmm mo. m¢. om. ~m\m Hummumsau swag: «v.a «m. mo.~ Hm\m muczoa HHmEm hm. hm. ma. om\o mumfluum can“: mm. mm. ma. v¢\oa Humfluum can“: NN.H «m. Ho.Hw H¢\H Hucsos HHmEm Xom Ron Hmmamm mcflcmmo wuouw >w>usm muflmmu me mo wpma Hmzom mcfl3mun ma mflm<fi HMOBm Nm>mbm Wm OZHmeO m0 Whfla QZ< mm30m UZHBRMQ .me¢m fiBHmdU mmm 117 .mhoum mmzsm 93 mo mswomu 332: 025 m QHSuHB mmawm muflmmu Hmma mm.a nm.a wm. nmwN muumucmu mcflaaozm amcofiumm mH.m m¢.~ ha. ¢m\m Huumucwo mcflmmosm Hmcoflmmm 5H.H mm. om. mm\m muumuamo mcflmmosm muflcseaoo vm.a mm. mm. mm\a H-umuamu mcflmmogm muflcseeou no. ¢n. H¢. om\a muumucmu mcflmmozm woonuonnmfimz vN.H mm. an. >m\m Hummucmu mcflmmogm oceanonnmflmz H¢.H mm. no.¢ nm\o mucgoa Hamem ¢¢.H @m. mo.~ Hm\m munzoa Hamsm mm.H wm. Ho.a H¢\H Alsace HHmEm mm. mm. mm. mm\m muumumsflu can“: no. m¢. mm. mm\m mnumumsao away: No. m¢. om. ~m\m Huumumsau can“: mm. av. mm. mm\H mumfluum can“: hm. hm. ma. om\m mumfluum can“: mm. mm. ma. m ¢¢\oa Humfluum aways kom Ammaflzv Eon Hmmamm mcwcmmo muoum wok/Ham “530m mcfismuo muflmmo me mo mama MAME mmOBm Mm UZHmeO m0 manna QEMNBOm GZH3H0u2H mocmumflm muoum >m>msm mmOBm >m>MDw QZ¢ A<>mmBZH moZdemHQ Nm mBMMm¢memDm m>HBHEmQZOU m0 mmmZDZ HN mqmflfi 119 om.H nm.a om. o muuwucmu mafimmosm Hmcofimmm mH.m N¢.N mm. H Hummucmu mcflmmonm HMGOHmwm na.a mm. om.a N muumucmo maflmmonm Nuflcsaeou vm.a mm. nn.H a Hummuamo mcfimmonm muflcsEEou no. «5‘ mm.a H muumucmo mcfimmonm noonnonnmfimz vm.a mm. mo.m o annmucmo mcflmmonm coonuonnmflmz H¢.H hm. mm.¢ o mucsoe Hamem fiv.a ¢w. Nm.N a Nlazoe HHmEm NN.H «m. mo.a H H-23os HHmEm mm. mm. mm.a o muumumsau sang: no. m¢. mm.a m Nuumumsao away: No. mw. Nm.a N HummumsHo GMQHD mm. av. mm.H o mumfluum away: mm. mm. mm.H H m-mfluum can“: mm. Nm. 0H.Hw A Humfiuum GMQHD mawz N\H muwamEmesm Rom X05 um mmamm w>auwummfiou mnoum mm>usm Hmzom mGHBMHQ muflmmo Hmm mo Hmnasz NN mamfifi mmOBm Vm>mDm Mm A<>mmBZH MUZHBHBmmSOU m0 mmmzbz 120 CHAPTER v SUMMARY AND CONCLUSIONS Introduction The purpose of Chapter V is to present the summary and con- clusions of the research findings. The chapter is divided into three sections. The first section evaluates the hypotheses presented in Chapter I in relation to the research findings. The second section relates these findings to the practical prob- lems of the supermarket location decision. The implications of the findings are discussed within the typical problem solving framework of the supermarket chain operator. Section three presents possible extentions of the research. Significant prob- lem areas not directly relevant to the research, as well as further refinements of the research are also presented and discussed. fizgluation of Hypothesis StorefComplex Hypothe§g§_ l. §§_£h§ product offering g§_§.£etail_complex increases (as measured by number and types of different stores) the drawing power of the supermarket increases. On the basis of Table 7 and Figure 8 and 9, the hypothesis is judged to be valid. In Table 7 a series of data are presented indicating an increase in drawing power by levels of store complex. 121 Table 8 offers additional support for the validity of the hypo- thesis by indicating the degree of consistency within each cate- gory of store type. The deviation in pattern shown by the Small Town type at the 90 per cent customer level (Table 7) can be explained by a comparison of mean and median values for the Small Town type store (Table 9). From the characteristics of the data in Table 7 and 8, three relatively homogeneous categories of drawing levels emerge. The Urban Strip and Urban Cluster types represent the first category; the Small Town, Neighborhood Shopping Center and Community Shop- ping Center types represent the second category; and, the Region- al Shopping Center type stands alone as a third category. 2..§ small town relatively isolated fromany other city demonstrates drawing power patterns similar 52 the medium sized shpppipg center (community). In the course of gathering basic data it was noted that the small towns selected for study had approximately the same number of retail businesses in their central business district as the large community shopping center (40-50 retail business units). Since the small town business area functions as a nucleated business district, it was hypothesized that the small town would demonstrate drawing power characteristics similar to the medium sized shopping center. S‘s: a? fl“ ‘- in 1...! es I...» u I “I ‘I ‘I I I I I I I I I n 122 On the basis of Table 7 and 8, hypothesis number two is judged to be a valid hypothesis. 3. As the product offering 23.2 retail complex increases (§§_measured by number and types 9: different stores) per capita sales gf_the supermarket increase. Per capita sales were calculated at three distance inter- vals, one-half, one and one-quarter and two miles. The dis- tnace categories used were subjected to analysis which indi- cated that the differences between distance intervals were highly significant (Figure 10). The hypothesis as stated is invalid at every distance in- terval. Both Table 13 and 14 present data which by inspection in- dicates that hypothesis number three would not hold. These tables (13-14) do, however, present a systematic pattern of data by distance interval. Per capita sales rise at every dis- tance interval up to the Small Town type and then fall; with the Neighborhood Shopping Center being lower than the Small Town type; the Community Shopping Center lower in per capita sales than the Neighborhood Shopping Center; and, the Regional Shopping Center being sharply lower from the Community Shopping Center level. It is also significant to note in Tables 13 and 14 the variation in per capita sales at the 1-1/4 and 2 mile distance shaal ¢!!. [IEEI illh R'II) ill) Ell. V'III {III ‘III III) illl III {III III iJIIrI:lI[ J 123 intervals. The wide deviation between the Small Town and other store types can probably be explained by the lack of competition in the small towns (Table 21). The deviation between the smaller shopping centers and the urban stores is more difficult to jus- tify on the basis of competition and are probably due to the degree of store complex. Store Size Hypotheses ling area) increases, per capita sales 9: the supermarket increase. The data relevant to the above hypothesis are presented in Tables 10 and 11. When the store size and drawing power varia- bles were correlated for the 70 per cent level of customer draw- ing power, a coefficient of correlation of +.575 was obtained. The coefficient of correlation is significant at the 95 per cent level of significance. When the store size and drawing power variables were correlated for the 90 per cent level of customer drawing power a coefficient of correlation of +.444 was obtained. The coefficient of correlation for the 90 per cent level of cus- tomer drawing power is significant at the 90 per cent level of significance, but not at the 95 per cent level. The store size hypothesis is judged as invalid on the basis of the distribution of data in Tables 10 and 11, and the low co- efficients obtained in the Correlation procedures. In Table 11 124 the data are classified and presented by store type. There are some indications that size of store has an influence upon draw- ing power within the store type classification, but further ob- servations wouldlxanecessary to test this observation. selling area) increases, per capita sales 9: the super- market increase. The above hypothesis refers to per capita sales within a two mile radius of the survey store. The importance and influence of the per capita sales dataame more effectively interpreted if the data are analyzed in reference to Table 12, which presents the percentage of customers living within a 2 mile radius of the survey stores. Data relevant to the hypothesis are presented in Tables 15 and 16. The variables of store size and per capita sales were correlated for the 1/2 mile, 1-1/4 miles and 2 mile distance in- tervals. The resulting coefficients of correlation were: -.540, +,l92 and +.l66, for the 1/2 mile, 1 1/4 miles and 2 mile dis- tance intervals, respectively. The coefficient of -.540 is sig- nificant at the 95 per cent level of significance. The coeffi- cients of +.l96 and +.l66 are not significant. The above hypothesis is judged to be invalid on the basis of the distribution of the data presented in Tables 15 and 16 and the low coefficients of correlation obtained in the Correla- tion procedures. Conclusions 125 The conclusions section is divided into two sections. The first section contains conclusions based upon the statistical findings contained in Chapter IV. The criterion of a 95 per cent level of confidence was established as the criterion of reliability for the conclusions. The second part of this section presents conclusions that emerge from the statistical and analy- tical treatment of the data. StatisticalfiFindings l. The relationship between drawing power (70 per cent cus- tomer level) and store complex is significant. 2. The Regional Shopping Center type is significantly dif- ferent in drawing power (70 per cent customer level) from all other store types. 3. The Urban Strip type is significantly different ing power (70 per cent customer level) from the ity and Regional Shopping Center types. in draw- Commun- 4. The relationship between drawing power (90 per cent cus- tomer level) and store complex is significant. 5. The Urban Strip type is significantly different ing power (90 per cent customer level) from the Town type. 6. The urban Strip type is significantly different ing power (90 per cent customer level) from the Shopping Center type. in draw- Small in draw- Community 7. 8. 10. ll. 12. l3, 14. 15. 126 The Urban Strip is significantly different in drawing power (90 per cent customer level) from the Regional Shopping Center type. The Urban Cluster type is significantly different in drawing power (90 per cent customer level) from the Small Town type. The Urban Cluster type is significantly different (90 per cent customer level) from the Community Shopping Center type. The Urban Cluster type is significantly different in drawing power (90 per cent customer level) from the Regional Shopping Center type. There is a significant relationship between drawing power (70 per cent customer level) and store size. There is no signficant relationship between drawing power (90 per cent customer level) and store size. The l/2 mile distance interval is significantly dif- ferent from the 1 1/4 and 2 mile distance interval in per capita sales. The 1 1/4 mile distance interval is significantly different from the 2 mile distance interval. There is no significant relationship between per capita sales at the 1/2 mile distance interval and store complex. l6. 17. 18. 19. 20. 21. 22. 23. 127 The Regional Shopping Center type is significantly dif- ferent than the Small Town type in per capita sales at the l/2 mile distance interval. There is a significant relationship between per capita sales at the 1-1/4 mile distance interval and store complex. The Urban Strip type is significantly different in per capita sales from the Community Shopping Center at the 1-1/4 distance interval. The Urban Strip type is significantly different in per capita sales from the Neighborhood Shopping Center type at the l-l/fijmile distance interval. The urban Cluster type is significantly different in per capita sales from the Community Shopping Center type at the 1-1/4 mile distance interval. The Urban Cluster type is significantly different in per capita sales from the Neighborhood Shopping Center type at the l-1/4 mile distance. The Regional Shopping Center type is significantly dif- ferent in per capita sales from the Community Shopping Center type at the 1-1/4 mile distance interval. The Regional Shopping Center is significantly different in per capita sales from the Neighborhood Shopping Center type at the 1-1/4 mile distance interval. 128 24. The Urban Strip type is significantly different in per capita sales from the Community Shopping Center type at the 2 mile distance interval. 25. The Urban Cluster type is significantly different in per capita sales from the Community Shopping Center type at the 2 mile distance interval. 26. There is a significant negative relationship between store size and per capita sales at the 1/2 mile dis- tance interval. Analyticalpponclusions Store Complex and Drawing Power Store Complex exerts a relatively significant influence upon drawing power, both at the 70 and 90 per cent drawing power level (Table 7). As the level of product offering at the retail site increases, the drawing power increases. However, the drawing power findings indicate that the increases in drawing power are not proportionate to the level of product offering at the retail site, but rather the data cluster into three general classifications. The first classification consists of the Urban Strip and Urban Cluster types. The second classification includes the Small Town, Neighborhood Shopping Center and Community Shopping Center types. The Regional Shopping Center type stands alone as the third category. The classification of drawing power by individual survey store further substantiates the classification 129 (Table 8). From these data the fact is deduced that there exists distinct levels of drawing power attraction clustered around different levels of store complex. These data are arrayed as follows to indicate the distinctive characteris- tics of drawing power by the above outlined classifications. TABLE 23 RANGE OF DRAWING POWER---GENERAL CLASSIFICATION BY STORE TYPE Classification Drawing Power 70 Per Cent 90 Per Cent l> .38 - .43 miles .56 - .62 miles 2 .79 — .87 miles 1.05 -l.38 miles 3 1.95 miles 2.53 miles If this type of classification is applied to the indi- vidual survey stores, the following array of data results: TABLE 24 RANGE OF DRAWING POWER---GENERAL CLASSIFICATION BY SURVEY STORE Classification Drawing Power 70 Per Cent 90 Per Cent 1 .32 - .49 miles .39 - .67 miles .54 - .98 miles .97 -l.44 miles 3 1.57 -2.42 miles 1.86 -3.15 miles 130 Thus, two general types of conclusions can be drawn on the basis of the findings regarding the relationship between store complex and drawing power. 