AN INVESTIGATION OE THE PRODUCT LIFE CYCLE CONCEPT AND ITS APPLICATION TO NEW PRODUCT PROPOSAL EVALUATION WITHIN THE CHEMICAL INDUSTRY Thesis for the Degree of Ph. D. MICHIGAN STATE UNIVERSITY MARTIN S. FREDERIXON 1969 o—“M 'h‘4;-. _ LIBRARY "' Michigan State University This is to certify that the thesis entitled AN INVESTIGATION OF THE PRODUCT LIFE CYCLE CONCEPT AND ITS APPLICATION TO NEW PRODUCT PROPOSAL EVALUATION WITHIN THE CHEMICAL INDUSTRY presented by MarTin S. Frederixon has been accepted towhrds fulfillment of the requirements for Ph.D. degree in Marketing—Finance (5542/ gZWLZé ’ Mefdr professor / - , 93%? /, Mg 0-169 ABSTRACT AN INVESTIGATION OF THE PRODUCT LIFE CYCLE CONCEPT AND ITS APPLICATION TO NEW PRODUCT PROPOSAL EVALUATION WITHIN THE CHEMICAL INDUSTRY By Martin S. Frederixon This quantitative investigation of industrial chemical product life cycles and their relationship to both performance and new product evaluation methodology has the major goals of: 1) classifying product life cycles by patterns on sales, profits, and related financial data for new industrial chemical products, 2) identifying those structural characteristics of new industrial chemicals which relate to performance, and 3) screening historical data of new industrial chemical product histories in order to formulate predictive multivariate models. This study specifically seeks to integrate product life cycle and capital budgeting theories. Documenting the structural characteristics of new industrial chemical products certainly promotes our fundamental understanding of product behavior; but it also sets the limits within which we may generalize from the experiences of major chemical marketers over the 1955-1964 period studied. An effective new product program demands high resolution of all relevant variables affecting the investment decision. And this research has confirmed the feasibility of using established structural characteristics as inputs in a statistical capital budgeting model for evaluating new product prOposals, thus facilitating a more optimum investment choice among a complex set of simultaneous alternatives . Martin S . Frederixon A detailed analysis of 27 new product offerings randomly drawn from among the new industrial chemical products introduced by participating firms between 1955 and 1960 revealed three basic shapes of sales patterns (five if we add the time dimension). These showed no measurable statistical differences on all performance variables except the growth rate for sales and the payback period. The overall shape of the sales cycle for a firm typically resembled that of the industry consumption pattern. This evidence runs counter to the premise that the profit cycle begins to descend while the sales curve is still rising during the maturity phase of the product life cycle: over fifty percent of the products studied had coincident or lagging profit life cycle structures. Known patterns of product behavior suggest a time horizon of at least eleven years for more accurately evaluating the experiences of new industrial chemical products, including an additional three years required for planning plant and equipment comitments to cover the time gap between authorizing capital expenditures and bringing facilities on-stream. Simple relationships were sought between performance and possible correlates of performance—factors associated with market structure, buyer behavior, product characteristics, and related intrafirm experiences. These factors were screened by appropriate statistical tests to reduce the number of possible determinants of product behavior. It was found that performance generally related to derived demand patterns, duplication difficulties by competitors, dependence on field coverage, impact of advertising on source selection, and product loyalty. The tests also generally supported relationships between performance and investment requirements, research and deve10pment expenditures, export patterns, Production scheduling experiences, aggregate marketing costs, and plant capacity utilization. Performance was not shown to have an association With a number of variables, including import patterns, patent protection, technological innovation, buyer purchasing patterns, merger activities, tI’Pe of distribution channel, marketing development approach, type of Product, and source of product discovery. . Mg. -. . ‘. . A“ ‘ . . ’ u. . ‘ . n s ‘- -u. . A-“ __ .,. ‘ u . ‘ u 4 .. _ .. . , ‘5.- ' I s ’ v... . _ . ._ . U“- ‘- . ' "r- Martin S. Frederixon Multivariate statistical models of product behavior were found to be tenable. For example, the number of years required for discounted present value sums to shift from negative to positive values depended on the following variables (all contributing individually to variance reduction): aggregate research and development commitment, number of minor consuming industries, promotional outlays, effect of industrial advertising on manufacturer selection, trade relations, product loyalty, customer backward integration, buyer acceptance of the product concept, technical service requirements, export patterns, and the orientation of the research and development program. But variable definition, data collection, and scoring must be further refined before improvements in predictions can be expected. AN INVESTIGATION OF THE PRODUCT LIFE CYCLE CONCEPT AND ITS APPLICATION TO NEW PRODUCT PROPOSAL EVALUATION WITHIN THE CHEMICAL INDUSTRY By Martin S. Frederixon A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Impartment of Accounting and Financial Administration Department of Marketing and Transportation Administration. 1969 " Copyright by MARTIN SHELTON FREDERIXON 1970 44.. __,-—v—-"“ AC KNOWLE DGMENTS In surveying possible dissertation subjects, I noticed a general dearth of any integrated approach to screening and evaluating new product marketing opportunities. Specifically, the literature revealed little empirical evidence on how resources were allocated for new industrial product development activities and what determinants of demand directly influence new product decisions. So I sought to define and structure new product behavior rigorously, first utilizing the product life cycle concept as a basis for financial and marketing measurement because this concept actively focuses on those strategies which may thrust a firm into a position of profitable leadership. Having limited this study to the industrial experiences of large chemical manufacturers, I investigated a number of product histories with the assistance of many individual company representatives: to all of those I owe my sincere appreciation. I owe much to the direction and stimulus provided by my dissertation chairman, Professor Adolph E. Grunewald, and the other members of my committee, Professors William J. E. Crissy and Thomas A. Staudt; their allowing me the freedom to test many concepts in action forced a more penetrating examination of the fundamental sources of product responses. Bruce Tracy reviewed drafts of the manuscript and offered many helpful suggestions and constructive comments. The manuscript was typed by Betty Hendricksen and Geri Galloner with utmost competence and patience. lam indebted to the ideas and encouragement given me by Professor Robert Headen and Edward Nepkie. Much of my motivation and enthusiasm in this endeavor, furthermore, was derived from my parents and the balanced family environment they provided, emphasizing individual deveIOpment and value formation. Needless to say, the present work required innumerable sacrifices from my family. I owe sincere appreciation especially to my wife, Nancy, whose thoughtfulness, understanding and patience have made my educational pursuits truly worthwhile. ii TABLE OF CONTENTS Page nmnmnmoomnns . . . . . . . ii LIST OF FIGURES. . . . . . . xii LIST OF TABLES . . . . . iriv LIST OF APPENDICES . . xxv Chapter I INTRODUCTION General Background . . . . . . . . . . . . . . . . 1 Research and DeveIOpment Expenditure Patterns . 1 Emphasis on New Product DeveIOpment as an Industrial Marketing Strategy . . . . . . . . . 6 Basic Structure of the New Product Evaluation Process . . . . . . . . . . . . . . . . . . . . . . 11 Identification of the Problem . . . . . . . . . . . . 14 Significance of the Study . . . . . . . . . . . . 16 Need for the Study . . . . . . . . . . . . . . . . . 17 II REVIEW OF SALIENT LITERATURE ON PRODUCT LIFE CYCLE THEORY Introduction . . . . . . . . . . . . . . . . . . 18 Linear Function . . . . . . . . . . . . . . . . . 18 Exponential and Logarithmic Functions . . . . . . . . l9 Gompertz Curve . . o o o o o o o O o o o o o o o o o 22 Conventional Product Life Cycle . . . . . . . . . . . 23 Possible Applications of Product Life Cycle Theory 28 III REVIEW OF SALIENT FINANCIAL CONCEPTS IN EVALUATING NEW PRODUCT PROPOSALS Introduction . . . . . . . . . . . . . . . . . . . 31 Increasing Emphasis on Capital Budgeting . . . . 31 Costs and Investments Considered . . . . . . . . . 32 Depreciation Policies Considered . . . . . 34 Salient Financial Concepts . . . . . . . . . . . . . 35 1. Cost of Capital . . . . . . . . . 35 A. Cost of Current Liabilities B. Cost of Long Term Debt C. Cost of Equity Capital . . . . . . . . 37 . . . . . . . . 37 2. Payback Period . . . . . . . . . . . . . . . 3g 3. Accounting Rate of Return . . . . . . . . . 41 4. Return on Investment . . . . . . . . . . . . 42 5 Internal Rate of Return . Chapter III . Present Worth Method . . . . . . . . . . Equivalent Rate of Return . Profitability Ratio . . . . Performance Index . . \OQNQ I O O O O O O O Sumry o o o o o 0 O O O O I O C I O O O C IV METHOD OF DATA COLLECTION Selection of Industry . . . Definition of New Product . Selection of Pepulation . . . . . . . . . . . . . . Performance Record of Major Chemical Companies V FUNDAMENTAL POSTULATES BEHIND THE RESEARCH EFFORT Postulate No. 1: New products are basic to the growth of any firm . . . . . . . . . . . . . . Postulate No. 2: The projections of demand estimates incorporated in new product evaluation proposals are the most uncertain inputs, and their uncertainty increases with time . . . . . . Postulate No. 3: The cost of capital concept aids in effectively evaluating the degree of financial uncertainty associated with new product decisions VI POSSIBLE INFLUENCES DETERMINING NEW PRODUCT BEHAVIOR General Hypothesis No. 1: No single representative nth order polynomial function best describes the sales patterns of new industrial chemical products General Hypothesis No. 2: The profit cycle for new industrial chemical products does not typically fit a declining exponential curve . . . . . . . . General Hypothesis No. 3: The profit cycle of new industrial chemical products does not typically descend while the sales curve is still rising in the maturity phase of the product life cycle . . General Premise No. l: The time horizon necessary to consider sales and profit contributions for new industrial chemical products exceeds five years beyond product introduction . . . . . . . Classification Schemes Used to Describe Performance Identified . . . . . . . . . . . Possible Factors Relating to the Performance of New Industrial Chemical Products . . . . . . . 1. Market Structure . . . . . . . . . . . . . A. Degree of Patent Protection . . . B. Demand Trends in Derived Demand Situations . . . . . . . . Page 45 47 49 SO 51 52 52 56 S8 75 8O 83 85 86 87 87 88 89 9O 9O 90 Chapter VI FOWMU KHH Duplication Difficulties by Competitors . . . . . . . . . . . . Extent of Capacity Utilization at Industry Level . . . . . . . . . . Import Patterns . . . . . . . . . . . Market Share . . . . . . . . . . Market Trends . . . . . . . . Minimum Corporate Asset Size of Competitors Required to Compete Effectively . . . . . . . . . . . . Number of Consuming Industries . . . Number of Significant Competitors . Supply Characteristics of Factors . Behavior . . . . . . . . . . . . . . Degree of Backward Integration . . . Degree of Required Deliberation . . . Dependence on Field Contact Work . Effect of Industrial Advertising on Source Selection . . . . . . . . Effect of Product Quality on Source Selection . . . . . . . . . . . . . Effect on Sales of Related Products . Extent of User Laboratory Evaluation Level of Product Loyalty . . . . . . Number of Contacts Required by Marketing and Technical Personnel Number of Product Sources . . . . . Number of Purchasers . . . . . . . . Number of Annual Purchases by Buyers Recognition Experiences of Product Advantages by Users . . . . . . . Reputation and Image of Manufacturer Time to Educate the User . . . . . . Trade Relations With Users . . . . uct Characteristics . . . . . . . . . CYC1ical Patterns . . . . . . . . . . Degree of Marketing Innovativeness Degree of Matching Between Technological Characteristics and Market Require- ments . . . . . . . . . . . . . . . End-Use Patterns . . . . . . . . Level of Technological Innovativeness Price Movements . . . . . . . . . . Product Differentiation Strategy . - Research and Development Harnessing Experiences . . . . . . . . . . . . Seasonal Patterns . . . . . . . . . . Specificity of Use . . . . . . . . Standard Industrial Classification (SIC) Code . . . . . . . . . . . . ‘1 Page 90 9O 91 91 91 91 92 92 92 92 92 93 93 93 93 94 94 94 94 94 95 95 95 95 95 96 96 96 96 96 97 97 97 97 97 97 98 98 _I .‘v Chapter VI WI Page L. Technical Service Requirement . . . . . 98 M. Trends in Cross Margins . . . . . . . . 98 N. Type of Product . . . . . . . . . . . . 98 0. Type of Product Demand . . . . . . . . 99 4. Related Intrefirm Experiences . . . . . . . . 99 A. Effectiveness of Product Flows in the Nays Distribution Channel . . . . . . . . 99 Export Patterns . . . . . . . . . . . . 99 Extent of Plant Utilization at Firm Level . . . . . . . . . . . . . . . . 99 Intensity of the Selling Effort . . . . 99 Investment Patterns . . . . . . . . . . 100 Length and Number of Production Runs 100 Level of Clarity of Product Demand . . 100 Level of Research and Development Expenditures . . . . . . . . Licensing Experiences . . . . . . . . Management Evaluation of Relative Success . . . . . . . . . . . . . 101 100 100 Marketing Costs . . . . . . . . . . . . 101 Merger Activities . . . . . . . . . . . 101 Mode of Production . . . . . . . . . 101 Orientation of Research and DeveIOp- ment Program . . . . . . . . . . . . 102 Output of Research Activities . . . . . 102 Product Concept Acceptance . . . . . . 102 Product Improvement Efforts . . . . . . 102 Promotional Media Strategy . . . . . . 103 Promotional Outlay Trends . . . . . . . 103 Source of Product Discovery . . . . . . 103 Suitability of Marketing Capabilities . 103 Supply Capabilities in DeveIOpment Sampling Programs . . . . . . . . . . 104 Technological Specialty Experiences . . 104 Type of Distribution Channel Used . . . 104 Type of Fixed Capital Employed . . . . 105 Type of Marketing Representation Used . 105 METHODS EMPLOYED IN DATA ANALYSIS Introduction . . . . . . . . . . . . . . . . . . . . 106 P"lymmiel Determination . . . . . . . . . . . 106 An Approach to New Product Evaluation . . . . . . . . 109 PTOgrem Analysis . . . . . . . . . . . . . . . . . . 111 Parametric Versus Nonparametric Methods . . 112 R°1m°80rov-Smirnov Two Sample Test . . . . 114 Kruskal-Wellis One-Way Analysis of Variance . . . . . Kiefer K-Semple Analogue Test . . . . . . . . . . . . 115 ah'Pkar KsSample Analogue Test . . . . . . . . . . . 115 115 SPeermen Rank Correlation Coefficient . . . . . . . 116 Mhltiple Regression . . . . . . . . . . . . . . . . . ll6 Linear Programming Regression . . . . . . . . . . . . 118 vi Chapter VIII RESEARCH FINDINGS Introduction . . . . . . . . . . . . . . . . . . . . General Hypothesis No. 1: No single representative nth order polynomial function best describes the sales patterns of new industrial chemical products General Hypothesis No. 2: The profit cycle for new industrial chemical products does not typically fit a declining exponential curve . . . General Hypothesis No. 3: The profit cycle of new industrial chemical products does not typically descend while the sales curve is still rising in the maturity phase of the product life cycle . . General Premise No. l: The time horizon necessary to consider sales and profit contributions for new industrial chemical products exceeds five years beyond their introduction . . . . . . . . . . . . . Discussion of Performance Variable Outcomes . . . . . 1. Sales Structure . . . . . . . . . . . Length of Sales Pattern . . . . . . . . . . . Profit Structure . . . . . . . . . . . . . . Growth Characteristics . . . . . . . . . . . Liquidity . . . . . . . . . . . . . . . . . . Return Structure . . . . . . . . . . . . . . . Present Value Structure . . . . . . . . . . . . Relative Contribution to Wealth . . . . . . . An Analysis of Possible Determinants of Product Performance . . . . . . . . . . . . . . . . . . 1. Market Structure . . . . . . . . . . . . A. Competitive Situation . . . . . . . . . B. Degree of Patent Protection . . . . . . C. Demand Trends in Derived Demand Situations . . . . . . . . . . . . . D. Duplication Difficulties by Competitors E. Industrial Plant Capacity Experiences . F. Import Patterns . . . . . . . . . . . . G. Market Share Statistics . . . . . . . . H I mummbuw . Market Trends . . . . . . . . . . . . . . Minimum Corporate Asset Size of Competitors Required to Compete Effectively . . . . . . . . . . . . . J. Profile of Consuming Industries . . . . K. Supply Characteristics . . . . . . . . 