1. As store complex increases, drawing power increases, but not proportionate to the level of product offering at the retail site. 2. There are distinct drawing power characteristics that tend to cluster at certain levels of store complex. Store Size and Drawing Power The independent variables of store size and store complex are related to some degree. That is, large supermarkets are built in a large shopping center and small stores are more likely to be found in a small town. When the findings are presented by rank order of store size (Table 10) there is no apparent rela- tionship between store size and drawing power. To remove the possible distortion caused by the interrela- tionship of store size and store complex, the data are also pre- sented by individual survey stores arrayed according to store type (Table 11). It would seem that by comparing store size and drawing power within the individual store type classifica- tions a more accurate reflection of the influence of store size upon drawing power might be obtained, since the store type variable would be held constant. H0wever, when the data are I I I I I I I IIIITIIHITIIIIJI 131 presented in this manner it again fails to indicate any systema- tic relationship between store size and drawing power. Thus, the conclusion must be drawn that on the basis of the findings there exists no systematic and reliable connection be- tween store size and drawing power. Store Type and Per Capita Sales Per capita sales estimates were made at the 1/2, 1-1/4 and 2 mile distance intervals from the survey store (Appendix C-4, E). Per capita sales outside a 2 mile radius from the survey store were not calculated since the major segment of customers for all store types with the exception of the Regional Shopping Center type were located within two miles of the survey store. In this regard, several special cautions should be exercised in the interpretation of per capita sales data. First, the Re- gional Shopping Center type draws over 65 per cent of its custo- mers from beyond the 2 mile distance interval. Hence, when per capita sales data are presented for the Regional Shopping Center type, it is relevant only to the 35 per cent of the customers of this store type. Another factor that should be noted is the lack of competition for the Small Town type stores (Table 22). The factor probably contributed to an unusually high per capita sales figure throughout the entire area under consideration. At the 1/2 mile distance interval, store complex has little effect upon per capita sales except in the case of the Regional 132 Shopping Center. There is a negative relationship between store complex and per capita sales when the total for the Regional Shopping Center type is contrasted to the remaining five types at the l/2 mile distance interval (Table 13). The convenience nature of the products sold by the supermarket probably accounts for the lack of variation in the data at the 1/2 mile distance interval. The low per capita sales of the Regional Shopping Center at the 1/2 mile distance interval is probably due to the fact that the typical consumer prefers not to become involved with the problems of large parking lots and heavy traffic if the sole objective of his shopping trip is for food purchasing. At the 1—1/4 mile distance interval a more distinct pattern of per capita sales emerges (Table 13). The influence of store complex upon per capita sales also becomes more evident. The data cluster in three general classifications at the l-l/4 mile distance interval. The first classification includes the Urban Strip, Urban Cluster and Regional Shopping Center types. The second category includes the Community and Neighborhood Shopping Center types. In the third category the Small Town type stands alone in per capita sales characteristics. The array and range of data at the 1-1/4 mile interval by store type are presented in Table 25. A similar presentation of data using per capita sales by individual survey store is presented in Table 26. TABLE 25 133 RANGE OF PER CAPITA SALES AT 1-1/4 MILE DISTANCE INTERVAL—- GENERAL CLASSIFICATION BY STORE TYPE Classification Range of per Capita Sales by Store Type 1 $.21 - .28 2 .62 - .71 3 2.64 TABLE 26 RANGE OF PER CAPITA SALES AT 1-1/4 MILE DISTANCE INTERVAL-- GENERAL CLASSIFICATION BY SURVEY STORE Classification Range of per Capita Sales by Survey Store 1 $ .15 - .31 2 .62 — .78 3 2.13 - 3.67 The distribution of per capita sales at the 1—1/4 mile distance interval indicates that store complex begins to influ- ence per capita sales somewhere between 1/2 and l-l/4 miles from the store. The negative relationship between the variables found in the Regional Shopping Center at the 1/2 mile distance interval continues to exert an influence at the 1-1/4 mile distance interval when contrasted with the per capita sales levels of the Neighborhood and Community Shopping Center types. 134 The consistency of these data is found both in store type (Table 13) and when classified by individual survey store (Table 14). The relationship between store complex and per capita sales becomesenmm.more apparent at the 2 mile distance interval (Table 13). Again, the data clusters around certain levels of store complex making possible distinct classifications. However, due to the fact that per capita sales for the Regional Shopping Cen- ter remain at a relatively high level, the classifications for- mulated for the 1-1/4 mile distance interval must be realligned. The general classification at the two mile distance interval is as follows: 1) Urban Strip and Urban Cluster; 2) Neighborhood, Community and Regional Shopping Center; and 3) Small Town. The range of per capita sales at the 2 mile distance interval using these classifications is presented in Table 27. TABLE 27 RANGE OF PER CAPITA SALES AT 2 MILE DISTANCE INTERVAL-- GENERAL CLASSIFICATION BY SURVEY STORE Classification Range of per Capita Sales by Store Type 1 $.02 - $ .06 2 .16 - .23 135 A similar distribution of data using the same classifications, but using the range of the individual survey stores within the general classification, is presented in Table 28. TABLE 28 RANGE OF PER CAPITA SALES AT 2 MILE DISTANCE INTERVAL-- GENERAL CLASSIFICATION BY SURVEY STORE Classification Range of Per Capita Sales by Store Type 2 .14 - .48 3 .67 - 3.76 It should be noted that both the high value for the range in classifications 2 and the low value for the range in classifi- cation 2 show a wide deviation from the mean value of the dis- tribution (Table 14). The distribution of per capita sales at the 2 mile dis- tance interval indicates that the influence of store complex on per capita sales becomes greater as the distance from the survey store increases. In summary the following conclusions are drawn from these findings: 1. All store types, with the exception of the Regional Shop- ping Center type, achieve relatively similar per capita sales in the 1/2 mile distance interval. 136 2. The congestion created by the number of stores and large parking areas provided by the Regional Shopping Center type effect a negative influence upon per capita sales in the 1/2 mile distance interval. 3. Per capita sales for the Small Town type are higher than other store types at all levels due to the fact that the consumer has a limited range of alternatives (stores) with which to fulfill food purchasing objec- tives. 4. Per capita sales for the Urban Strip and Urban CluSter types drop more rapidly than for the Community and Neighborhood Shopping Center types at the 1-1/4 and 2 mile distance interval. 5. The influence of store complex on per capita sales be- comes more prominent at the 1-1/4 and 2 mile distance intervals. Store Size and Per Capita Sales As was indicated in the above section on store size and drawing power, there is probably a relationship between store size and store complex. If the data are arrayed by.rank order of store size (Table 15) there appears to be a negative relation- ship between store size and per capita sales at the 1/2 mile distance interval. Correlation of the series indicates a 137 coefficient of correlation of -.540 which is significant at the 95 per cent level of confidence. Correlation between store size and per capita sales at the l—l/4 and 2 mile distance intervals produce coefficient of correlations of +.192 and +.l66 respec- tively. Neither of these coefficients are significant at the 90 or 95 per cent level of significance. These same findings are presented by store type and store size in Table 16 in order to minimize distortions due to the re- lationship between store size and store type. There is no dis- tinct pattern of store size influence even when this relationship is presented by store type classification. In summary it is concluded that on the basis of these find— ings there is no systematic and reliable connection between store size and per capita sales. Implication of Findings for Location Policy Introdpption The supermarket chain operator faces two types of problems when seeking to formulate effective policy and strategy for mar- ket development. They are: l) The problem of optimal network expansion where the objec- tive is to construct a network of stores that will pro- vide optimum sales and profit in any given market territory. 2) The problem of developing individual distribution points iiil iii! Ihil iii! hil‘ IEII 138 where the objective is to develop a profitable individual distribution unit. These two problems are closely related and the development of any individual location must be evaluated both on its profitability as an individual site and on its contribution to network expansion. For any given location decision, both the objective of optimal network expansion and individual store profitability should be achieved. However, in the competitive market place location stra- tegy might dictate that the objective of individual store prof- itability be subordinated to the objective of optimal network ex- pansion. That is, it is realistic to assume that some individual locations are developed with the prospect of future profitability and/or present market representation rather than immediate prof- itability.1 The factor of uncertain competitive adjustments to the loca- tion decision add a demension of strategy to the supermarket chain operator's location policy. The twin objectives of optimal net- work expansion and individual store profitability must be sought in an uncertain competitive environment. The degree of freedom of competitive reactions to the location decision is almost un- limited. A.practica1 limitation is provided by the fact that all competitors presumably formulate location policy in order to in- sure survival in the marketplace. 1"The Structure of Retail Competition in the Philadelphia Market," op. cit., p. 51. Egéi iL—i i—JF ifidl ‘lzi' lzéI ‘Esfl IEHI EEI’ IEII imam -pna. 1223 leak 139 If all competitors in the market place were aware of the economicslinvolved in the location decision, a more orderly and efficient competitive environment than currently exists in most metropolitan areas would result. That is, in almost every large city there are areas that are overstored and have too many supermarkets servicing the population of the area to allow a reasonable return on investment for the stores in the area. This type of situation would not develope if all supermarket operators sought to achieve the objectives of optimal network expansion and individual store profitability within an economi- cally justified framework. It is necessary that all competitors be aware of the eco- nomics of location development. One reckless competitor com— mitted to expansion at any cost can disturb the economic struc- ture of the market for all of the competitors in the market place. In summary it appears that one of the most pressing prob- lems facing the supermarket industry is an awareness of the economics involved in the location decision. Store Complex and Location Policy The concept of store type can be usefulin.developing sound location policy in several ways. The conclusion that a store placed in a certain retail complex will perform differently both 1 . . . The term "economics" as it is used here refers to the con- cept of return on investment. 