2. Buyer Behavior . . . . . . . . . . . . . . A. Annual Purchasing Patterns by Buyers . B. Degree of Backward Integration . . . . C. Degree of Required Deliberation . . . . D. Dependence on Contacts by Company Representatives . . . . . . . . . . . ii Page 120 121 123 123 126 132 132 137 140 146 152 155 161 167 172 173 173 175 178 179 181 182 184 186 187 189 191 192 192 193 195 196 Chapter Page ‘H11 E. Educational Requirements of Users . . . 198 F. Effect of Industrial Advertising on Manufacturer Selection . . . . . . . 199 G. Effect of Product Quality on Source Selection . . . . . . . . . . . . . 201 H. Effect of Reputation and Image of Manufacturer . . . . . . . . . . . . 203 Effect on Sales of Related Products . . 204 Extent of User Laboratory Evaluation . 205 Field Contact Requirements . . . . . . 206 Level of Product Loyalty . . . . . . . 207 Number of Purchasers . . . . . . . . . 209 Product Source Intelligence . . . . . . 210 Recognition Experiences of Product Advantages by Users . . . . . . . . . 211 P. Trade Relations . . . . . . . . . . . . 212 3. Product Characteristics . . . . . . . . . . . 213 . Cyclical Patterns . . . . . . . . . . . 213 . End-Use Patterns . . . . . . . . . . . 215 . Innovation in Marketing . . . . . . . . 216 . Innovation in Technology .p. . . . . . 217 Matching of Technological Characteristics with Market Requirements . . . . . . 218 F. Price McVements . . . . . . . . . . . . 219 G. Relative Technical Service Requirements 224 H. Research and Deve10pment Harnessing Experiences . . . . . . . . . . . . . 225 1. Seasonal Patterns . . . . . . . . . . . 227 J. Specificity of Use . . . . . . . . . . 228 K. Standard Industrial Classification (SIC) Code . . . . . . . . . . . . . 229 OZZFWQH MUOW? L. Strategy on Product Differentiation . . 230 M. Trends in Cross Margins . . . . . . . . 231 N. Type of Product . . . . . . . . . . . . 233 0. Type of Product Demand . . . . . . . . 234 P. Valuation of Byproducts . . . . . . . . 235 4. Related Intrafirm Experiences . . . . . . . . 236 A. Action Concerning Product Improvements 236 B. Basing of Research and Development Program . . . . . . . . . . . . . . . 238 C. Clarity of Product Demand . . . . . 239 D. Effectiveness of Channel for Product Flows . . . . . . . . . . 240 E. Effect of Supply Capabilities in DeveIOpmental Sampling Programs . . . 242 F. Existence of Licensing Arrangements . . 243 G. Export Patterns . . . . . . . . . . . . 244 H. Extent of Plant Capacity Utilization . 246 ii Chapter Page \HII 1. Fruits of Research and DevelOpment . . 248 J. Intensity of the Selling Effort . . . . 249 K. Investment Patterns . . . . . . . . . . 250 L. Merger Activities . . . . . . . . . . . 253 M. Mode of Production . . . . . . . . . . 254 N. Plant Capacity Experiences for Family of Related Products . . . . . 255 0. Product Concept Acceptance . . . . . . 256 P. Production Run Information . . . . . . 258 Q. Promotional Media Strategy . . . . . . 261 R. Promotional Outlay Trends . . . . . . . 262 S. Relative Marketing Commitment . . . . . 264 T. Research and DevelOpment Expenditure Patterns . . . . . . . . . . . . . . 266 U. Source of Product Discovery . . . . . . 269 V. Strategy Concerning Sales Force Size . 270 W. Subjective Measure of Success or Failure: A Management Evaluation . . 271 X. Suitability of Existing Marketing Personnel . . . . . . . . . . . . . . 276 Y Technological Specialty Experiences . . 277 Z. Type of Distribution Channel . . . . . 278 AA. Type of Fixed Capital Employed . . . . 279 BB. Types of Marketing Representatives Utilized . . . . . . . . . . . . . . 280 Types of Sales Patterns . . . . . . . . . . . . . . . 282 Shape and Timing of Sales Patterns . . . . . . . . 291 1. Group 1: Conventional Product Life Cycle Model . . . . . . . . . . . . . . . . . . . 292 2. Group 11: Extended Conventional Product Life Cycle Model . . . . . . . . . . . . 297 3. Group III: Linear Product Life Cycle Mode1 . 301 4. Group IV: Rapid Penetration Model . . . . . 306 5 Group V: Extended Rapid Penetration Model . 311 H PROPOSED RESULTS OF MULTIVARIATE MODELS Introduction . . . . . . . . . . . . . . . . . . 323 Mbltivariate Model of Performance Variable 10, Payback Period . . . . . . . . . . . . . . . . . . 323 1. Criterion . . . . . . . . . . . . . . . . . . 323 2. Model . . . . . . . . . . . . . . . . . . . 324 3. Presentation of Selected Statistical Calculations . . . . . . . . . . . . . . 324 4. Listing of Variables and Scoring Routines Used in the Model . . . . . . . . . . . . . 325 5- Analysis of PrOposed Model . . . . . . . . . 328 Chapter Page IX Multivariate Model of Performance Variable 11, Equivalent Rate of Return . . . . . . . . . . . . . 330 1. Criterion . . . . . . . . . . . . . . . . . . 330 2. Model . . . . . . . . . . . . . . . . . . . 330 3. Presentation of Selected Statistical Calculations . . . . . . . . . . . . . . . 330 4. Listing of Variables and Scoring Routines Used in the Model . . . . . . . . . . . . . 331 5. Analysis of PrOposed Model . . . . . . . . 336 Multivariate Model of Performance Variable 13, Return on Investment . . . . . . . . . . . . . . . 338 1. Criterion . . . . . . . . . . . . . . . . . . 338 2. Model . . . . . . . . . . . . . . . . . . . . 338 3. Presentation of Selected Statistical Calculations . . . . . . . . . . . . . . 338 4. Listing of Variables and Scoring Routines Used in the Model . . . . . . . . . . . . . 340 5. Analysis of Proposed Model . . . . . . . . 341 Multivariate Model of Performance Variable 147, Annualized Discounted Present Value Sum . . . . . . 343 l. Criterion . . . . . . . . . . . . . . . . . . 343 2. Model . . . . . . . . . . . . . . . . . . . . 344 3. Presentation of Selected Statistical Calculations . . . . . . . . . . . . . . . 344 4. Listing of Variables and Scoring Routines Used in the Model . . . . . . . . . . . . . 344 5. Analysis of PrOposed Model . . . . . . . 348 Multivariate Model of Performance Variable 152, Annualized Net Sales . . . . . . . . . . . . . . . 350 1. Criterion . . . . . . . . . . . . . . . . . . 350 2. Model . . . . . . . . . . . . . . . . . . . . 350 3. Presentation of Selected Statistical Calculations . . . . . . . . . . . . . . 350 4. Listing of Variables and Scoring Routines Used in the Model . . . . . . . . . . . . . 352 5. Analysis of PrOposed Model . . . . . . . . 353 Multivariate Model of Performance Variable 153, Annualized Net Profits After Taxes . . . . . . . . 356 1. Criterion . . . . . . . . . . . . . . . . . . 356 2. Model . . . . . . . . . . . . . . . . . . . . 356 3. Presentation of Selected Statistical Calculations . . . . . . . . . . . . . . 357 4. Listing of Variables and Scoring Routines Used in the Model . . . . . . . . . . . . . 358 5. Analysis of PrOposed Model . . . . . . . . . 361 .pu .— Chapter IX Mbltivariate Mbdel of Performance Variable 154, Critical Turning Point for Present Value Calculations . . . . . . . . . . . . . . . . . . . l. Criterion . . . . . . . . . . . . . . . . . . 2. Model . . . . . . . . . . . . . . . . . . . . 3. Presentation of Selected Statistical Calculations . . . . . . . . . . . . . . . 4. Listing of Variables and Scoring Routines Used in the Model . . . . . . . . . . . . 5. Analysis Of PrOposed Model . . . . . . . . X SUMMARY AND FURTHER COMMENTS ON THE INDUSTRIAL CHEMICAL DEVELOPMENT PROCESS Screening Chemical Entities for End-Use Characteristics . . . . . . . . . . . . . . . . . . An Assessment of New Product Activities for Industrial Chemicals from a Financial Viewpoint . . The Appropriate Time Horizon in New Product Proposal Evaluations for Industrial Chemical Products . . The Selection of New Product Entries . . . . . . . The Meaning of Identified Product Life Cycle Structures to Product Management . . . . . . . . . A Tactical Position on Lead Time . . . . . . . . . . Identifying the Structural Characteristics of New Industrial Chemical Products Which Relate to Performance . . . . . . . . . . . . . . . . . . . . 1. Factors relating to performance which were generally supported by empirical research . 2. Factors relating to performance which were partially supported by empirical research . 0n Structuring Multivariate Models of Performance . . Limitations of the Study . . . . . . . . . . . . . . Possible Directions of Future Research . . . . . . . APPENDICES . . . . . a . . g a a a a a a a a a o a o e a a a o a a BIBLIOGRAPHY a . , . . g a a a s a a a a a a a a a a a a a a a a a Page 364 364 364 364 366 368 370 371 372 374 374 375 376 377 377 378 379 380 382 483 Figure 1-1 2-1 32 L4 2-5 2-6 5-1 8-1 8-2 8-3 8-4 8-5 8-6 88 8-9 8-10 8-11 8-12 LIST OF FIGURES Decay Patterns of New Product Ideas . Cumulative Expenditures and Time By Stage of Evolution-A11 Industry Average . . Linear Model . . . Exponential Model "A" . . . Exponential Model "B" 2nd Degree Parabola Gompertz Model . . . . . . . Conventional Product Life Cycle Basic Life Cycle of Products . Contribution of New Products to Expected Growth, 1963-1967 - New Product Evaluation Form . Group 1: Group 1: Group 1: Group 2: Group 2: Group 2: Group 3: Group 3: Group 3: Group 4: Group 4: Group 4: Selected Selected Selected Selected Selected Selected Selected Selected Selected Selected Selected Selected Performance Performance Performance Performance Performance Performance Performance Performance Performance Performance Performance Performance xii O O 0 Indices . Indices Indices Indices . Indices . Indices . Indices Indices . Indices . Indices . Indices . Indices . Page 12 13 19 20 21 21 22 24 77 79 110 294 295 296 298 299 300 303 304 305 308 309 310 Figure Page 8-13 Group 5: Selected Performance Indices . . . . . . . . . 313 8-14 Group 5: Selected Performance Indices . . . . . . . . . 314 8-15 Group 5: Selected Performance Indices . . . . . . . . . 315 .1 .p. .4. Table 1-1 1-2 1-3 1-4 1-5 1-6 1-7 4-2 4-3 4-4 4-7 4-8 4.9 LIST OF TABLES Page Relationship of Industrial R&D to Economic Aggregates: 1953-1968 . . . . . . . . . . . . . . . . 3 Industrial Research and Development Funds By Industry and Source: 1957-1966 . . . . . . . . . . . 4 Estimated Research and DevelOpment Expenditures of U.S. Industry By Type: 1967-1970 a s a a a a a a o a 7 Research and DevelOpment Expenditures of U.S. Industry By State of Research: 1966 . . . . . . . . . 7 Extent of Acquisition Activity of Major Industry (hDupS: 1960-1967 a o a a a a a a a o a a a a o a o a 9 MainPurposeofR&DPrograms.............. 10 Rate of Commercial Success . . . . . . . . . . . . . . . 14 Performance Profile of Major Chemical Firms in Study: 1966 O O O O O O O O O O O O O O I O O O O O l O O 0 O 64 Performance of Major Chemical Firms in Study on Net sales: 1955-1966 e s a a a a a a a a a a o a a a s a 65 Performance of Major Chemical Firms in Study on Net Profit to Common: 1955-1966 . . . . . . . . . . . . . 66 Performance of Major Chemical Firms in Study on Earnings Per Share: 1955-1966 . . . . . . . . . . . . 67 Performance of Major Chemical Firms in Study on Cash Flow Per Share: 1955-1966 . . . . . . . . . . . 68 Performance of Major Chemical Firms in Study on Average Market Price: 1955-1966 or Otherwise Indicated...................... 69 Performance of Major Chemical Firms in Study on Price-Earnings Ratio: 1956-1966 or Otherwise Indicated...................... 70 Performance of Major Chemical Firms in Study on Long Term Debt to Total Capital: 1956-1966 . . . . . . . . 71 Performance of Major Chemical Firms in Study on Selected Profitability Ratios: 1956-1966 . . . . . . 72 xiv Table Page 4-10 Performance of Major Chemical Firms in Study on Retention Rates: 1955-1966 . . . . . . . . . . . . . 74 5-1 Estimation of Outcomes: A Case Study . . . . . . . . . 81 8-1 Polynomial Determination of Industrial Chemical ProductSalesData.................. 122 8-2 Profit Indexes For New Industrial Chemical Products . . 124 8-3 Polynomial Determination of Industrial Chemical ProductPrOfitData................. 125 8-4 Timing of Profit Cycle in Relation to Sales Patterns . . 126 8-5 Minimum Number of Years Required For Net Income on Product as a Percentage of Net Sales to Match or Exceed Its Corporate Performance Record . . . . . . . 128 8-6 Sampling Distribution of Payback Period . . . . . . . . 129 8-7 Minimum Number of Years for Return on Investment Measure to Match or Exceed Its Company's Cost of Capital at the Time of Market Introduction . . . . . . . . . . . 130 8-8 Critical Turning Point in Dollar Flows . . . . . . . . . 131 8-9 Sales Structure on Annualized Net Sales . . . . . . . . 133 8-10 Spearman Rank Correlation Tests of Annualized Net Sales (Research Code Variable 152) Against Selected Variable Measurements . . . . . . . . . . . . . . . . 135 8-11 Nonparametric Statistical Tests of Annualized Net Sales (Research Code Variable 152) Against Selected Group Variables . . . . . . . . . . . . . . . . . . . 136 8-12 Timing of Identified Product Sales Cycles . . . . . . . 137 8-13 Spearman Rank Correlation Tests of Timing of Sales Cycle (Research Code Variable 8) Against Selected Variable Measurements . . . . . . . . . . . . . . . . 139 8‘14 Nonparametric Statistical Tests of Timing of Sales Cycle (Research Code Variable 8) Against Selected Group Variables......................I40 Table 8-15 8-16 8-17 8-18 8-19 8-20 8-21 8-22 8-23 8-24 8-25 8-26 8-27 8-28 8-29 8-30 8-31 8-32 Profit Structure on Annualized Net Profits (Losses) AfterTaxes TimingofProfitCycle....... Spearman Rank Correlation Tests of Annualized Profits (Losses) After Taxes (Research Code Variable 153) Against Selected Variables . . . . . . . . . . . . . Nonparametric Statistical Tests of Annualized Net Profits (Losses) After Taxes (Research Code Variable 153) Against Selected Group Variables . . . . . . . . . Spearman Rank Correlation Tests of Timing of Profit Cycle (Research Code Variable 9) Against Selected Variables . . . . . . . . Median Growth Rate of Net Sales . .. . . O O O O O O C Pbdian Growth Rate of Net Profits . . . . Median Growth Rate of Net Losses . . . . . . . . . . . . Spearman Rank Correlation Tests of Growth Characteristics Against Selected Variables . . . . . . . . . . . . . bhnparametric Statistical Tests of Growth Characteristics Against Selected Group Variables . . . . PaybaCk PeriOd O O O O O O O O O O O O C O O O I O O Spearman Rank CorreLation Tests of Payback Period (Research Code Variable 10) Against Selected Variable Measurements . . . . . . . . . . . . . . Nonparametric Statistical Tests of Payback Period (Research Code Variable 10) Against Selected Group Variables . . . . . . . . Accounting Rate of Return . . . . . Equivalent Rate of Return . . . . . Median Return on Investment . . . . Spearman Rank Correlation Tests of Investment Return Measures Against Selected Variables . . . . . . . . Nonparametric Statistical Tests of Investment Return Measures Against Selected Group Variables Page 141 141 143 144 145 146 147 147 1119 151 152 154- 155 155 ISES 1156 '158 16“) 'a a 'A....., aD—.‘ h...‘l1 "o. Table 8-33 8-34 8-35 8-36 8-37 8-38 8-39 8-40 8-41 8- 42 8-43 8-44 8-45 8-46 3-47 8-48 Page Discounted Present Value Sum . . . . . . . . . . . . . . 161 Annualized Discounted Present Value Sum . . . . . . . . 162 Critical Turning Point For Present Value Calculation . . 162 Spearman Rank Correlation Tests of Present Value Structure Against Selected Variables . . . . . . . . . 165 Nonparametric Statistical Tests of Present Value Structure Against Selected Group Variables . . . . . . 167 PrOfitability Ratio 0 O O O O O I O O O O O O O O D O O 168 Spearman Rank Correlation Tests of Relative Contribu- tion to Wealth Against Selected Variables . . . . . . 169 Nonparametric Statistical Tests of Relative Contribu- tion to Wealth Against Selected Group Variables . . . 171 Nonparametric Statistical Tests of Direct Competitive Situation (Research Code Variable 27) AgainSt Selected Performance Criteria . . . . . . . . . . . . 174 Nonparametric Statistical Tests of Indirect Competitive Situation (Research Code Variable 28) Against Selected Performance Criteria . . . . . . . . . . . . 175 Nonparametric Statistical Tests of Process Patent Protection (Research Code Variable 24) Against Selected Performance Criteria . . . . . . . . . . . . 176 Nonparametric Statistical Tests of Demand Trends in Derived Demand Situations (Research Code Variable 99) Against Selected Performance Criteria . . . . . . . . 179 Nonparametric Statistical Tests of Duplication Difficulties by Competitors (Research Code Variable 78) Against Selected Performance Criteria . . . . . . . . 180 Nonparametric Statistical Tests of Industrial Plant Capacity Experiences (Research Code Variable 65) Against Selected Performance Criteria . . . . . . . . 182 Nonparametric Statistical Tests of Import Patterns (Research Code Variable 115) Against Selected PerformanceCriteria................. 133 Nonparametric Statistical Tests of Market Share Statistics (Research Code Variable 20) Against Selected Performance Criteria . . . . . . . . . . . , 185 l‘.’ »- 9a... a . \; L '50., Table 8-49 8-50 8-51 8-52 8-53 8- 54 8-55 8-56 8-57 8-58 8-59 8-60 Page Nonparametric Statistical Tests of Market Trends (Research Code Variable 95) Against Selected PerformanceCriteria................. 187 Nonparametric Statistical Tests of Minimum Corporate Asset Size of Competitors Required to Compete Effectively (Research Code Variable 94) Against Selected Performance Criteria . . . . . . . . . . . . 