140 in drawing power and per capita sales than when placed in another form of retail complex has some obvious and signifi- cant implications for optimal network expansion. The super- market chain either consciously or unconsciously has a great variety of store types within any metropolitan area. If these types are isolated and analyzed and their drawing power charac- teristics determined an optimum network of distribution points can be established and maintained by adding or closing individ- ual distribution points. Regarding the location decision to build a new store the implication of this research would be to build a store in the store complex situation that would match the geographical limits FIGURE 13 HYPOTHETICAL DISTRIBUTION NETWORK 141 StoreJTyp§_ Regional Shopping Center Community and Neighborhood Shopping Center Small Town Urban Strip and Cluster , Y, Z = Proposed Sites N:U O w W II of market opportunity. In Figure 14 a network of stores for a hypothetical city is outlined. The size of the hexagons corresponds to relative drawing power of store types. Points X, Y, and Z represent proposed additions to the network. Proposed store "X" is located on a major traffic artery in an urban area. The proposed store type is on Urban Strip and it appears that given this segment of undeveloped market opportunity that the development of this location would con- tribute to the network as a whole. Proposed store "Y" is lo— cated on the fringe of town in a new Neighborhood Shopping Center. As illustrated this location would also contribute to optimum network expansion. Proposed location "2" is loca- ted in a new Regional Shopping Center. The drawing power of the Regional Shopping Center would cause a store at this point seriously to compete with other segments of the network and hence a store developed at this point would be detrimental to optimum network expansion. The hypothetical analysis above assumes that there is sufficient potential available for market development and that competition is evenly distributed throughout the market area. .IJIIII‘IIIIII‘IIIII'IIIIIIIIIIII'II'IIE ..I- I I; n- _ 142 Since it has been established that different store types have different patterns of per capita sales, this concept may also be useful in estimating sales volume for a proposed site. If enough observations were made for a given store type, the following procedure might produce valid results in estimating sales volume for a proposed site. TABLE 29 PROCEDURE FOR CALCULTION ON ESTIMATED WEEKLY SALES FOR PROPOSED SITE Distance Per Capita Total Interval Population Sales galgg 1/2 10,000 $1.55 $15,500.00 1-1/4 20,000 .21 4,200.00 2 50,000 .02 1,000.00 Sales under 2 miles $22,700.00 Sales over 2 miles 2,110.00 Total weekly Sales $24,810.00 The basic formula used for the estimation procedure outlined above is: W = + + Total eekly Sales (Pa X PCSa) (Pb X PCSb) (Pc X PCSC) + (TSx) Where: Pa = Population within 1/2 mile of survey store. PCSa = Per capita sales figure for urban Strip type l/2 mile distance interval. Pb = Population in 1 1/4 mile distance interval PCSb = Per capita sales figure for Urban Strip type at 1 1/4 mile distance interval. 143 P = Population in 2 mile distance interval. PCS = Per capita sales figure for urban Strip type at 2 mile distance interval. TS = Total sales over two miles--The total is cal- culated in Table 30 by the following: 91. 5% + $22Jjoo TSx = 8.5% X Where: 91.5% = Per cent of customers from under 2 miles. 8.5% = Per cent of customers from over 2 miles. $22,700 = weekly sales within 2 miles of survey store X = weekly sales over 2 miles from survey store. The degree of accuracy of the procedure would be influenced by several factors. First, it seems to be a generally accepted fact that a supermarket's average weekly sales will decline over time. Estimating sales on the basis of currently operating supermarkets might then produce a disproportionately low sales estimate. The factor of declining weekly sales could be adjusted for in the formula quite easily, if the rate of sales decline were a constant. A second factor influencing the usefulness of the per capita sales measurement for estimating sales volume for proposed sites would be the amount and degree of competition. If further re- search can accurately establish the influence of competition upon per capita sales, then the factor of competition could also be 2 . . . This gradual decline in weekly sales was proposed as a common phenomena by a number of supermarket chain real estate executives interviewed by the author. 144 incorporated into a formula for estimating sales volume of proposed sites. Store Size and Location Policy There is no connection between store size and drawing power or per capita sales for the survey stores analyzed in this research. In general, the larger a store the greater in- vestment required in facilities. As the amount of investment in the location increases, a larger amount of sales are re- quired to maintain a constant return on investment. The increasing return on investment would presumably lead the developer to use a minimum size of store criterion in his development policy. Both a more profitable total network and more profitable individual locations would result from a min- imum store size criteria in development policy. On the basis of these research findings it would seem difficult to justify the increasing size of store on the basis of performance. There are probably promotional factor and long range planning factors that enter into store size deci- sions. Hewever, it would seem that the size of store decision should be critically evaluated by supermarket management in its' relation to profitable network and site development. WIIIIIJIIIIIII-INI.E J - .__‘.~.. 'Eifl Eéfil ii! on ad 145 The Dynamics of Location Policy and Strategy The establishment and maintenance of an optimal network of distribution points is a continous challenge due to a num- ber of dynamic factors in the marketplace. When the decision is made to develop an individual location, the site becomes a fixed target for competitive adjustments. Due to the fixed nature of the costs incurred in developing a site an error, once made, is likely to develop into a situation where the supermarket chain continues to pay both in terms of profits and poor market representation for a number of years. The huge shifts of population to the suburbs along with an increasing number of competitors cause a constant shift in the structure of optimal network. The alert developer must constantly reassess his market structure both in terms of profitability and in terms of how well his market is being served. The concept of store type and the implications of the store size variable provide some insights into a possible framework for planning and evaluating the distribution struc- ture of the modern large scale supermarket chain. The conclu- sions,while drawn on the basis of a limited number of obser- vations,provideaamethodology and approach to the problem of evaluating both the objective of optimal network expansion and individual store profitability. _—- —-— ‘E Ira-:5 I=I 146 Suggested Areas for Further Research There are a very limited number of publications dealing with location research. The researcher working in this field is handicapped to some extent by the lack of established meth- odology and the lack of comparative data. With the increasing emphasis upon mass merchandising and large scale retailing some of the traditional competitive silence on location pro- cedures and study should be breached. Three broad areas for further research are suggested. It is believed that a significant research contribution could be made in broadening the concept of store type. Six selected levels of store complex were chosen for this research. As noted above the levels of store complex range on a continuum from the free standing store to the Regional Shopping Center. A research study designed to isolate and classify additional store types would provide the supermarket industry with addi- tional tools for intelligent and efficient market development. Research devoted to exploring in depth the various classifica- tions of store profitability type would also prove of value in evaluating individual store profitability. A second area suggested for further research is research regarding the influence of the store size variable. The store size variable is not established as significant in this research. However, within the store type classifications there is some indication that this variable is important. Further observations 147 using a greater number and range of store size might contribute valuable additional insights into the economic justification of the large supermarket. A third broad area of inquiry is the study of the competi- tive environment of the supermarket industry. More specifically, what is the influence of competition upon the drawing power and per capita sales of the supermarkets? Several vague general- ization or rules-of-thumb are accepted in the supermarket indus— try regarding competition. There are some indications based on the present study that these rules would not stand the rigor of close observation. As the marketplace becomes more saturated and over storing becomes a more common situation the necessity of an answer to the question of competitive influence becomes imperative. If a contribution to more efficient competition is to be made by increased amounts of location research, several pre- requisites are necessary. First, the firms must stand ready to provide their retail units as laboratories for the collection of data. They must further be ready to share the results of their research through an independent research institute of integrity with competitor and non-competitor alike. The re- searcher must publish clear and concise methodology and results so that his research might be duplicated and his results validated. 148 Lastly, the contributions of many desciplixuas must be in- tegrated into the study of the retail structure of the market. The related and unrelated disciplines of marketing, transpor- tation and traffic, economics, economic geography and others have theoretical and empirical insights to offer the location researcher. 149 APPENDIX A Delineating Retail Trading Areas in urban Areas Hi. — _ 150 Delineating Retail TradingpAreas in Urban Areas Purpose The purpose of this appendix is to present several methods for delineating retail trading areas in urban areas. The pre- sentation of these techniques is divided into "procedure" and "evaluation". The procedure section gives a brief outline of how these techniques are employed, while the evaluation section points up the advantages and disadvantages of the different ap- proaches. This appendix presents techniques designed to delineate the trading areas of single sites or distinct clusters of stores (shopping center). Other techniques are available for delin-. eating large retail trading areas such as the trading area of an entire city. Some of the techniques presented here can be used in both situations, but they are presented here within the framework of single-site or cluster analysis. A comparative study of the effectiveness of various tech- niques for delineating the trading area of an entire community has been done by Isodore V. Fine.1 In the study, Fine uses seven techniques for delineating the trading area of a community including: l . .. Isodore V. Fine, Retail Trade Area Measurement Techniques as Applied to Fort Atkinson, Baraboo and west Bend, unpublished doctoral dissertation, Columbia University, 1953. 5!! ‘EE! SEQ! ses 151 1. Law of retail gravitation. 2. Automobile license plate analysis. 3. Bank check clearance analysis. 4. Credit record analysis. 5. Newspaper circulation analysis. 6. Retailer determined delineation. 7. Consumer interviews. He establishes the consumer-interview technique as the most accurate delineation and measure deviation from this model._ He concludes that automobile license plate analysis is the second most accurate form of community trade area delineation. Reilly's "Law of Retail Gravitation" was also first conceived as a method of establishing trading areas for large areas such as towns or cities. However, since its first formulation, it has undergone alterations which some authors claim has extended its usefulness to single-site and cluster locations. The envolve- ment of these procedures and current applications will be dis- cussed in the second part of this appendix. Introduction The delineation of retail trading areas is undertaken for many different purposes. These purposes range from a projected rejuvenation of the central business district to providing a framework for projecting sales volume of a proposed store. The 152 available techniques for delineation also vary widely in sophis- tication, accuracy, and cost. The existence of alternative tech- niques requires that the objective of the practical delineation problem be clearly defined so the most effective techniques may be employed in the research. In a business environment, the clinical approach to the re- search problem is often applied to trading area analysis. That is, in contrast to the scientific method, the data is collected, and the problems, relationships, and influences are deduced from the data without benefit of prior hypotheses. There are certain disadvantages to the clinical approach in business research with one of the main weaknesses being its heavy dependence upon the skill (or lack of skill) of the diagnostician. There are two basic approaches to the delineation of retail trading areas: l)the empirical approach, 2) the gravitational approach. The first of the two approaches, the empirical approach, depends upon primary data and, hence, is generally a more costly process. Depending, however, on the degree of accuracy required, it Can often yield the best results at the lowest net cost. Other techniques might yield results at lower cost, however, these results might not possess the degree of accuracy speci- fied by the research problem. The gravitational approach gen- erally relies upon secondary data and, for this reason, is often less costly in application. 