188 Nonparametric Statistical Tests of Number of Major Consuming Industries (Research Code Variable 30) Against Selected Performance Criteria . . . . . . . . 189 Nonparametric Statistical Tests of Number of Minor Consuming Industries (Research Code Variable 31) Against Selected Performance Criteria . . . . . . . . 190 Nonparametric Statistical Tests of Number of Annual Purchases by Buyers (Research Code Variable 82) Against Selected Performance Criteria . . . . . . . . 193 Nonparametric Statistical Tests of Degree of‘Backward Integration (Research Code Variable 109) Against Selected Performance Criteria . . . . . . . . . . . . 19S Nonparametric Statistical Tests of Degree of Required Deliberation (Research Code Variable 108) Against Selected Performance Criteria . . . . . . . . . . . . 196 Nonparametric Statistical Tests of Dependence on Contacts by Company Representatives (Research Code Variable 89) Against Selected Performance Criteria . . 198 Nonparametric Statistical Tests of Time to Educate the User (Research Code Variable 112) Against Selected PerformanceCriteria................. 199 Nonparametric Statistical Tests of Effect of Industrial Advertising on Manufacturer Selection (Research Code Variable 90) Against Selected PerformanceCriteria................. 201 Nonparametric Statistical Tests of Effect of Product Quality on Source Selection (Research Code Variable 86) Against Selected Performance Criteria . . . . . . . . 202 Nonparametric Statistical Tests of Effect of Reputation and Image of Manufacturer (Research Code Variable 77) Against Selected Performance Criteria . . 203 . C 7 r...,_.. . .. _ -.,_ . . I.. ~.,_ . v... V a u.__ .. ' ...Q. ‘ . '-..._ . a. . . . f“: - 'D . ,. , 7. . '. \"c. . . an . \. v.‘ . . Table 8-61 8-62 8-63 8-64 8- 65 8-66 8-67 8-68 8-69 8- 70 8-71 8- 72 Nonparametric Statistical Tests of Effect on Sales of Related Products (Research Code Variable 87) Against Selected Performance Criteria . . . . . . . . Nonparametric Statistical Tests of Percentage of User Laboratory Evaluation (Research Code Variable 76) Against Selected Performance Criteria . . . . . . . . Nonparametric Statistical Tests of Number of Contacts Required by Marketing and Technical Personnel (Research Code Variable 113) Against Selected Performance Criteria....................... Nonparametric Statistical Tests of Level of Product Loyalty (Research Code Variable 107) Against Selected PerformanceCriteria. .. . . . . . . . . . . . . . . Nonparametric Statistical Tests of Number of Purchasers (Research Code Variable 101) Against Selected Performance Criteria . . . . . . . . . . . . . . . . . Nonparametric Statistical Tests of Number of Product Sources by Buyers (Research Code Variable 100) Against Selected Performance Criteria . . . . . . . . Nonparametric Statistical Tests of Recognition Experiences of Product Advantages by Users (Research Code Variable 81) Against Selected Performance criteria 0 O O O O O O O O O O O O O O O O O O O O O O Nonparametric Statistical Tests of Trade Relations (Research Code Variable 104) Against Selected PerformanceCriteria................. Nonparametric Statistical Tests of Cyclical Patterns (Research Code Variable 103) Against Selected PerformanceCriteria. . . . . . . . . . . .. . . .. Nonparametric Statistical Tests of End-Use Patterns (Research Code Variable 134) Against Selected PerformanceCriteria. .. . .... ... . Nonparametric Statistical Tests of Degree of Marketing Innovativeness (Research Code Variable 84) Against Selected Performance Criteria . . . . . . . . . . . Nonparametric Statistical Tests of Degree of Technological Innovativeness (Research Code Variable 85) Against Selected Performance Criteria . . . . . . . Page 204 206 207 208 210 211 212 213 214 216 217 218 n .- ... o. I - I ...-ov- ‘1 I Table 8-73 8-74 8-75 8-77 8-78 8-79 8-80 8-81 8-82 8-83 8- 84 Nonparametric Statistical Tests of Matching of Technological Characteristics with Market Require- ments (Research Code Variable 83) Against Selected Performance Criteria . . . . . . . . . . . . Nonparametric Statistical Tests of Action Regarding Price Changes (Research Code Variable 63) Against Selected Performance Criteria . . . . . . . . . . . . Nonparametric Statistical Tests of Causes of Marked Price Declines (Research Code Variable 34) Against Selected Performance Criteria . . . . . . . . . . . . Nonparametric Statistical Tests of Relative Technical Service Requirements (Research Code Variable 114) Against Selected Performance Criteria . . . . . . . . Nonparametric Statistical Tests of Research and Develop- ment Harnessing Experiences (Research Code Variable 73) Against Selected Performance Criteria . . . . . . . . Nonparametric Statistical Tests of Seasonal Patterns (Research Code Variable 102) Against Selected PerformanceCriteria..... ............ Nonparametric Statistical Tests of Specificity of Use (Research Code Variable 32) Against Selected PerformanceCriteria................. Nonparametric Statistical Tests of Standard Industrial Classification (SIC) Code (Research Code Variable 133) Against Selected Performance Criteria . . . . . . . . Nonparametric Statistical Tests of Strategy Concerning Number of Product Offerings For Same End-Use (Research Code Variable 45) Against Selected Performance Criteria 0 O C O O O O O C O O I O O O O C O O O O 0 O Nonparametric Statistical Tests of Trends in Cross Margins (Research Code Variable 120) Against Selected PerformanceCriteria. ..... ........... Nonparametric Statistical Tests of Type of Product (Research Code Variable 130) Against Selected PerformanceCriteria... .............. Nonparametric Statistical Tests of Type of Product Demand (Research Code Variable 97) Against Selected PerformanceCriteria . . . . . . . . . . . . . . . Page 219 221 223 225 226 227 228 230 231 232 233 235 Table Page 8-85 Nonparametric Statistical Tests of Valuation of Byproducts (Research Code Variable 121) Against Selected Performance Criteria . . . . . . . . . . . . 236 8-86 Nonparametric Statistical Tests of Action Concerning Product Improvements (Research Code Variable 55) Against Selected Performance Criteria . . . . . . . . 237 8-87 Nonparametric Statistical Tests of Basing of Research and DevelOpment Program (Research Code Variable 123) Against Selected Performance Criteria . . . . . . . . 239 8-88 Nonparametric Statistical Tests of Clarity of Product Demand (Research Code Variable 98) Against Selected Performance Criteria . . . . . . . . . . . . . . . . . 240 8-89 Nonparametric Statistical Tests of Effectiveness of Channel For Product Flows (Research Code Variable 110) Against Selected Performance Criteria . . . . . . . . 241 8-90 Nonparametric Statistical Tests of Effect of Supply Capabilities in DevelOpmental Sampling Programs (Research Code Variable 75) Against Selected PerformanceCriteria.... . .. 243 8-91 Nonparametric Statistical Tests of Licensing EXperiences (Research Code Variable 79) Against Selected Performance Criteria 0 O O O O C O O O O O O O O C O C D O O O O O 244 8-92 Nonparametric Statistical Tests of Export Patterns (Research Code Variable 116) Against Selected Performance Criteria . . . . . . . . . . . . . . . . . 245 8-93 Nonparametric Statistical Tests of Extent of Plant Capacity Utilization (Research Code Variable 126) Against Selected Performance Criteria . . . . . . . . 247 8-94 Nonparametric Statistical Tests of Fruits of Research and DevelOpment (Research Code Variable 74) Against Selected Performance Criteria . . . . . . . . . . . . 248 3'95 Nonparametric Statistical Tests of Strategy Concerning Number of Sales Calls Per Unit of Time (Research Code Variable 41) Against Selected Performance Criteria.......................250 8‘96 Nonparametric Statistical Tests of Median Cumulative Investment Requirements (Research Code Variable 53) Against Selected Performance Criteria . . . . . . . . 