153 This appendix is devoted to a consideration of the two approaches outlined above. The Empipical Approach The empirical approach, as mentioned above, relies upon primary data. The data may be collected at the store site (or proposed store site) or they may be collected through field interviewing of the estimated trading area. Following are some common procedures for both methods of empirical analySis. Site Survey 1. Customer Interview2 Procedure - Normally, the customer interviewing takes place within the store after the customer has completed his purchase. In the case of a self-service store, the customer is interviewed as he leaves the check-out stand. In the case of a personal- service type store, the customer in interviewed either While his purchase is being wrapped or after he has completed his purchase and is leaving the store. The interviewing is done randomly with a quota sample or a stratified sample can be established. The type of sampling would depend upon the purposes of the interview, if any, in addition to delineating the trading area. 2 . . . . The customer-interView procedure was used in this re- search. A copy of instructions to the interviewers follows in Appendix B. 154 If the interviewing period covers extended periods of time or more than one interviewer is employed, caution should be taken so as not to interview a customer more than once. In the case of a quota sample, a five to seven per cent allowance should be made to compensate for customers who have given non-existent addresses and recording errors. If the quota is exceeded, the excess customer interviews can be eliminated, using a table of random numbers. Evaluation - The principal advantage of the customer-inter- view procedure is its flexibility. In addition to the customer address, a whole range of other information can be obtained. Information such as: size of purchase, types of purchase, reasons for purchases, shopping habits, competitive data, etc. A pro- perly designed project can probe for useful attitudes, Opinions and habits as well as provide the necessary data for delineating the trading area. The main disadvantage of this procedure is its expense. Costs are incurred both for the interviewing and the spotting of customers" homes on a map of the area. The costs involved depend 3For examples of this technique see: Bart J. Epstein, "Eval- uation of an Established Planned Shopping Center," Economic Geo- graphy, January, 1961, pp. 12-21: William Applebaum and Richard F. Spears, "HOw to Measure A Trading Area," Chain Store Age, January, 1951, pp. 149-154, and Bart J. Epstein and Howard J. Green, "Store Location Analysis," Marketing Research in Action, Studies in Business Policy, No. 84, National Industrial Con- ference Board, Inc., 1957, pp. 85-87. 155 upon the experience of the interviewers and individuals doing the map work as well as the size of the store and dispersion of customers. However, using a quota sample of one interview per one hundred dollars of sales per week, as used in this research, 'would result in costs of between seventy-five and one hundred and fifty dollars per store. 2. Automobile License Plate Analysis Procedure - Prior to the actual recording of customer li- cense plates all employees' plates are recorded. Owner and employee automobiles have a small influence on individual stores, but in the case of shopping centers, they could provide enough influence to cause some distortion of data. In this procedure, a recorder is placed at every entrance to the parking lot. His job is to record all license plates entering the center or lot. In the case of a store with no off street parkings, this task becomes more difficult. Normally. the procedure when collecting data involving on-street parking is to analyze the parking situation and establish a parking zone for the store or cluster of stores under study. After the zone is established it is checked at short intervals (no longer than 10 minutes apart) and all new licenses are recorded. In both cases commerical vehicles are excluded from the sample or census. The next step is to obtain the addresses of the car owners from the state license plate bureau. (It should be noted that 156 the list of names and addresses obtained can also be used for mail questionnaire purposes.) These addresses are then plotted on a map and the trading area of the site delineated. Evaluation - The license plate check assumes that the auto; mobile is in the area for shopping purposes. The assumption seems entirely logical in the case of a shopping center or store parking lot. However, in the case of on-street parking, it is difficult to attribute the automobiles presence to any single store, or even group of stores. In the case of an off- street parking situation, area analysis rather than site analysis would provide the most accurate delineation. The outstanding advantage of the automobile license plate analysis is that the desired information can be obtained in a relatively short period of time. An added advantage is that license plate data can also be obtained for a competitive loca- tion and with it a competitive trading area can be delineated. One outstanding disadvantage of the automobile license plate analysis can be its expense. Many state motor vehicle depart- ments charge a fee on a per name basis for this information. For example, in Michigan there is a charge of fifty cents per license. If a license plate check of a Regional Shopping Center were made, it could cost three to five thousand dollars to obtain the names and addresses for the recorded license plates. In other states the individual has access to the records or can 157 hire a firm specializing in clerical recording service at a reasonable fee. Another disadvantage of automobile license plate analysis is that unless followed up with a mail questionnaire, nothing can be deduced about the size of purchase or place of purchase. 3. Prize Contests Procedure - The technique of delineating trading areas using prize contests is a fairly simple procedure. Usuallly a list of prizes is established appealing to a fairly wide range of customers. The rules for the contest are established. In some cases the customer is required to make a guess or estimate of something, and in other cases the customer has only to regis- ter to be eligible. He is given a slip of paper inside the store, or is offered the opportunity of clipping it out of a newspaper or circular and bringing it to the store. The coupon, when completely filled out, has the customer's name, address and phone number, along with any of the other procedural data re- quired by the contest. The addresses on these coupons are recorded and plotted on a map of the area and the trading area is delineated. Evaluation - The chief advantage of the prize contest ap- proach to trading area delineation is that the trading area can be delineated while at the same time store traffic is being in— creased. While this in a sense is an advantage, it is also a 158 disadvantage in that the promotional aspects of the contest might distort normal shopping patterns hence, distort the actual trad- ing area under normal circumstances. Another problem generally associated with the prize con- test type of approach is that it is difficult to select a list of prizes that will appeal to all age groups and sexes. If the prize list appeals to one segment of the potential customers more than another segment, a distorted trading area may result. 4. Check Cashing and Clearing Procedure - Payroll check and personal check used in pay- ment for merchandise provide another source of customer addresses. These addresses are recorded over a period of time. If payroll check constitutes a significant part of total checks taken in, then the recording period is established over the entire pay- roll cycle. In a factory area the payroll cycle would probably be one week in a largely white collar area it would probably be either a two week or monthly payroll cycle. Evaluation - The credit record approach also has the advan- tage of being a simple and inexpensive method of tentatively de- lineating the trading area. On the other hand, it also possesses a similar weakness to credit record analysis in that the use of checking accounts is not uniform throughout every area. The use of a checking account is influenced by the individuals’ age, marital status, occupation, etc. 159 Area Survey 1. Heme Interviewing Procedure - The home interview is generally used in connec- tion with the analysis of a proposed site rather than an existing location. Its purpose is usually to determine present shopping habits and willingness to change under certain circumstances. A probability sample can be effectively employed if the proper lists of data are available.4 After the sample is drawn, field interviews are made regarding shopping habits, attitudes, etc. Upon completion of the data collection phase, the data are proportionately expanded to cover the entire area and esti- mates of potential business are made. Evaluation - The home interviewing approach to retail trad- ing area delineation is undoubtedly the most accurate and useful form of analysis. It provides the researcher with both a geo- graphic and per capita sales projection for a proposed site. It is also useful in attacking specific problems for existing sites, such as shrinking sales volume in certain areas. The major disadvantage of the home interviewing approach is its costs. HOwever, when a site is being considered for 4 . "Sampling Methods for a Small Heusehold Survey" Public Health Monograph No.340, U.S. Department of Health, Education and welfare (Washington: U.S. Government Printing Office, 1956). 5 . ‘ . For a graphic example of this type of procedure see: Rich- ard L. Nelson, The Selection of Retail Locations (New York: F. W. Dodge Corporation, 1958) pp. 160-163. 160 development, the survey cost as per cent of total development cost might be relatively small in the long run. 2. Telephone Interview Procedure - Using properly trained personnel it is possible to obtain useful results through telephone interviewing. Using the cross index phone directory available in most larger cities the same sampling procedure used in the home inter- viewing is established. Then, instead of an actual visit, the names selected for the sample are contacted by telephone. The telephone interviewers use a standard interview guide to insure comparability of data and records all responses as they are made. After the telephone interviewing portion is complete the remaining procedure is the same as in (1) above. Evaluation - Telephone interviewing is more economical than field interviewing because it eliminates costly call-backs and can be done as time permits by personnel with less training than the field interviewing personnel would require. The major problem in telephone interviewing is the ease with which the person.being interviewed can terminate the interview. Refusal to answer any questions would presumably be higher by telephone than by personal interviewing, also an experienced field interviewer can be more flexible on turn-downs and more objective in evaluating the quality of a response. 161 The Gravitational Approach - The gravitational approach is based upon the concept of comparative advantages which states that customers direct their patronage to the retail site where the maximum utility returns per dollar and time unit may be obtained. Implicit in the presentation of a retail gravita- tional model is some level of consumer knowledge of what is . . . 6 available in the marketplace at alternative supply pOints. The Development of Reilly's Law of Retail Gravitation7- The original formulation of the "law" of Retail Gravitation was made by William J. Reilly and was stated as, "under normal conditions two cities draw retail trade from a smaller inter- mediate city or town in direct proportion to some power of the population of these two large cities and in an inverse proportion to some power of the distance of each of the cities from the smaller intermediate city.8 The formula presented in the original formulation was: pg = (g)N (212)N Bb (Pb) (Da) 6 For a general discussion of gravity models in regional analysis see: walter Isard (ed.), Methods of Regional Analysis (New York: John Wiley, 1960) ' It is more literally correct to term "Reilly's Law" "Reilly's Principle" of retail gravitation since by his own definition it does not cover all possible situations (Reilly, op. cit., p. 16) 8William J. Reilly, op. cit., p. 16. 162 Where - Ba = the business which City A draws from intermediate Town T. Bb - the business which City B draws from intermediate Town T. Pa = Population of City A. Pb = Population of City B. Da = Distance of City A from Intermed- iate Town T. Db = Distance of City B from Intermed- iate Town T.9 A simplified version of this law was developed by P. D. Converse. Breaking Point between A and B, miles from B = Distance between A and B l + ngulation of Town A Population of Town B Thus, the formula presented by Reilly is used to determine where the trade of an intermediate town will go between two com- pleting cities. The Converse formulation by assuming Ba = Bb in the Reilly formula is used to determine the breaking point for trade flow between two towns. Another application was developed by Converse in order to determine waht percentage of trade was kept within a city. For- mulations up until this point include only the factors of pop- ulation and distance and are only applicable in the case of 9Ibid., p. 48. 