251 Table 8-97 8-98 8- 99 8-100 8-101 8-102 8-103 8-104 8-105 8-106 8-107 8-108 Nonparametric Statistical Tests of Median Annual Incremental Investment Requirements (Research Code Variable 54) Against Selected Performance Criteria....................... Nonparametric Statistical Tests of Merger or Combination Experiences (Research Code Variable 92) Against Selected Performance Criteria . . . . . . . . Nonparametric Statistical Tests of Mode of Production (Research Code Variable 129) Against Selected PerformanceCriteria.. .. ... .. . . .. . . . . Nonparametric Statistical Tests of Plant Capacity Experiences For Family of Related Products (Research Code Variable 66) Against Selected Performance Criteria....................... Nonparametric Statistical Tests of Product Concept Acceptance (Research Code Variable 111) Against Selected Performance Criteria . . . . . . . . . . . . Nonparametric Statistical Tests of Length of Production Run (Research Code Variable 125) Against Selected Performance Criteria . . . . . . . . . . . . Nonparametric Statistical Tests of Number of Annual Production Runs Scheduled (Research Code Variable 127) Against Selected Performance Criteria . . . . . . . . Nonparametric Statistical Tests of Outlay Trends For Product Promotion (Research Code Variable 88) Against Selected Performance Criteria . . . . . . . . . . . . Nonparametric Statistical Tests of Relative Marketing Costs (Research Code Variable 23) Against Selected PerformanceCriteria.. .. ... ... .... .. . Nonparametric Statistical Tests of Aggregate Research and Development Expenditures (Research Code Variable 21) Against Selected Performance Criteria . . . . . Nonparametric Statistical Tests of Strategy Concerning Research and DevelOpment in Related Areas (Research Code Variable 51) Against Selected Performance Criteria....................... Nonparametric Statistical Tests of Source of Product Discovery (Research Code Variable 132) Against Selected Performance Criteria . . . . . . . . . . . xxii Page 252 253 254 256 257 259 260 263 265 267 268 270 Table 8-109 8-110 8-111 8-112 8-113 8-114 8-115 8-116 8-117 8-118 8-119 8-120 9-1 9-2 9-4 Page Nonparametric Statistical Tests of Strategy Concerning Sales Force Size (Research Code Variable 39) Against Selected Performance Criteria . . 271 Spearman Rank Correlation Tests of Subjective Measure Of Success or Failure (Research Code Variable 122) Against Selected Performance Criteria 273 Nonparametric Statistical Tests of Subjective Measure of Success or Failure (Research Code Variable 122) Against Selected Performance Criteria . 275 Nonparametric Statistical Tests of Suitability of Existing Marketing Personnel (Research Code Variable 91) Against Selected Performance Criteria 276 O C O O 0 O Nonparametric Statistical Tests of Technological Specialty Experiences (Research Code Variable 72) Against Selected Performance Criteria 277 Nonparametric Statistical Tests of Type of Distribution Channel (Research Code Variable 119) Against Selected Performance Criteria . . 279 Nonparametric Statistical Tests of Type of Fixed Capital Employed (Research Code Variable 124) Against Selected Performance Criteria . . . 280 Nonparametric Statistical Tests of Types of Marketing Representatives Utilized (Research Code Variable 131) Against Selected Performance Criteria 281 O O O O I O I O Identified Shapes of Sales Patterns . . . . . . . . . . 282 Ranking Results of Selected Variables Against the Types of Sales Patterns (Research Code Variable 150) . 287 Identified Shapes and Timings of Sales Distributions . . 291 Ranking Results of Selected Variables Against the Shape and Timing of Sales Patterns (Research Code Variable 151) . . . . . . . . Historical Fit of Model For Variable 10 . Historical Fit of Model For Variable 11 . . Historical Fit of Model For Variable l3 . . . Historical Fit of Model For Variable 147 . . . C Q 0 a . vviii -0~'~ u _..-.aa' ..sau‘ -..-u. . ...-on. Table Page 9-5 Historical Fit of Model For Variable 152 . . . . . . . . 353 9-6 Historical Fit of Model For Variable 153 . . . . . . . . 361 9-7 Historical Pit of Model For Variable 154 . . . . . . . . 368 10-1 Rank Correlation Coefficients of Discounted Present Value Sum Figures Through Time . . . . . . . . . . . . 373 , 'O-v.. LIST OF APPENDICES Appendix Page A LiSting Of variables 0 O O O O O O O O O O O O O O O O O 383 B Nonparametric Statistical Tests of Group Variables Against Selected Performance Criteria . . . . . . . . 389 C Listing of Calculating Formulas For Selected StatisticalTests..................421 D Questionnaire.....................427 E ProgramAnalysis....................477 A. - '-vu-... ‘11 v. a” ' _— , : ..V..~.‘. vs, a 'I. '- ‘-. I .... b. ‘ . \ " -.~ . 'I..»I ‘ u. ’I ‘Q- 1. ~ '9 .- ~, ' -- a. .r _‘ v ,__V ~ b\. u... I“ m‘ “-s ‘4 . A. . N -. ...- I‘m. ' ._ \. _ “I ._ . " “~. \. ‘. :‘I .. \h - ... o . ". I. . .‘ 'V“.’ ~ . . ._ ‘I ~.. .-. .! . . . Q '- A. ‘ t CHAPTER I INTRODUCTION General Background Most major chemical firms confront each year an avalanche of new product proposals, more than they can eXploit profitably, for the emphasis on chemical research and deve10pment generates an ever-increasing output of new chemical intermediates and Specialties. There are two basic reasons for the accelerating stream of new products. The above-average growth record of the chemical and allied products industry has been a source of funds for reinvestment in research and deve10pment in both eXploratory and applied research ventures. The chemical industry has long emphasized the need for innovation. As products become technologically obsolescent. inPuts for research and deve10pment have included in-house observations and experimentation as well as an increasing awareness by technical and marketing pe0p1e of the needs of industrial users. Many technical and professional peOple directly involved in the evaluation of new products make important decisions influencing the future growth and PErformance of the total enterprise. So this study focuses on one Of the key problems facing management, how to best handle the new PrOdUCt deve10pment effort. Specifically, this research seeks to integrate prOdUCt life cYcle and capital budgeting theories, since they are viewed as having Pertinence to an understanding of new product planning and management. R \esearch and DevelOpment Expenditure Patterns Industrial research and development has emerged in the last decade as a major source of technological change and economic 81‘0““, with actual exPendj-tul'cs for industrial research nearly tripling between 1953 and 1960 (see Table 1‘1)- The long term trend of research performed by industrial \L organizations since 1953 equates to a 11.07. annual compound growth rate. The significance of these figures is enhanced by comparing this industrial research and development growth with that of the Gross National Product, though such growth in this decade has generally paralleled that of the general economy. Research by industrial firms consists largely of scientific investigations having comercial orientations: approximately three-fourths of R&D spending in the United States in 1966 was limited to development work defined as the "systematic use of scientific knowledge directed toward the production of useful materials, devices, systems or methods, including design and deve10pment of prototypes and processes."1 Industrial firms c00perate in research and deve10pment work in many sectors of the economy, including the aerospace, defense, and chemical industries, a diversity indicated in Table 1-2. \ "“10““ Science Foundation, Reviews of Data on Research and M’ N°° 41 (Washington, D.C.: U.S. Government Printing Office, 38 te 3c: aber’ 1963): Po 10. National Science Foundation, Review! of Data on "we R - 8 Government W "31" 68-5 No. 12 ashin ton D.C.. U. - Printing office . Elfin-ran I :I‘fl \ .a CS" 8 ’ , - ' 0 SOURCES FOR TABLE 1-1 National Science Foundation, Basic Research, Applied Resegggh‘_ggg ngglogment in Industry, 196§, Surveys of Science Resources Series, NSF 67-12 (Washington, D.C.: U.S. Government Printing Office, June, 1967), p. 20. National Science Foundation, National Patterns of R80 Resources. Eundg gnd flengowg; in the united States gl953-682, NSF 67-7 (Washington, D.C.: UISI GOV.rnment Printing Office, April, 1967), p. 22. National Science Foundation, Reviews of Data on Science Resougges, NSF 68-5, No. 12 (Washington D.C.: U.S. Government Printing Office, January, 1968), p. A. U.S. Department of Commerce, Office of Business Economics, The National Income and Product Accounts of the United States, 1929-1965 EWZBhington, D.C.: U.S. Government Printing Office, August, 1966), pp. U. S. Department of Commerce, Office of Business Economics, Survgx 9§_Q2££§n£_§ggigg§g0 Vol. 48, No. 1 (Washington, D.C.: U.S. Government Printing Office, January, 1968), p. S- 1. U. S. Department of Commerce, Office of Business Economics, Survey “98 . Vol. 48, No. 11 (Washington, D.C.: U.S. Government Printing Office, November, 1968), p. S-l. TABLE 1-1 RELATIONSHIP OF INDUSTRIAL R&D TO ECONOMIC AGGREGATES: 1953-1968 N _ Industrial R&D Nonresidential Percent of Expenditures GNP Percent Fixed Investment Nonresidential (Millions of (Billions of of (Billions of Fixed Year dollars)1a2 dollars) GNP dollars)3 Investment 19684 17,300 860.9 2.0 89.4 19.8 1967 16,6104 789.7 2.1 83.6 20.1 1966 15,541 747.6 2.1 81.3 19.2 1965 14,197 684.9 2.1 71.3 20.0 1964 13,512 632.4 2.1 61.1 22.1 1963 12,630 590.5 2.1 54.3 23.3 1962 11,464 560.3 2.1 51.7 22.2 1961 10,908 520.1 2.1 447.0 23.2 1960 10,509 503.7 2.1 48.4 21.7 1959 9,618 483.7 2.0 45.1 21.3 1953 8.389 447.3 1.8 41.6 20.2 1957 7,731 441.1 1.8 46.4 16.7 1956 6.605 419.2 1.6 43.7 15.1 1955 4,6404 398.0 1.2 38.1 12.2 1954 4.0704 364.8 1.2 33.6 12.1 1953 3.630 364.6 1.0 34.2 10.6 \ Data exclude company-financed work contracted to outside organizations. 