10 . Paul D. Converse, Harvey W. Huegy and Robert V. Mitchell, Elements of Marketing, 6th Edition, (New York: Prentice-Hall, 1958) p. 30 (First Present in 2nd Edition, 1935, p. 792). 11P. D. Converse, Retail Trade Areas in Illinois (urbana, Business Study No. 4, University of Illinois, 1946) p. 30-31. 163 shopping goods. The formulation by Converse is intended to predict the amount of fashion goods business that should be retained in any town. Mathematically, the formula is stated 12 as: a: - <22) (1)2 Bb (Hb) (d) Where - Ba = Proportion of trade going to the outside town. Bb = Proportion of trade retined by the home town. Pa Population of the outside town Hb = Population the home town d = Distance to the outside town 4 = Inertia factor The inertia factor was calculated from Mr. Converse's re- search of Illinois trading areas and trade flows. While pop- ulation and distance are the basic variables in these formula- tions, Converse offers the following alternatives: "The attraction of a shopping district may be measured by the population of the town, the volume of sales, or the square footage in fashion goods stores. The time and ex- pense factor may be represented by car or bus fare and time when public carriers are used." 12Paul D. Converse, "A New Application of the law of Re- tail Gravitation," Opinion and Comment, August, 1947 and Paul D. Converse, "New Laws of Retail Gravitation," Journal of Marketing, Vol. XIV, No. 3 (Oct. 1949) pp. 382-383. 3 1 Paul D. Converse, Harvey W. Huegy and Robert V. Mitchell, op. cit., p. 29. 164 In a recent article, Jung attempts to show how Reilly's Law does not properly predict trade flow in Columbia, Missouri. In his original publication, however, Mr. Reilly makes the fol- lowing statement which seems to negate Jung's exception. "In other words, every city presents an individual case with its characteristic differences, and the retail trade territory of any given city is the resultant of a highly complicated inter-relationship of a large number of factors rather than the resultant of the influence of one or two or three or fbur factors.15 Stating the quote above as his reason, Reilly then proposes a "provisional list rather than an arbitrary classification" of factors that may influence the retail trade territory of any given city.16 They are: TABLE 30 OUTLINE OF FACTORS INFLUENCING RETAIL TRADING AREAS 1. Lines of Transportation A. Public Highways B. Railroads and railraod rates--including special rates to commuters C. Electric Lines--regular and special rates 14 Allen F. JUng,"Is Reilly's Law of Retail Gravitation Always True?" Joprnal of Marketing, October, 1959. pp. 62-63. 15William J. Reilly, op. cit., p. 21. lerid., p. 21. 165 D. Bus lines--regular and special rates E. Waterways--regular and special rates F. Express and parcel post rates--regular and special G. Air lines Lines of Communication A. Circulation of the daily newspaper (1) Number of papers distributed (2) Geographical territory covered (3) Classes of people reached B. Telephone and telegraph lines and rates The Class of Consumer in the Territory Surrounding the Market Density of Population in the Territory Surrounding the Market Proximity of the Market to a Larger City Market The Business Attractions of the City A. The nature of the leading stores of the city (1) The kinds of goods and selections of goods offered by stores in the market (2) The delivery, credit, and other services offered by these stores (3) The general reputation of these stores as style- goods centers B. The extent to which the city offers storage and a market for the sale and redistribution of goods produced in the surrounding territory C. The banking facilities of the city The Social and Amusement Attractions of the City Theaters Educational institutions and facilities Musical attractions Athletic events Church, society, or fraternal gatherings Fairs and expositions WINDOWS, 166 8. The Nature of the Competition Offered by Smaller Cities and Towns in the Surrounding Territory A. The kinds of goods and selections of goods offered by stores in smaller locations. B. The general attitude of these surrounding cities and towns toward the larger city. 9. The Population of the City 10. The Distance Which Prospective Customers Must Travel in Order to Reach the Market, and the Psychology of Distance Prevailing in That Part of the Country 11. The Topographical and Climatic Conditions Peculiar to the City and its Surrounding Territory 12. The Kind of Leadership Offered by the Owners of Managers of Various Business Interests of the City.17 Reillyjs Law and Shopping Centersls- In 1953, James W. Rouse, President of the Moss-Rouse Company presented the possibility of the application of Reilly's law to planned shopping centers.19 The Rouse contribution is noteworthy in that it is perhaps the earliest attempt to adapt Reilly's law to a retail cluster with- in a large urban area. 17Source: Ibid., pp. 21-22. 18In 1951, Baker and Funaro made the following comment: "When measuring the pulls of a large Regional Center, Reilly's law may be applied just as aptly as it has been to downtown shopping centers." They did not elaborate or give example but from what followed this comment, the reader is lead to assume that the original formula can be applied rather than an adaption of the original formula. (Geoffrey Baker and Bruno Funaro, Shopping Centers: Desigp and Operation (New Yerk: Reinhold, "Progressive Architecture Library," 1951), p. 18. 9 James W. Rouse, "Estimating Productivity for Planned Re- gional Shopping Centers" News and Trends in City Development - urban Land Insititute. (November, 1953) pp. 1-5. 167 Rouse suggested that by substituting retail presentation of shopping goods in square foot area for size of city and convert- ing distance to driving time distance the principle of retail gravitation could be applied in urban areas.20 He restates the principle as: "Retail Shopping Centers and districts in a metropolitan area attract trade from the neighborhoods and communities com- prising the area in direct proportion to the shopping goods presentation at the district or center and in inverse propor- tion to the square of the driving time - distances between the ratail districts and centers and neighborhoods and communities." The technique presented is also proposed as useful in deter- mining the level of shopping goods sales for a proposed center. The conversion of the basic principle into this conclusion is vaguely presented by the author with little substantive comment. In a later article Mr. Leon W. Ellwood, moving from Mr. Rouse's article, restates the principle as: "The principle retail districts within a metropolitan trad- ing area attract trade from the residential sections of the area approximately in direct proportion to the size of the retail dis- tricts and in inverse proportion to the square of the driving 20Ibid., p. 3. 211bid.. p. 4. 168 time distance from each residential section to the retail districts."22 This presentation also suffers from vagueness in that "size of the retail districts" is not precisely defined. Basically, the Ellwood formulation differs from the Rouse formulation in that potential sales or per capita sales must be calculated. Then a modification of Reilly's Law is: . 2 Distance from A to B 3 Size A Size B applied to calculate relative "pulling" power between com— peting centers and the proposed center. The buying power within the pulling range of the proposed site is then cal- culated. Certain adjustments are made to this calculation in order to delete from consideration those stores and ser- vices that will not be included in the center. After the deletion adjustment has been made the remainder represents the potential volume of the proposed centers. Using the Ellwood formulation an optimum, relative size of proposed shopping center, is obtained. A third somewhat similar approach was presented by Harry J. 22 . . . Leon W. Ellwood, "Estimating Potential Volume of Pro- posed Shopping Centers," The Appraisal Journpl, Vol. XXII, No. 4 (October, 1954). pp. 581-589. 23 Ibid., p. 584. 169 24 . . Casey, Jr. a few years later. The follow1ng mathematical formula summarizes his approach: Fa 2 . _ (Dia) Bla Fa Fb Fc Fd Fe X ?1 2 etc. (35a)2 (63b) (63c) (Bi—M (Tie) Where i Bl Buying Power of Neighborhood 1 Bia- Purchases made by residents of Neighborhood 1 in the shopping Center A the square feet of retail space in the shopping centers A, B, C, etc. driving time distances between neighborhood 1 and other retail centers. Fa, Fb, Fc, etc. Dia, Dib, Dic, etc. The three adaptations of the Principle of Retail Gravita- tion outlined above have several things in common. First, all three methods are applicable to the retail complex commonly known as the shopping center. Secondly while most adaptations of Reilly's "law" have been concerned with calculating proposed sales volume or proposed size reflected by sales volume. A third factor and probably a weakness is that in every case, little, if any, solid empirical research is offered in support of the validity of the individual formulations. In summary, the gravitational approach to trading area de- lineation was originally developed to measure retail trade flows between tOWns of different sizes. In recent years new applications of the gravitational approach have been developed 4 2 Harry J. Casey, Jr., op. cit., p. 82. 170 which have application for delineation of the trading area of shopping centers. There are at least two additional areas of research in which contributions to the development of a relia- ble gravitational model can be made. Further empirical re- search is necessary to validate the reliability of the applica- tion of gravitational models to shopping center analysis.’ In the three gravitational applications cited in this section, hypothetical data is used for illustration and no evidence of raliability is presented. A second opportunity for contribu- tion is the extention of the application of valid formula to other types of retail store complexes in addition to shopping centers. 171 APPENDIX B INSTRUCTIONS FOR CUSTOMER SPOTTING Appendix B contains the instructions for customer spotting given to store interviewers. These instructions were provided through the generous cooperation of Professor Saul B. Cohen, Department of Geography, Boston University, Boston, Massachu- setts. 172 Instructions fongustomer Spotting Introduction Customer spotting is a method of getting customer ad- dresses and other related information po_aid_ip_store location. Procedure The following pages give an explanation of the proce- dure to be used to accomplish the tasks. Listed below is a brief summary of the information found on each page. Page 2 and 3 - Customer Interviewing Procedure. An explanation of how to interview, what to ask, and hours of interview. Page 4 - Copy of the Questionnaire. Page 5 - An explanation of the questionnaire. ill-1.1.5! I I I FEE I ' I — ' m 173 Customer Interviewing Procedure Position Stand back of the checkout booths where you can observe the customers, the total amount of their purchases, can encounter customers as they leave the checkout booth or are standing wait- ing for their order to be packed. Try to Stand in a position that will not block the normal flow of traffic. Pre-Interview Data While the customer is getting her order checked out, the following information can be entered on the interview card: Time, Sex, Departments Patronized and Sales Amount. Interview The checkout procedure is not to be interrupted and the customer is not to be approached until the cashier has com- pleted the cash transaction. Be courteous and smile as you approach the customer. Your conversation with each customer will be as follows: "Good morning. we are making a little survey for this store. Wbuld you mind telling me whether you walked or drove to this store today?“ (Fill in column headed Transportation on the interview card.) "How often do you shop at this store?" (Fill in column marked How Often on the interview card.) "How long have you traded at this store?" (Fill in column marked How Long on the interview card.) - 174 "And what is your home address, please?" (Fill in column marked Address on the interview card.) "Thank youi" If the customer makes inquiry concerning the purpose of the interview, you may say: "we are trying to see if our store is conveniently located near the homes of our customers." Reply to Customer's Inquipy Regarding Purpose of Interview (NOTE: Not for solicitations of any kind.) (NOTE: Do not record or ask for names.) A tactful manner on your part will easily secure 99% of the interviews for you. Never insist on an answer or subject a customer to "pres- sure." A refusal or misunderstanding on the part of any cus- tomer can be dismissed easily by merely saying, "well, thank you just the same.“ CUSTOMER ADDRESS IS THE SINGLE MOST IMPORTANTpITEM ON THE SURVEY FORM. LATER EACH ADDRESS WILL BE MARKED ON A MAP AND THE EXACT ADDRESS IS VITAL TO THIS OPERATION. ALWAYS ASK LOCAL CUSTOMERS FOR STREET AND NUMBER. ASCERTAIN WHETHER THE AD- DRES IS A STREET, AVENUE, ROAD, PLACE, ETC. IF A CUSTOMER GIVES ONLY AN RFD ZONE, BE SURE TO ASK FOR THE STREET NAME ALSO. REPEAT ALL ADDRESSES AS THEY ARE GIVEN. FOR OUT-OF- TOWN CUSTOMERS, THE NAME OF THE TOWN WILL BE SUFFICIENT UN- LESS YOU ARE INSTRUCTED OTHERWISE. d- -—'_— ‘ Q _~__ g 175 Do not "pick" customer, but take them as they come so that we will have a representative sample. Be especially sure that customers with small orders are not missed. Slow business--few checkouts working--try to get every cus- tomer. Steady business-~some checkouts working--4 consecutive in- terviews per checkout booth. Start at first open booth and ro- tate booths to cover all open booths before coming back to first open booth. Fast business--all checkouts working steadily--same as Steady Business above. Hours of Interview Friday - 10:00 A.M. - 1:00 P.M. - 2:00 P.M. to 7:00 P.M. - Total day - 8 hours Numbeg of Interviewers One interviewer, if store volume is under $20,000 per week. Two interviewers, if store volume is over $20,000 per week. 176 STORE INTERVIEW FORM BRANCH DATE STORE Time Sex Department Sales Amount 9 3 F A11 10 4 M GM 11 5 Ch (under 15) GP“ $ 12 6 Couple MP 1 7 Multiple G 2 8 M P CUSTOMER-ADDRESS Transportation How Often pep Week How Long lst time -3M 3 - 6M 6M - 1 yr. 1 yr.- 2 yrs. 2 yrs.- 5 yrs. 5-plus yrs. 6+ Since opening 6 w 2 > OWU'IQLAJNI-‘H lst Time Branch Store Date Time Sex Department Sales Amount Address Transportation How Often Week How Long 177 EXPLANATION OF QUESTIONNAIRE -Branch territory where survey is being conducted. -To include name and address. -Date that interview takes place. -Time of interview. Circle the hour nearest time of interview. -Sex refers to all persons in the shop- ping party, not only to the person who is interviewed. -Observe what the customer buys and mark the department(s) shopped. Watch the flags on the cash register to record meats, groceries and produce. -Cash register total sales. Record the sales figure in the space provided. -Address of the customer interviewed. Get the complete address—-Street, Ave- nue, Road, Place, Rural Route, etc. Don't accept names of apartments, but insist upon full address. -Record the method of getting to the store--auto, taxi, streetcar, bus, etc., are all important. —Record the number of times the cus— tomer visits this store in a normal week's time. -Record the length of time that the cus- tomer has been shopping in this store. Note: Carefully encircle each applicable item on the form. Also be sure to write each address legibly. ‘ !t! ‘ -- ~I; -! ! E g Q ~——4 --H ---I l-II IIII III III. III. IIII IIII III. ‘II' fill] 178 APPENDIX C FORMS' WORKSHEETS AND MEASUREMENT DEVICES USED IN RESEARCH PROCEDURE Appendix C presents copies of the forms, worksheets and measurement devices developed for this research. The appendix is divided into five subsections. C — 1 Store Survey Data - The store survey form was designed to record the basic data necessary for analysis of the survey store. The site map was drawn to reflect the mix of stores with 1/3 mile of the site for unplanned sites and the entire shopping center plan for planned sites. The area map was constructed by mounting a three mile diameter section of a street map on the sheet and spot- ting competition by color code on this map. Customer Spotting Overlay - In its original form, Ap- pendix C-2 was drawn on a transparent plastic overlay and scaled to the customer spotting map. Customer Spotting Record BasigpProcedure for Popplation Estimates - The graphic presentation illustrated in Appendix C-4 was originally drawn directly on to a detailed street map of the area. The Census Tract and/or Enumeration Districts were then also outlined on the map and the population and housing _i___———-—Islsl'aulsflh!lI-I1IHE; 179 percentage estimated by tract or district were recorded on Appendix C-5. C - 5 Population Estimate Form 180 No. Date Time STORE SURVEY DATA I. CITY and SITE DATA ADDRESS CITY SIZE NEAREST CROSS STREETS STREET WIDTH II. STORE DATA STORE NO. .1 TYPE STORE SIZE SELLING AREA STORAGE AND WORK SUB TOTAL BASEMENT TOTAL SPECIAL DEPARTMENTS -- DATE OPENED,f DATE REMODELED DATE SURVEYED 181 III. WEEKLY SALES ORDER NUMBER OF INTERVIEWS FIRST WEEK SECOND WEEK 3 AVERAGE SALES: (Total Sales) 4 THIRD WEEK NO. OF INTERVIEW8= (Average Sales) FOURTH WEEK 100 TOTAL IV. COMPETITI ON I NAME LOCATI ON E STIMATED S I ZE PARKING REMARKS Jill—hlhllllflflflflflfififi. u 1 I - ‘- i . 182 V. STORE COMPLEX NUMBER TYPE LOCATION ESTIMATED SIZE PARKING WI I. I. I I I I II I I. I. I I. I. I, -: 183 SITE MAP VI. I, n“. 184 VII. AREA MAP \flj/ KK/K/KQN .II- IIIIIIIIIIIIIIIII APPENDIX C-3 186 Store No. Page 0f . . uad . ad #0. Dlstance ‘and yNo. Dlstance'—g—~ No. Dlstancem'gu Dlr. No. D1r. No. D1r. N04 I El. T1. T1. II.I-IIIIIIIIIIIIIII 90 APPENDIX C-4 187 = Quadrant 180° /\ k A Survey Store Site 0 270 —%——_‘ Store No. APPENDIX C-5 Page Mile Zone 188 of Census Tract E.D. Quad Total P & H Pop. Housing Percentage Corrected P & H Pop. Housing nun-IIIIIIIIIIIIII! 189 APPENDIX D STATISTICAL PROCEDURES Appendix D describes the statistical procedures, formula and identities used in the application of the Analysis of Variance and Multiple Range Tests to the data collected for this research. D-l 'Analysis of Variance (Single Classification Problem) D-2 Analysis of Variance (Two-way Classification Problem) D-3 Multiple Range Tests 190 APPENDIX D-l ANALYSIS OF VARIANCE (Single Classification Problem) The fundamental identity for a one way classification is; TSS = Between SS + Within SS, where, SS represents sum of squares. ~ .th . Let Xi represent the 1 observation. The formulae are: 2 - 2 - - 2 Between SS = n1(A - X) + n2(B - X) . . . . nm(Z-- X) 2 2 _A___i Bi £3 n1 n2 nm 2 (2x.) - 2 2 1 Total $5 = 2(x. - X) = 2X. - , l 1 Nr Within SS = Total SS - Between SS . .th . Where xi = 1 observat1on X = average of all observations nl = number of observations in first column n2 = number of observations in second column . . »th nm = number of observat1ons 1nrn column A = average of first column B = average of second column 5 = average of gm'column A = sum of first column B = sum of second column 191 t sum of B h column [’13 ll 2 ll total number of observations The degrees of freedom are computed as follows: Total = N - 1 Between = Number of columns - 1 Within = Total degrees of freedom - Between DF . . . SS The mean square column 1s computed us1ng the rat1o BE'for the between and within categories. Between Mean Square T F a ' t' . . he v lue is computed by the ra 10 of W1th1n Mean Square III=IJIIIIIIIIIIIII I I I I I I I I I I I I I -‘u‘-I-I--!!!S!!gg 192 APPENDIX D-2 ANALYSIS OF VARIANCE (Two-Way Classification) The fundamental identity for a two-way classification is; TSS = Row SS + Column SS + Error 88, where, SS represents sum of the squares. .th . Let Xi represent the 1 observat1on. The formulae are: 2 2 (in) S = Z . - —--—- Total 5 x1 N 2 ZR.2 (in) Row SS = ‘1 -'——7;—- ni 2 2ci (in)2 Column SS = - -—-—-—- n N R c Error SS = Total SS - Row SS - Column SS where: .t . Xi = the 1 h observat1on Ri = sum of the ith row Ci = sum of the ith column n.C = number of columns nR = number of rows N = total number of observations -‘-‘-----!!!!!!a 193 The degrees of freedom are computed as follows: Total = N - 1 Row = number of rows - 1 Column = number of columns - 1 Error = total DF - Row DF - Column DF . . . SS The mean square column 13 computed us1ng the rat1o SE for each category. The F value is computed by the ratios, Row Mean Square Column Mean Square Error Mean Square ' Error Mean Square 194 APPENDIX D-3 MULTIPLE RANGE TEST When it is ascertained from the Analysis of Variance that there is a significant difference in the data, a device 1 is utilized for establishing termed the Multiple Range Test these significant differences between the individual columns. Using specially constructed "Significant Studentized Ranges for the 5% levelz‘ the values for the number of items to be compared are found. This table is entered to the degrees of freedom of the error term (within a single classification). The values found in this table are multiplied by a constant \/%§§', where EMS = error (within) mean square; SE = R%I 2 (Zn - éfir'); where K = number of classes and n = number of observations per class. EMS . Table . The above computed value (1(—SE- Value_) 18 then tested against the average of the classes. If A, B . . . . E represents the average of each class and A plus the computed value is less than B it can be stated that they are significantly different. All possible combina- tions are tested in this manner and the significance or lack of significance between classes established. 1David B. Duncan, "Multiple Range and Multiple F Tests," Biometrics, Vol. ll, 1955, pp. l-42. zIbid., pp. 3-4. g 195 APPENDIX E EMPIRICAL DATA SURVEY STORE POPULATION ESTIMATES BY QUADRANT AND DISTANCE INTERVAL Appendix E presents the population estimates upon which per capita sales are based by survey store. These data are presented both by quadrant and distance interval by individual survey store. 196 APPENDIX E SURVEY STORE POPULATION ESTIMATES BY QUADRANT AND DISTANCE INTERVAL Distance Quadrant Survey Store Int al Total erv 1 2 3 4 Urban Strip-1 1/2 mile 2537 1953 1953 2962 9405 1-1/4 mile 12854 8174 4367 10753 36148 2 mile 18988 5354 5167 12216 41725 Urban Strip-2 1/2 mile 1275 2278 2094 1471 7118 1-1/4 mile 9647 9797 13160 5118 37722 2 mile 18235 16218 26937 15697 77087 Urban Strip-3 1/2 mile 4160 4241 2315 2151 12867 1-1/4 mile 14841 22834 19506 9863 67044 2 mile 29799 29625 31024 24489 114937 Urban 1/2 mile 1442 1901 2443 2850 8636 Cluster-l . 1-1/4 mile 11600 8774 11778 9627 41779 2 mile 15592 19886 18499 6554 60536 Urban 1/2 mile '2924 2205 3247 4103 12479 Cluster-2 1-1/4 mile 21618 15969 11000 18403 66990 2 mile 36600 28135 19757 29595 114087 Urban 1/2 mile 4886 3980 2320 2948 14134 Cluster-3 1-1/4 mile 15537 20310 15300 9738 60885 2 mile 14419 41734 30485 19994 106632 Small Town-l 1/2 mile 832 400 1032 1063 3327 1-1/4 mile 75 86 129 320 610 2 mile 172 127 126 178 598 Small Town-2 1/2 mile 992 992 992 991 3967 1-1/4 mile 158 339 337 316 1148 2 mile 299 637 637 449 2022 Small Town-3 1/2 mile 439 200 438 676 1753 1-1/4 mile 150 150 100 200 600 2 mile 100 100 75 150 425 197 APPENDIX E. Continued Distance Quadrant Survey Store I t al Total ner" 1 2 3 4 Neighborhood 1/2 mile 408 273 536 619 1836 Shopping 1-1/4 mile 1293 1685 3168 3292 9438 Center-1 2 mile 1472 5447 2615 1988 11522 Neighborhood 1/2 mile 118 1788 771 2001 4678 Shopping 1-1/4 mile 2049 4878 6796 6569 20292 Center-2 2 mile 17375 10964 7523 11243 47105 Community 1/2 mile 755 982 439 306 2482 Shopping 1-1/4 mile 5715 7489 5090 1472 19766 Center-2 2 mile 5037 16547 13342 7838 42764 Community 1/2 mile 1486 1746 1091 1614 5937 Shopping 1-1/4 mile 5941 6325 8388 7951 28605 Center-2 2 mile 6124 10883 8174 6983 32164 Regional 1/2 mile 431 1257 891 76 2655 Shopping l-l/4 mile 8816 7807 11746 2638 31007 Center-1 2 mile 14711 19153 19274 2513 55651 Regional 1/2 mile 530 732 623 1496 3381 Shopping 1-1/4 mile 6381 9434 14441 10978 41234 Center-2 2 mile 14142 16832 19718 15841 66533 !!-! !F!! !!5 a!!! ll .! _ 1 1 198 APPENDIX F EMPIRICAL DATA SURVEY STORE CUSTOMERS BY DISTANCE TRAVELED TO SURVEY STORE AND QUADRANT Appendix F presents summary results of the customer spotting maps. A distance interval of one-eighth of a mile was established for measurement purposes. When the distance is specified as .125 miles the customers' home was located between 0 - .125 miles from the survey store. The tables in Appendix F are arranged so that at least 95 percent of the survey store customers are identified by distance traveled with only five percent or less of store customers in the "over" classification. - - ‘ . . . m g ! APPENDIX F-l 199 EMPIRICAL DATA - SURVEY STORE CUSTOMER BY DISTANCE TRAVELED TO SURVEY STORE AND QUADRANT (URBAN STRIP) Distance Interval Urban Strip-l Urban Strip-2 Urban Strip-3 Quad Quad Quad Quad Quad Quad Quad Quad Quad Quad Quad Quad (MlleS) 1 2 3 4 l 2 3 4 1 2 3 4 .125 4 1 2 3 2 3 5 4 3 3 2 3 .250 9 3 10 11 4 5 13 14 17 15 22 16 .375 6 7 5 10 9 1 12 11 23 27 10 9 .500 9 5 8 10 10 6 18 13 24 28 3 17 .625 4 7 5 6 4 5 3 8 - 23 27 1 20 .750 2 8 5 10 6 3 4 5 13 23 5 7 .875 - - 2 1 - 2 2 4 7 21 10 6 1.000 1 2 — - - 3 - - - 2 5 - 1.125 - 1 3 1 4 - - 3 - 1.250 1 2 3 2 2 1 3 2 1 1.375 2 2 1 1 7 - 1 4 1.500 ‘ 1 2 1 2 1 - - 1.625 1 - - 1 1 1 2 1.750 1 1 - 2 - 1 - 3 1.875 2 1 - - - 1 2.000 1 1 1 3 1 - 1 2.125 1 1 l - 1 1 2 2.250 1 1 1 - - 2.375 3 - 3 1 1 2.500 1 1 1 - 1 1 2.625 1 1 1 2.750 1 1 1 2.875 1 1 3.000 4 1 3.125 3.250 1 3.375 1 1 3.500 1 1 4 1 > 3.5 miles - 2 - 2 5 l l 10 - 5 6 Total 36 36 39 54 50 50 68 74 143 160 .72 104 200 APPENDIX F-2 EMPIRICAL DATA - SURVEY STORE CUSTOMERS BY DISTANCE TRAVELED TO SURVEY STORE AND QUADRANT (URBAN CLUSTER) . Urban Cluster-1 Urban Cluster-2 Urban Cluster-3 Distance §;::::?1 Quad Quad Quad Quad Quad Quad Quad Quad Quad Quad Quad Quad 1 2 3 4 1 2 3 4 1 2 3 4 .125 1 3 5 3 4 - 3 7 5 5 2 .250 4 1 3 28 13 21 7 9 23 20 16 17 .375 5 10 10 46 17 26 12 13 21 29 34 27 .500 6 5 13 26 17 27 12 19 13 35 22 21 .625 2 8 12 17 12 6 8 13 16 9 12 12 .750 1 10 13 10 16 1 14 11 9 16 1 14 .875 2 4 6 6 4 4 7 16 14 6 4 7 1.000 - 5 7 5 1 4 2 9 9 10 1 6 1.125 - 3 3 2 2 11 9 6 2 3 1 1 1.250 2 1 10 1 2 5 4 10 1 5 1 8 1.375 - 3 5 - - 1 3 9 - 2 - 5 1.500 1 3 6 - 1 2 9 l7 1 6 4 3 1.625 - 1 1 - - 2 2 2 2 9 - - 1.750 2 2 2 1 - 3 3 2 3 3 2 2 1.875 2 1 - l 1 3 8 4 1 - 1 2.000 3 2 1 - 1 - 2 2 3 - 2 2.125 1 2 - - - - 2 1 - 3 2.250 1 2 2 2 1 1 2 6 - 1 2.375 1 1 - - - 2 1 - 2.500 1 2 2 1 - 1 - 2 - 2.625 1 1 1 2 1 2.750 1 - 1 2 2.875 1 1 1 1 2 1 3.000 1 1 3.125 1 1 3.250 1 2.375 1 1 1 1 3.500 1 1 1 2 l > 3.5 miles - 6 — 4 - - 10 11 6 l 2 2 Total 29 76 107 156 92 121 107 167 139 180 109 136 APPENDIX F-3 201 EMPIRICAL DATA - SURVEY STORE CUSTOMERS BY DISTANCE TRAVELED TO SURVEY STORE AND QUADRANT (SMALL TOWN) Distance Interval (Miles) Small Town-1 Small Town-2 Small Town-3 Quad Quad Quad Quad 1 2 3 4 Quad Quad Quad Quad 1 2 3 4 Quad Quad Quad Quad 1 2 3 4 .125 .250 .375 .500 .675 .750 .875 1.000 1.125 1.250 1.375 1.500 1.625 1.750 1.875 2.000 2.125 2.250 2.375 2.500 2.625 2.750 2.875 3.000 3.125 3.250 3.375 3.500 Alhablard IO‘OON H Irate: I-‘l NU’ll-‘O‘IDO‘NI INCDNO‘U'IUJUJ ll-‘IwQJNH I I—‘ I raulnlh'o»mubcuw>01UIN H ::ara: l-‘ONNI l—‘Nl—‘I—‘Nl Il-‘Nl-‘l-‘INIW I WbI-‘WNOI ball-“QM F‘l Nlarardhab-N mud I H P‘ P‘OIP‘I w l-" NI—‘NINI HI‘F‘EJQJI H IFA¢>UImIA HEJP‘BJF‘O)fild ll G\G\P'N moonmo H bum N vald lehJP‘ ‘3 1... "lll ‘lll llll APPENDIX F-3. Continued 202 A I ‘II! III! III! 