2 Research and development work includes basic and applied research in a dthe natural sciences, including the medical sciences and engineering, n deVElOPment. InCIUdes the net acquisition of fixed capital goods by private bu ' Blue88 and nonprofit institutions. 4 Esiiilnated. A a. '- and-......- Use-unwa- —_.—c.. a‘\) - Aemsanucoov own mmq mom mom o¢¢.m Nam.q ofim.m omm.u moawmmwa mam ummuouw< onm man woo “Hm HNM.H omo.~ own men unusafisvo coHuouuodmcmuu . nonuo was moaoaso> HOuoz moq.a amo.s new moo ohm.m oom.~ ~mm.~ aom.~ nonsmoncsasoo mom uawansvo Hmowuuooam wmm moN mmm mmm Hom.H ,mmm mom moo huoafisomz qu oNH mod no sea mmH qu mmH muosooua Hmums moumogunmm Hmm qua NoH moH mum mmH nma moH mamuoa muwswum nNH as .<.z .¢.z ems ooH mm «as muoaeoca mmmam 6am .amau .6a68m 0mm Ham mm cm «ma 0mm HNH mofi nauseoud Happen mwm cow cum ooN qu mam com Ham meowu nomuuxo was mcwcwwou asoaouuom awn.“ soo.~ now one an.H an~.H 0mm nee nauseous cusses was masonaoso mm me .<.z mm mm on em mm muosvoud commas can pause .<.z .<.z m «H «H AH oH «H ousuasuam mam .muoovoua woos .uonfidg .<.z mm am «a we on mm mg Hoummam mam moawuxoa non .<.z no «m 88H one see an mongoose success was soon coma mead coma snag coma mama coma mmma huuosoca Assesses no maoaaaazv Hawumaaoo «6 neonafiazv ‘11 comm huumsocH wum>wum mocsm amDOH womathmwd "Envy—30m QZ< NMHMDQZH Wm mQZDh ESCHEMQ Q24 ZUMawm—mnmd Aoo .m.D H.U.Q .GOumaasmmBV NH .02 .mrmo mmz .mmousomom mocofiom so muma mo mBmH>wm..:ofiumocsom oucowom HmcoHumz . .om new mm .68 .Aaooa .mcsu .monm we a: n ucmEch>ou .m.: u.U.Q .aOuwcflsmmzv.mHumo mmz .mmfiumm moonsommm mocmwom mo mko>usm .momwo.huwwswcm. cm ucoedoHo>oQ mam «noumomom vmwfiqmd «noumomom ofimmm .coHumwcsom mocmwom Hmoowumz "mMUMDom .mdnmfiwm>m uoa mm vmwMfiomam mumv muum: c an m zflflmowwwomam uoc mmfiuumsvcw wcHusuommscmEcoc mam wafiusuommscme nonuo Ham pom mumm mmwswoWH v ume m .cowumocsom mocmfiom Hmcoflumz onu ma woumswumm .mmwcmano Esoaouuoa mo mmwuw>wuom coflumuoHaxo Hmowmxndoow can HmowmoHoow moosaoxmo m .mocsw Hmuoomm mo :oHumnHuucoo movsfloxm N .moHuH>wuom Hmowccomucoa nocuo was .ooH>uom mmamm .ooHuoEoua moamm .soumommu umxume .wcflummu uosooua oaflusou .xuo3 Houucoo zuwamsd moosfioxo mhuumsmCH cwcufis voEuomuoa xuos usoanHo>ov mam .soummmou voHHQam .noumomou ofimmn mmwsaocH a .Hmuou ow ommsfioofi use .zHoumuwmow oHAmHHm>m uoz n .¢.z am~.a oom.m m~¢.s emm.m Hem.ma cmo.~a oom.oH Hma.h 44508 ma~ 5mg moH .<.z oma omm Nam .<.z mmoauumsoca nonuo mom mom 05H 08H «as omm mum mom muaoasuumaw oqufiucofiom was Hmaofimmowoum ooafi mama ocmfi nmma oomfi momfi coma mmaa muumsocH Nauumanoe «6 macaHHaZM. Hmwumfieoo no «nonsense awash huumnvcH oum>aum awash amuoa Aconcaucoov methmmH "WDMDOm DZ< WEmDDZH Mm mQZDh Hzgoanmsma 92¢ IUMdVMmWM QHMHmDQZH NIH agfi . . . -n. .. a r .o-. ~ «v 1 - .-.-- ...... ... ..- ...,.A . -.>. —-....., ,,.‘ I u a ' ~‘.. . 5 1 .4 ._ \ a‘ .... _. . ‘QI - . . - a .. .a. _~ . '- -u. -.. . ‘ ‘Il._ .a. ., . 1 .. . . ‘a \ .\1 . 1 Once we delete the effect of federally supported funds, the chemical and allied products industry accounts for one of the largest amounts of total private funds spent on research and development. These expenditures have been growing rapidly, more than doubling since 1957. Most of its influence on the economy emerges in changes in productivity induced by technological change. Research and development activities in the chemical industry have received ever-increasing emphasis. Having thought to accelerate new product introductions as well as to increase the obsolescence rate of existing products, applied new product development has altered existing market structures and competitive forces profoundly. Recent projections by the McGraw-Hill Department of Economics indicate continued increases in research and development outlays: by 1970, they estimate the chemical and allied products industry will be spending $1.84 billion.2 Of such funds spent in 1966, the chemical industry committed approximately 137. to basic research, 417. to applied research and the remaining 467. to development (see Tables l~3 and 1-4). Emphasis on New Product DevelOpment as an Industrial Marketing Strategy Chemical manufacturing firms have relied on four basic types of activities to generate product growth: Acquisition, merger, or combination arrangements, Product improvement work, New process discoveries, and New product deve10pment. Dubai—- 0... 2 "R&D Looms Big in Fiscal Budgets," Business Week, No. 1967 (May 13, 1967). PP. 68-69+. -.-: .....- ... . ‘ ' - "'9. .-- . '0. .’ .... ..‘ a . ------.. '. . . .. 1 . .... \s. ‘3 .. 5,, I , -. "H ‘- '0-.‘ \ ' l. "'\u ..., 1 ,- \.‘ ' . 1.1 a. a I. g‘ . ,~ ‘. . u \ ‘- TABLE 1- 3 ESTIMATED RESEARCH AND DEVELOPMENT EXPENDITURES OF U.S. INDUSTRY BY TYPE: 1967-1970 T (Millions of dollars) 1967 1970 Chemicals and allied products 1,561 1,842 Petroleum and coal products 488 571 Rubber products 176 208 Stone, clay, and glass products 151 205 Nonferrous metals 99 117 Paper and allied products 88 104 ALL INDUSTRY 16,605 20,792 SOURCE: "R&D Looms Big in Fiscal Budgets," Business Week, No. 1967, May 13, 1967, pp. 68-69+. TABLE 1-4 RESEARCH AND DEVELOPMENT EXPENDITURES OF U.S. INDUSTRY BY STATE OF RESEARCH: 1966 (Percent) 1966 1966 Basic Applied 1966 Research Research Development Chemicals and allied products 12.5 41.1 46.4 Petroleum and coal products 9.9 39-0 51.1 Rubber products 5.0 36.2 58.8 Stone, clay and glass products 6.2 45.1 48.7 Nonferrous metals 9.9 51.4 38. 7 Paper and allied products 8.3 39.3 52,4 ALL INDUSTRY 7.2 24.1 68.7 SOURCE: "Rousing Forecast for Research," W, Vol. 98, N0. 20, May 14, 1966, p. 61. In 1962, a committee of Congress studying the acquisition eXperiences of major industrial and merchandising firms concluded that "chemical companies have joined forces in order to exploit joint interests and garner captive sources of raw material."3 This reflects the growing concern over the impact of merger activity by major firms on the general economy and the welfare of the ultimate consumer. Table 1-5 indicates the relative extent of recent acquisition activities by major industrial manufacturing sectors. Clearly many firms within the chemical industry have taken the merger and acquisition route to broaden marketing horizons and improve financial performance records, as well as relying on internal growth. While future antitrust action by governmental special interest groups may reduce its significance as an alternate industrial strategy for major firms, acquisi- tions and mergers will continue as an open alternative to the internal development of new products. U.S. Congress, House of Representatives, Select Committee on 81111111 Business, Mergers and Superconcentration: Acquisitions of 529 Largest Industrial and 50 Largest Merchandising Firms, 87th Congress (Wazhington, D.C.: U.S. Government Printing Office, November 8, 1962), P- 3. '- .. ..._ ‘ ‘~ ....,._ ‘ " ~-. " '- ... 1' 1 ' ‘L ~ -... ‘- . ,’ “.. --.. .. ~\ '3 \1. —. TABLE 1-5 EXTENT OF ACQUISITION ACTIVITY OF MAJOR INDUSTRY GROUPS: 1960-1967 Major Industry 0mm; of Average Acquiring Company 1955-1959 1960 1963 1966 1967 Food and kindred products 58 61 67 69 95 Tobacco manufactures 4 2 6 9 5 Textile mill products 30 42 37 27 22 Apparel 6 11 25 37 45 Lumber products, except furniture 11 26 21 15 24 Furniture and fixtures 3 6 8 14 16 Paper andvallied products 31 52 16 21 36 Printing and publishing 13 26 31 23 33 Chemicals 57 68 78 105 123 Petroleum 14 10 14 13 10 Rubber and plastics 14 14 15 29 Leather products 4 1 6 6 7 Stone, clay and glass 23 27 15 27 35 Primary metals 34 29 35 33 65 Fabricated metal products 42 45 46 50 87 Nonelectrical machinery 86 77 88 102 155 Electrical machinery 64 113 109 145 257 Transportation equipment 51 67 46 64 103 Professional and scientific 24 35 23 50 92 Miscellaneous and ordnance 16 30 26 16 22 MANUFACTURING 576 742 716 841 1 , 261 _ 1 Data limited to mergers and acquisitions reported by Moody's Investors Service, Inc., and Standard and Poor's Corporation. SOURCE: Bureau of Economics, Federal Trade Commission, News Release, March 18, 1968, p. 13. p , .-.-... ... ..- s" 5 o..- _... u. 'v \I a... 