'III |lll III! III! III! III. In. Distance Interval (Miles) Small Town Small Town-2 Small Town—3 Quad Quad Quad Quad 1 2 3 4 Quad Quad Quad Quad 1 2 3 4 Quad Quad Quad Quad 1 2 3 4 3.625 3.750 3.875 4.000 4.125 4.250 4.375 4.500 4.625 4.750 4.875 5.000 5.125 5.250 5.375 5.500 5.625 5.750 5.875 6.000 F‘H I F" F'H'H H I H I N NII—‘NINI—‘NH Over 6 miles Total 16 22 33 35 52 33 64 84 32 47 34 75 203 APPENDIX F—4 EMPIRICAL DATA - SURVEY STORE CUSTOMERS BY DISTANCE TRAVELED TO SURVEY STORE AND QUADRANT (NEIGHBORHOOD SHOPPING CENTER) Neighborhood Shopping Center-1 Neighborhood D. . 1stance Shopping Center—2 Interval (Miles) Quad Quad Quad Quad Quad Quad Quad Quad .rfil _ __._,I I1 f 1 _I : - - - - - « _ r' ~e‘ H a z I I ‘III IIIII III“ 1 2 3 4 l 2 3 4 .125 .250 .375 .500 .625 .750 .875 1.000 1.125 1.250 1.375 1.500 1.625 1.750 1.875 2.000 2.125 2.250 2.375 2.500 2.725 2.750 2.875 3.000 3.125 3.250 3.375 3.500 3.625 3.750 3.875 4.000 prays: H II»: k)wIUFJUJF'HI§P‘O\UIw H'NI‘F‘P‘F‘H I H-HF‘| W I II k)U|H II I PIHINIHIoIoka¢>FJUIGIw NIHIo Ira: H H‘H IIDIbInLDCDbaUIK>o btdtvrdtnkd HIUF‘N I RIO HIMUJWIQK)NIHO\GIHKJH II FIJI H HIaId ocncura N M II Wtthl hquJF‘NI‘kJfilflko P‘H NIUI ®I4¢>®(D\JWDQ\J® IF‘N I F‘N Ik‘b IUJHIQI H‘H H'w.brokarabao-H1mcu~dr4\1m blvkotfi¢>l III-IIII-I - I I I I I, I. IIIIIIIIIII I. 204 APPENDIX F-4. Continued Distance Interval (Miles) Neighborhood Shopping Center-l Neighborhood Shopping Center-2 Quad Quad Quad Quad 1 2 3 4 Quad Quad Quad Quad 1 2 3 4 4.125 4.250 4.375 4.500 4.672 4.750 4.875 5.000 00% Over 5 miles Total 58 40 19 94 45 100 108 92 205 APPENDIX F-S EMPIRICAL DATA — SURVEY STORE CUSTOMERS BY DISTANCE TRAVELED TO SURVEY STORE AND QUADRANT (COMMUNITY SHOPPING CENTER) Community ’ Community Distance Shopping Center-l Shopping Center—2 Interval (Miles) Quad Quad Quad Quad Quad Quad Quad Quad 1 2 3 4 l 2 3 4 .125 .250 .375 .500 .625 .750 .875 1.000 1.125 1.250 1.375 1.500 1.625 1.750 1.875 2.000 2.125 2.250 2.375 2.500 2.625 2.750 2.875 3.000 3.125 3.250 3.375 3.500 3.625 3.750 1 3.875 2 1 4.000 3 1 I I qu I H‘ H AJUJG\Q)N>¢-w H F‘F‘F’P‘P’H UIN<3<3uJoxunw FJA)P‘P‘H'H \JGINIHIaFJGIm HIH I" U'I F‘HI»FHG\\lb-htdh3¢>mnbboI I I I II—‘I-‘II-‘II-‘NNI IN N IlNI—‘I—‘NNI—‘II—‘I I IHHIWHNHbNIbmWOWWHLflwwLflIbID-H P'H I I I I IN I I INWHwanw-fiLAJIWUIUIOOWWKO I—‘I—‘I—‘LdNhNHNNH-PWWI—‘NHWWWGCDNI—‘Ihwww l—‘NU'IU'II-‘wNWNHHNme-DINPNIDINNUIONOCDIfiLfiI-J n: kah‘k'MIaraI O\F‘h'N-b~JPJ\IwIOLuF4uIm APPENDIX F—S. Continued 206 Community Community Distance Shopping Center-l Shopping Center-2 i;§:::?l Quad Quad Quad Quad Quad Quad Quad Quad 1 2 3 4 1 2 3 4 4.125 2 1 4.250 2 l 3 1 4.375 1 l 2 4.500 1 4.625 1 l l 2 4.750 1 4.875 1 1 5.000 2 l 1 5.125 2 5.250 1 1 5.375 1 1 5.500 1 1 5.625 5.750 2 1 5.875 1 l 1 6.000 2 1 Over 6 miles 3 2 - - 3 l 2 Total 89 98 59 16 57 146 197 126 APPENDIX F-6 207 EMPIRICAL DATA - SURVEY STORE CUSTOMERS BY DISTANCE TRAVELED TO SURVEY STORE AND QUADRANT (REGIONAL SHOPPING CENTER) Distance Interval (Miles) Shopping Center-l Regional Regional Shopping Center-2 Quad Quad Quad Quad 1 2 3 4 Quad Quad Quad Quad 1 2 3 4 .125 .250 .375 .500 .625 .750 .875 1.000 1.125 1.250 1.375 1.500 1.625 1.750 1.875 2.000 2.125 2.250 2.375 2.500 2.625 2.750 2.875 3.000 3.125 3.250 3.375 3.500 3.625 3.750 3.875 4.000 4.125 4.250 4.375 NI—‘N-pUTI-‘QI—‘NWI-‘I IF‘FJOIWIHLQBJquIb IRJRIH.brd¢>uIM I H U'INOCDU‘INIDIIPU'IO‘IU'lnpNNUJUImIp-QO‘U'IU'IWIhWQNI—‘I lI-‘UJI NWWO‘WIfiNwwNO‘IQO-DUINW\IO‘MNNWGO‘QWWNNHI | F‘kIPIH-fi-btut‘¢>U1UIN II‘k‘h‘k‘bINIthJ>4>¢>P‘l H Iran I N NMNHNI—‘meme-prbWNle I INI—‘UJI-‘I-‘I I NIAIAIHI I-‘NNII—‘NI-‘I—‘WNI—‘WI—‘Ulthbmthmmw F‘ H P'b-DtflCDthJGJCIGIJCNLnO\k)\luI® mcptnua\IUImsocn~JO\HIwI»roI H IWI-‘fl-bU'IUJI—‘NI II-‘Il-‘I-‘INNI ====gzaa STE _-===—u=u=n=zgggst=5= 208 APPENDIX F-6. Continued Regional Regional Distance Shopping Center-l Shopping Center-2 Interval (Miles) Quad Quad Quad Quad Quad Quad Quad Quad 1 2 3 4 1 2 3 4 4.500 4.625 4.750 4.875 5.000 5.125 5.250 5.375 5.500 5.625 5.750 5.875 6.000 6.125 2 6.250 6.375 6.500 6.625 6.750 6.875 7.000 7.125 7.250 7.375 7.500 7.625 7.750 7.875 8.000 NI—‘Nl—‘LAJI IN I-‘I NNII-‘I I—‘I—‘NbNWNfiN H p l P‘P‘H ltptnl F‘H‘N Irardl dim H N NwHHIHthI—‘fi-NwmmlwbI—‘LDHN INNNNHHHHNwHIN|HNHHH| NINNI—‘I—‘NH-fi-NWIh-l It‘l ual N I HInrdI H NNNI—‘bNI I-‘ I I-‘ I-" Over 8 miles 6 18 7 8 - - 2 Total 89 214 162 111 79 173 256 40 I] I I II [J I II II I I l I I l I I I I I 209 APPENDIX G EMPIRICAL DATA LOCATION OF COMPETITIVE SUPERMARKETS BY SURVEY STORE The locationscfifcompetitive supermarkets are presented in Appendix G. The competitive stores are identified by distance and degree using the intersection of the N—S and E—W quadrant lines as the reference point. Identification - letters (i.e. A, B, C, D, etc.) indicates competitive chain organizations. That is, all of the stores labeled "A" belong to the same chain of supermarkets operating in the survey area. The let- ter "X" denotes a sister store of the survey store. 210 APPENDIX G LOCATION OF COMPETITIVE SUPERMARKETS BY SURVEY STORE . Chain Survey Store Dtiiigg? Degree Identi- fication Urban Strip-1 .40 1050 D .70 140 A 1.00 35 B 1.13 210 C 1.50 269 B 1.50 310 .A 1.50 342 D 1.63 271 A 1.88 302 B 2.00 7 C Urban Strip-2 .45 272° c .55 287 A .63 142 D .90 190 A 1.05 21 C 1.14 88 C 1.25 220 B 1.30 250 D 1.60 330 C 1.70 128 B 1.70 321 A 1.80 310 X Urban Strip-3 .70 570 B .80 66 A .80 302 B .93 298 A 1.15 252 A 1.20 262 B 1.45 17 A 1.55 16 X 1.67 79 X 1.70 139 X 1.70 309 X 1.75 64 A 211 APPENDIX G. Continued Distance Chain Survey Store . Degree Identi- (Miles) . . fication Urban Strip-3 (Continued) 1.80 1490 A 2.00 154 C 2.00 260 A Urban Cluster-l .05 2700 C .38 268 A .88 9 B .88 90 C 1.17 268 C 1.25 7 X 1.38 230 A 1.63 225 C 1.65 183 D 1.90 121 C 2.00 235 B 2.00 302 D Urban Cluster-2 .25 2730 D .40 81 A 1.23 40 A 1.40 255 C 1.50 248 D 1.55 157 A 1.65 287 A 1.60 279 C 1.65 253 B 1.75 204 A 1.75 330 A 1.80 208 C 1.85 225 C 1.90 210 X 2.00 185 D 2.00 302 A 0 Urban Cluster-3 .75 162 A .75 272 A .80 150 B 1.25 182 B 1.30 300 B 1.50 178 A pi.» A E; g» __._ EEEI I55! I!!! ggg! 212 APPENDIX G. Continued Distance Chain Survey Store (Miles) Degree Identi- fication Urban Cluster-3 (Continued) 1.60 750 X 1.65 145 X 1.65 80 A 1.75 208 A 2.00 110 B 2.00 112 A Small Town—1 .15 90° A Small Town-2 .20 900 A Small Town-3 - - Neighborhood Shopping Center-1 1.05 00 D 1.20 58 A 1.22 60 B 2.00 237 C Neighborhood Shopping Center-2 .20 270° D 1.20 37 B 1.25 314 B 1.25 212 A 1.40 40 D 1.50 46 A Community Shopping Center—1 .25 2700 A 1.55 122 B 1.55 132 X 1.58 270 B 1.65 230 D 1.60 186 D 2.00 83 C Community Shopping Center-2 .20 2680 A .40 260 D .60 255 B 1.25 215 D 1.50 56 C 1.65 131 B 1.90 297 A 213 APPENDIX G. Continued Distance Chain Survey Store (Miles) Degree Identi- fication Community Shopping Center-2 ‘1,95 242° c (Continued) 2.00 241 A 2.00 298 B Regional Shopping Center-1 .45 2000 B 1.25 185 X 1.30 195 C 1.35 202 A 1.40 305 C 1.42 296 A 1.42 151 C 1.75 222 C 2.00 262 D Regional Shopping Center—2 .80 2800 D 1.05 78 B 1.40 340 B 1.45 274 B 1.65 190 B 1.75 135 C 1.75 150 A Ilyfl I .l-=-l l I l I”! I I Iw I I I 214 BIBLIOGRAPHY Public Documents Chenoyer, Helen G. Selecting a Store Location. washington, D.C.: Bureau of Foreign and Domestic Economic Series, No. 56, 1946. Hansen, Lawrence A., Measuring A Retail Market. U. S. Depart- ment of Commerce, Trade Information Bulletin No. 272. Washington: Government Printing Office, 1924. "How American Buying Habits Change," U. S. Department of Labor, Washington: U. S. Government Printing Office, 1959. Miller, Nelson A. Selecting a Location: Grocery Store. Wash- ington, D.C.: U. S. Government Printing Office, 1947. Rolph, Inez K. Location Structure of Retail Trade. U. S. De— partment of Commerce, Domestic Commerce Series No. 80. Washington: U. S. Government Printing Office, 1933. "Sampling Methods For A Small Household Survey," Public Health Monograph No. 40, U. S. Department of Health, Education and welfare, Washington: U. S. Government Printing Office, 1956. Staff Report to the Federal Trade Commission, Economic Inquiry into Food Marketing: Part I Concentration and Integration in Retailing, Washington: U. S. Government Printing Office, 1960. Books Alderson, Wroe. Marketing Behavior and Executive Action. Home- wood, 111.: Richard D. Irwin, Inc., 1957. Baker, Geoffrey and Bruno Funaro. Shopping Centers. New York: Reinhold Publishing Co., 1951. Barlowe, Raleigh. Land Resource Economics, New YOrk: Prentice Hall, 1958. 2555 F! 215 Brink, Edward L. "A Simplified Solution for the Location of Marketing Outlets Using an Analog Computer," Marketing}s Role in Scientific Management. Chicago: American Mar- keting AssoCiation, 1957,.pp. 468-75. Brown, Lyndon O. Marketing_and Distribution Research. 2nd edition. New York: Ronald Press, 1949. Chamberlin, Edward H. The Theory of Monopolistic Competition, 7th ed. Cambridge: Harvard University Press, 1956 (Origi— nal 1933). Clark, J. M. Studies in the Economics of Overhead Costs, Chicago: University of Chicago Press, 1923. Cohen, Saul B. (ed.)’ Store Location Research For The Food Industry. New YOrk: National-American Wholesale Grocers Association, 1961. Converse, Paul D., Huegy, Harvey W. and Mitchell, Robert V. Elements of Marketing. 6th edition. New YOrk: Prentice Hall, 1958. Dobriner, William M. The Suburban Community. New York: G. P. Putnam's Sons, 1958. Epstein, Bart J. and Green, Heward J. "Store Location Analysis," Marketing Research in Action, Studies in Business Policy,' No. 84, National Industrial Conference Board, 1957, pp. 85-87. "Factors in A Purchase Decision," Convenience Goodg Purchasing: Needed Research. Ann Arbor: The Foundation for Research on Human Behavior, 1957. "Facts About New Supermarkets Opened in 1958," Supermarket Institute, 1959. (Reprint) "Facts in Grocery Distribution," Progressive Grocer, April, 1960, p. F7. Fine, Isodore V. Retail Trade Analysis, Bureau of Business Re- search and Service, Madison: University of Wisconsin, 1954. Fisher, Heward T. "Shopping Centers Lure Customers Away from Downtown Districts," American Business, XXII, No. 3 (March, 1952). pp. 16-17. rI III IIIIIIII I I I I I I III- 216 Fisher, walter D. "Economic Aggregation as a Minimum Distance Problem," Econometrica, Vol. 25, No. 3. (Abst. p. 363.) Fitzpatrick, P. "What Factors Make a Successful Store Location?" Journal of Retailing, October, 1933, pp. 90-92. Flint, Lucius S. "How Chains Are Selecting Locations," Chain StOre Age, Vol. XXII (March, 1946), p. 276. ”Fort werth Department Store Abandons Downtown for the Suburbs," Business week, No. 1405 (August 4, 1956), p. 83 Gibbs, M. G. "HOW a Prominent Chain Picks Store Locations," Printers' Ink, Vol. CXLI (November 10, 1927), pp. 108-109. Goloff, Joseff M. "What Customers Think about Branch Stores," Journal of Retailing,0ctober, 1948, pp. 103-108.' Green, Robert F. "HOw a Smaller Chain Analyses Store Locations," Chain Store Age, July, 1959, pp. 33-35. Greenhut, Melvin L. "Space and Economic Theory," Regional Science AssociationLWPapers and Proceedings, Vol. 5, 1959. pp. 267-280. Greenhut, Melvin L. Plant Location. Chapel Hill: University of North Carolina Press, 1956. Gruen, Victor and Smith, Larry. Shgpping_Towns U.S.A.‘New York: Reinhold Publishing, 1960. Heward, John A. Marketing Management. Homewood, 111.: Richard D. Irwin, Inc., 1957. Holdren, Bob R. The Structure of a Retail Market. Englewood Cliffs: Prentice Hall, Inc., 1960. Hoover, Edgar M. Location of Economic Activity. New York: McGraw-Hill Book Company, 1948. Isard, walter. Location and Space Economy. New York: John Wiley & Sons, 1956. Isard, Walter. et. a1. Methog§_of Regional Analysis. New Yerk: John Wiley & Sons, 1960. Jonnassen, C. T. Downtown vs, The Suburbs. Columbus: Bureau of Business Research, Ohio State University, 1955. 217 Kelley, Eugene J. ShoppingACenters: Locating_Controlled Re- gional Centers. Saugatuck: The Eno Foundation for High- way Traffic Control, 1956. Losch, August. The Economics of Location, Translated by welf- gang F. Stolper, 2nd ed. revised. New Haven: Yale Univer— sity Press, 1954. Lundy, Robert D. "How Many Service Stations Are 'Too Many'?" Reavis Cox and wroe Alderson, (eds.), Theory in Marketing, Chicago: R. D. Irwin, Inc., 1950. Maynard, Harold H., Beckman, Theodore H., and Davidson, William R. Principles of Marketing, 5th ed. New York: Ronald Press, 1957. McCarthy, E. Jerome. Basic Marketing: A Managerial Approach, Chicago: R. D. Irwin, Inc., 1960. Mossman, Frank H. and Newton Morton. Principles of Transporta- tion. New York: Ronald Press Co., 1957. Nelson, Richard L. The Selection of Retail Locations. New York: F. W. Dodge Corp., 1958. Proudfoot, Malcolm J. The Major OutlyinggBusiness Centers of Chicago. Chicago: University of Chicago, 1938. Ratcliff, Richard U. "The Problem of Retail Site Selection'" Michigan Business Studies. Ann Arbor: University of Michigan, Vol. 9, No. 1, 1939. Selltiz, Claire, Marie Johoda, Morton Deutsch, and Stuart W. Cook. Research Methodggin Social Relations. New Yerk: Henry Holt & Co., Inc. Smith, Paul E. Shgpping Centers: Planning and Management. New York: The National Retail Dry Goods Association, 1956. Troxel, Emery. Economicgyof Transport. New York: Rinehart, 1955. wehrly, Max S. and McKeever, J. Ross, (eds.). The Community Builders} Handbook. Washington: Urban Land Institute, 1960. Walker, Helen M. and Joseph Lev. Statistical Inference. New York: Henry Holt and Co., 1953. gnu-Il-IH-‘ _ r..__. EH: 218 Wingate, John W., and Corbin, Arnold. Changing Patterns in Retailing. Chicago: R. D. Irwin, Inc., 1956. Articlee and Periodicals "A New Dimension: Economic and Marketing Geography," Food Topics, March, 1961. pp. 6-10. Applebaum, William and Cohen, Saul B. "Store Location Strategy in a Changing Market," Proceedings of the 1961 Mid—Year Conference, Chicago: The Supermarket Institute, 1961. Applebaum, William and Cohen, Saul B. "The Dynamics of Store Trading Areas and Market Equilibrium," "The Annals of the American Aesociation of Geographers. Vol. 51, No. 1 (March, 1961) pp. 73-101. Applebaum, William and Spears, Richard F. "How to Measure a Trading Area," Chain Store Age, January, 1951. (reprint) Applebaum, William. "Evaluating Store Sites and Determining Store Rents," Chain Store Age, Vol. XXXIV (March 5, 1958), pp. 15.-17. Applebaum, William. "Studying Customer Behavior in Retail Stores," Journal of Marketing, October, 1951, pp. 172-178. Aspinwall, Leo. "The Functional Organization of Cities," Cost and Profit Outlook, Vol. IX (July, 1956), pp. 1-3. Balderston, A. J. "Assortment Choice in Wholesale and Retail Marketing," Journal of Marketing, Vol. XXI (October, 1956), pp. 175-183. Baumol, W. J. and Ide, E. A. "Variety in Retailing," Management Science, Vol. 3, No. 1 (October, 1956). Bennett, Victor W. "Consumer Buying Habits in a Small Town Located Between Two Large Cities," The Journal of Marketing, VIII, No. 4 (April, 1944), 405-416. Bliss, Perry. "Supply Considerations and Shopper Convenience," Journal of Marketing, Vol. 25, No. 1 (July, 1960), pp. 43145. 219 Blankertz, D. F. "Customer Patronage of a Parent and New Branch Store," The Journal of Marketing, XVI, No.2 Bowden, William K., and Cassady, Ralph, Jr. "Decentralization of Retail Trade in the Metropolitan Market Area," The Journal of Marketing, No. 3 (January, 1941), pp. 270-275. Bratcher, Austin S. "A Method of Delineating Retail Trade Zones," Journal of Marketing, January, 1939, pp. 262-264. Bratcher, Austin S. A Method of Delineating Retail Trade Zones. University of Arizona, School of Business Research, Study No. 5. Tucson: University of Arizona, 1939. Breyer, Ralph F. "Some Preliminary Problems of Sample Design for a Survey of Retail Trade Flow," Journal of Marketing. Vol. 10, No. 4 (April, 1946), pp. 343—353. Carroll, J. D., Jr. "Defining Urban Trade Areas," Traffic Quarterly. April, 1955. pp. 149—161. Casey, Harry J., Jr. "The Broadening Perspective of Marketing,‘ American MarketingeAeeociation, Chicago, 1956. Casey, Harry J., Jr. "The Law of Retail Gravitation Applied to Traffic Engineering," Traffic Quarterly. July, 1955. pp. 313-321. Cassady, Ralph, Jr., and Bowden, W. K. "Shifting Retail Trade Within the Los Angeles Metropolitan Market," The Journal of Marketing, VIII, No. 4 (April, 1944), pp. 398-404. "Centers Open on Target in '60'," Chain Store Age, January, 1961, P. E26. "Chains Reveal Rules-of-Thumb for Choosing Store Locations," 'Chain Store Age, January, 1960, pp. E33—E38. Cohen, Saul B. and Applebaum, William. "Evaluating Store Sites and Determining Store Rents," Economic Geography, Vol. 36, No. 1 (January, 1960). pp. 1-35. Coles, Jessie V. "The Pattern of Retail Food Stores in a Small City," The Journal of Marketing, XIII, No. 2 (October, 1948), pp. 163-171. — FE. I I I H.,: .v-] .. ] —___:=l.=::l 220 Comish, Newel. "What Influences Customer Choice of a Food Store," Journal of Retailing; (Summer, 1958), pp. 90-99. Converse, P. D. A Study of Retail Trade Areas in East Central Illinois. Bureau of Business Research, Business Study No. 2. Urbana: University of Illinois, 1943. Converse, P. D. "Is There A Law of Retail Gravitation?" Printers Ink, Vol. 204 (Sept. 10, 1943) p. 36. Converse, P. D. "New Laws of Retail Gravitation," The Journal of Marketing, XIV, No. 3 (October, 1949), pp. 379-384. Converse, P. D. Retail Trade Areas in Illinois. Bureau of Business Research Business Study No. 4. Urbana: University of Illinois, 1946. Cox, Reavis. "Consumer Convenience and The Retail Structure of Cities," Journal of Marketing, Vol. 23, No. 4 (April, 1959) pp. 355-362. Curran, Frank B. and Stegmaier, Joseph T. "Travel Patterns in 50 Cities," Bulletin 203, Highway Research Board, Washington: National Academy of Sciences, 1959. "Department Store Branches in Suburbs Succeed, Multiply," Business week, No. 56 (October 1, 1930), pp. 10-11. "Department Stores Hurrying to suburbs,” Business week, No. 944 (October 4, 1947), pp. 24-26. Doherty, Richard P. "Decentralization of Retail Trade in Boston," The Journal of Marketing] VI, No. 3 (January, 1942), pp. 281-86. Doherty, Richard P. "The Movement and Concentration of Retail Trade in Metropolitan Areas," The Journal of Marketing, No. 4 (April, 1941), pp. 395-401. Doherty, Richard, and Cox, Stuart. An Analysis of Retail Trade and Consumer Buying Habits in the Boston Retail Market. Bureau of Business Research. Boston: Boston University, 1932. Douglas, Edna. "Measuring the General Retail Trading Area--- A Case Study: 11," The Journal of Marketing, XIV, No. 1 (July, 1949), 46-60. Eberle, George J. "Metropolitan Decentralization and The Re— tailer," Journal of Retailing, XXII, No. 4 (December, 1946), pp. 9l-94. u (I. In. I I? l llvllllrfllr I'LII IF I, IIrL 2:23 2:3 22:! Ell EEE! Illl !!I IIII III' lill III' IIII I—.-_=:I.::=:J::l::3 221 Ellwood, Leon W. "Estimating Potential Volume of Proposed Shopping Centers," The Appraisal Journal, Vol. 22, No. 4 (October, 1954). pp. 581-589. Harb, Ray 0. "Why Three Out of Four Stores Should Consider A New Location," Progressive Grocer, Vol. XXXII (August, 1953), pp. 48-52. Huff, David L. "A Topological Model of Consumer Space Pre- ferences," Occasional Paper No. 11, Bureau of Business Research, Seattle: University of Washington, 1959. Isard, Walter. "The General Theory of Location and Space Econ- omy," Quarterly Journal of Economics, Vol. XLIII (NOvember, 1949). pp. 476-506. ' Jonassen, C. T. "Shopper Attitudes," Special Report ll-A, Highway Research Board, Washington: National Academy of Sciences, 1955. Jones, Fred M. A Survey of a Retail Trading Area. Bureau of Business Research, Bulletin No. 44. Urbana: University of Illinois, 1932. Jung, Allen F. "Is Reilly's Law of Retail Gravitation Always True?" Journal of Marketing, Vol. 24, No. 2 (October, 1956) pp. 62-63. Kelley, Eugene J. "The Importance of Convenience in Consumer Purchasing," Journal of Marketing, July, 1958, pp. 23-38. Marble, Duane F. "Transport Inputs at Urban Residential Sites," The Regional Science Association--Papers and Proceedings, Vol. 5, 1959. pp. 253-266. "Marketing Definitions: A Glossary of Marketing Terms," Ameri- can Marketing Association, Chicago, 1960. McCarthy, L. J. "Development of Trading Areas," National Marketing Review. 1935, pp. 80-82. Metropolitanization of the United States. Research Monograph No. 2.Washington: Urban Land Institute, 1959. Muller, Robert W. "1959 Grocery Sales up 4.2% as Total Tops $50 Billion," Progressive Grocer, April, 1960, p. 2 (reprint). 222 Muller, Robert W. "What's the Best Location for Today's New Store?" Progressive Grocer, Vol. XXXII (February, 1953), p. 50. Nahl, Perham C. Application of the Interview Method to Trading Area Survey of Stillwatery Oklahoma. Stillwater: Oklahoma A & M College, 1939. 1960 Greater Milwaukee Shopping Habits, The Journal Co., 1960. "Operational Facts About New Supermarkets Opened In 1960," Pro- gressive Grocer, April, 1961. pp. 120-128. Proudfoot, Malcolm J. "The Selection of a Business Site," Journal of Land and Public Utilities Economics, NOvember 14, 1938, p. 372. Predohl, Andreas. ”The Theory of Location in Its Relation to General Economics," Journal of Political Economy, Vol. XXXVI (June, 1928), pp. 371-389. "Profile of Supermarket Customers," Super Value Study, Progres- sive Grocer, August, 1958. pp. 8106-111. Pyle, J. F. "Determination of Location for Retail Concerns," Harvard Business Review, Vol. IV (April, 1926), pp. 303-312. Ratcliff, Richard V. "100 Per Cent Concept in Retail Location Analysis," Journal of Land and Public Utility Economics, Vol. xv (August, 1939), pp. 350-353. Ratcliff, Richard V. "Some Principles of Site Selection in Outlaying Retail Subcenters," Journal of Marketing (Fall), 1935. Reed, Virgil D. "Some Statistical Possibilities of Defining Market Limits," Journal of Marketing. July, 1938, pp. 39-43. Rester, Leon and Harold H. Kantor. Optimal Location Allocation. Sixth Annual Meeting of the Operations Research Society of America, May 16, 1958. Rouse, James W. "Estimating Productivity for Planned Regional Shopping Centers," Urban Land, Vol. 12 No. 10 (November, 1953.) pp. 1-6. Samuelson, Paul A. "Spatial Price Equilibrium and Linear Pro- gramming," American Economic Review, Vol; 42, 1952. pp. 283—303. nu: _::: :2: .::3 :::1 L::3 2:2] 2:: 22:] 2:3 I!!! 253! .251 ISII Ell llll [Ell IEjl llll 223 "Shopping Centers: Checking the Location Map," Chain Store Age, Vol. XXXIV (May, 1958). pp. 19-20. "Shoppers Buy in Two Stores or More," Supermarket Merchandising, February, 1958, p. 122. Smithies, Arthur. "Optimum Location in Spatial Competition," Journal of Political Economy, Vol. XLIX (1941), pp. 423-439. "Super Value Study," Prggressive Grocer, August, 1958, p. S-1l (reprint). "The Importance of General Merchandise," (The Dillon Study: Part 5), Progressive Grocer, August, 1960, pp. 57-69. "The True Look of the Supermarket Industry, 1959,“ Supermarket Merchandising, May, 1960. (reprint) Sharpe, Gordon B. "Travel to Commercial Centers of the Washington, D.C. Metropolitan Area," Bulletin 79, Highway Research Board, Washington: National Academy of Sciences, 1953. Reilly, William J. "Method for the Study of Retail Relationships," Research Monograph No. 4, University of Texas Bulletin No. 2944, Austin: University of Texas Press, 1929. Staniford, Edward F. Business Decentralization in Metropolitan Los Angeles. Los Angeles: Bureau of Government Research, University of California, 1960. Steadman, Gordon H. "The Rise of Shopping Centers," Journal of Retailing, Vol. XXXI, No. 1. (Spring, 1955. pp. 14415). Valavanis, Stefan. "Losch on Location," American Economic Re- view, Vol. XLV (1955), pp. 637-644. Vidale, Marcello L. "A Graphical Solution of the Transportation Problem," Operations Research, Vol. 4, No. 2 (April, 1956L pp. 193-203. Voorhees, Alan M., Gordon B. Shorpe, and J. T. Stegmaier. Shop- ping Habits and Travel Patterngp Technical Bulletin No. 24, Urban Land Institute, Washington, D.C., March, 1955. weiss, E. E. Highway Retailing-—The Next Great Retail Revolution. New York: Doyle--Dane-—Dernbach, Inc., 1958. iii I::I E21 IEEI {2:3 IEEI E33 ICE! :23 3:23 EEEI :2: .::3 ;::3 _::j :‘t! i223 —I_'.:= 224 "What Will It Cost To Open A New Supermarket in 1961?", Chain Store Age, December, 1960, pp. 38-39. wo1ff, Reinhold P. "Estimating the Market Potential of a Float— ing Population," The Journal of Marketing, XIX, No. 1 (July, 1954). pp. 12-17. Young, George A., Jr. "Do's and Don'ts in Selecting Supermarket Locations," Chain Store Age, Vol. XV (February, 1939), p. 29. Zimmerman, M. M. "The Supermarket and the Changing Retail Struc- ture," The Journal of Marketing, V, No. 4 (April, 1941), Unpublished Material Barnewall, Gordon Gouverneur. "Factors Affecting the Optimum Location and Layout of Planned Suburban Business Districts." Unpublished Ph. D. Dissertation, The Ohio State University, 1953. "Basic Report on Consumer Behavior," Philadelphia: Alderson and Sessions. (Mimeographed). Epstein, Bart J. "The Quincy Food Market: A Study in Marketing Geography," Unpublished Doctoral Dissertation, Clark Univer- sity, 1956. Fine, Isodore V. "Retail Trade Area Measurement Techniques as Applied to Fort Atkinson, Baraboo and west Bendf'Unpublished Doctoral Dissertation, Columbia University, 1953. Kelley, Eugene J. "Location of Controlled Regional Shopping Centers." Unpublished Ph. D. Dissertation, Graduate School of Business Administration, New York University, 1955. Oaks, Clinton Lewis. "A Critical Analysis of Selected Problems in the Development, Organization, and Control of Large Sub- urban Branches of Pacific Coast Department Stores," Unpub- lished Ph. D. Dissertation, Graduate School of Business, Stanford University, 1955. If?! I!!! Iii! E23 E!!! F’S I!!! “7T3 iffj 2:15 7:23 T223 _;_3 _'_'::J ___J ‘34:] 225 Robinson, James George. "Retail Decentralization in the Los Angeles Metropolitan Area and Some Attendent Problems with Special Emphasis Given to Selected Downtown Department and Specialty Stores." Unpublished Ph. D. Dissertation, Grad- uate School of Business Administration, New York University, 1951. Other Sources Interview with Dr. Saul B. Cohen, Boston University, January 18, 1961. I I | I I 5 I II I I I I I I I I I A . LIBRRRIES IGRN S V (Ill IllllllWllIlIllllllllMlllll 1293007536349 TQTE UNI HICH