'a. 10 The objectives proposed by executive management for research and development work should establish the relative importance of product improvements, new processes, and new products as industrial strategies. AMcGraw-Hill Department of Economics survey showed that new products were usually one goal of the R&D efforts made by the respondents: Table 1-6 shows the breakdown of responses by industry. Note the difference in emphasis on new product deve10pment among the various industries listed. TABLE 1- 6 MAIN PURPOSE OF R&D PROGRAMS (Z of companies responding) Improving New Existing New Industry Products Products Processes Chemicals and allied products 70 20 10 Petroleum and coal products 27 33 40 Rubber products 17 83 -- Stone, clay and glass products 41 41 18 Nonferrous metals 39 44 17 Paper and allied products - 37 41 22 ALL INDUSTRY 45 41 14 M SOURCE: "Rousing Forecast for Research," Chemical Week, Vol. 98, No. 20, May 14, 1966, p. 62 The nature of the product generally dictates the strategies employed and may itself be a limiting factor. In any event, new products are fundamental to corporate growth, most new product programs in fact receiving a disprOportionate amount of attention by management because of the time required for the evaluation process and program implementation. ... - . .,. -n.... .. ‘ .. ,_' v s .‘h' .6. -..__. .. ‘ s. v - :'. "a . ..- ,\._- ._1 ‘u. q. ‘ . 11 Basic Structure of the New Product Evaluation Process New products are not commercialized randomly: each firm somehow has specific procedures for screening prOposals, which may start as nothing more than notions. There is an attrition of new product ideas during the various phases through which prOposals are channeled, with this decay rate presented in Figure 1-1. Most preposals are eliminated during preliminary and revised financial and marketing analyses. The Commercial Chemical DevelOpment Association has estimated that for each 540 new industrial chemical ideas proposed, only one product is O O 4 commerc1a11zed. The stages faced by new product introductions can be understood in terms of the expenditures incurred. Each stage towards commercialization becomes increasingly more expensive, requiring more of a firm's resources to insure proper preparation of the product for introduction and subsequent buyer acceptance. Booz-Allen and Hamilton, one management consulting organization that has had the opportunity to analyze a number of new product programs, has drawn up the industry average cost relationships through time shown in Figure l--2.S Notice how costs accumulate through time in the various stages. Since the major expenditures come after the decision has been made to deve10p the new product, it becomes imperative to exhaustively analyze all proposals on their financial and marketing merits before making a decision to develOp; the cost of failing to make such an analysis can be too burdensome for any company, regardless of size. The final construct considered in the new product evaluation process, the success-failure rate of new products, again is reflected in the latest analysis of thousands of new product histories compiled by Booz.A11en and Hamilton. It is noteworthy to find few interindustry differences in success-failure rates among the six industrial sectors covered in the analysis (intrafirm differences within any industry are M 4 __ As reported by Conrad Berenson (ed.), The Chemical Industry: Viezpoints and Perspectives, (New York: Interscience Publishers, 1963), p‘ 5s 5 Management of New Products, (4th edition, New York: Booz.A11en .W< ‘A‘I I‘LVOIII1‘1 a w 11111111 12 92 38:85:95 856m .25 8:6» 365 was“: SO. oza 5.83m; . E. 302. 1.40.2on 1mm...» .5228 2850 E95 8834 8:9... .x. Duet—zmam m4_...<4:2:o o oo. om 8 ca 8 o mzme_ mmzmaxm 8 ca ammmEazmaxH t o. 8 /mmmstozmaxm r .252. On w 4.322.323. ..x. .mmmstozmaxm 295.65 .58 4 _ _oo_ m0< >mhmboz_1.jm mo moflb >m 22.... w mmmatozmnxm m>_._.<15230 N1. umber. .~..- // 7. . ,4 u \ a. I. .- 'u u_‘ . u,_ ... . to a 14 a different matter). Table 1-7 indicates that the average rate of commercial success for chemical products was 597.. Since most new products at the development stage are failures according to some criterion or another, we need greater selectivity in the evaluation process, requiring more information and more extensive marketing analysis. Yet this success- failure rate further suggests the unusually high uncertainty associated with the comercialization process. The firm that fails to take into account these financial and marketing uncertainties severely limits improvement in its success-failure rate. TABLE 1 - 7 RATE OF COMMERCIAL SUCCESS (As percentages of success) Product New New Product DevelOpment Products Industry Classification Ideas Projects Introduced Chemicals 2 18 59 Consumer packaged goods 2 ll 63 Electrical machinery 1 13 63 Metal Fabricators 3 11 71 Nonelectrical machinery 2 21 59 Raw material processors 5 14 59 —.__ SOURCE: Management of New Products, (4th edition, New York: Booz-Allen and Hamilton, Inc., 1965), p. 12. identification of the Problem This research derived its inspiration from a number of factors relating to new product evaluation. Of all inputs used in any quantita- tive model, demand estimation is generally the most uncertain in new Product evaluation; and the uncertainty of these estimates generally increases as the projected time span increases. But demand estimates appear to be the critical input in any capital budgeting model for evaluat- ing new products; since material, labor, and related costs each vary with -._~. O... ‘- ,. .'.‘ I» ... ..Za 7‘ “n~a 15 the quantity produced, the projected level of demand is the single most important consideration. Any rejection or acceptance decision is based on the worth of a new product, which is directly influenced by the level of existing demand . Other investigators have attempted in the recent past to establish criteria for new product selection; Bertram Schoner constructed a stochastic model for the selection of applied research and deve10pment projects.6 He attempts to represent mathematically a maximization process through utility theory which accounts for interaction between projects. Two important assumptions in this model limit its applicability to a given firm or industry situation: 1. Rapid obsolescence makes it unwise to consider sales from a product as extending further than five years beyond a product's introduction. 2. Contributions to profit are highest in the first year of a product's introduction and decline exponentially thereafter.7 Before any such model for the selection of research and deve10pment projects can be logically employed in any industrial situation, empirical research must test these underlying assumptions. Nevertheless, the acceptance of product life cycle theory within the firm is critical in the evaluation process if models are to be built around this concept with the expectation of having applicability in Specific future product situations . Basically this research seeks to: 1. Classify product life cycles of both sales and profits for new industrial chemicals. 2. Identify the important structural characteristics for new industrial chemicals. 3. Examine historical data of new chemical product histories for predictive content. 6 See Bertram Schoner, "The Selection of Research and Development Projects," (unpublished Ph.D. dissertation, Graduate School of Business, Stanford University, 1965)- 7Ibid, p. 78. '1 -.~ ' u. .1- ... ..A .nu . ‘ .. -.. ~ .- - o o ., _ b.-. . .—.. ~,_' . ~. .: 5 --l -\ “-.. ‘ u-u‘ '. v.‘ e. "I., '~.._ ... '- ..- .. _ .. . 9v 'o l. '. - .. ‘u ‘. ,. .. .i n. \ . y. . i . \ § '\ . v. . ~Q ‘. L' .~ 3 . . I 1 I 16 Since chemical firms make large research and deve10pment outlays, historical performance should be closely scrutinized to find patterns that may lead to more refined inputs for better executive judgment in the future. Signifigagce of the Study This research focuses on the relationships among product life cycle theory, financial planning, and the performance of selected new industrial chemical products. Admittedly, a juxtaposition of various product classes in the analysis could have proven valuable, although the research design required for such comparisons would be burdensome. A design of the present type could apply to other product groups, however, with few basic modifications. The basic groundwork has been laid. Both financial and marketing planning vitally affect overall company performance, and will likely expand in the future. Corporate planning has broadened from its short term emphasis to include long-range planning. Even though sales have been rising in recent years, the chemical industry has shown declining profit margins and returns on investment.8 Because of the rising trends expected in new capital outlays and research and development expenditures, then, operating personnel and management have an even greater responsibility to improve techniques in product analysis . One purpose of this dissertation—to provide a suitable classification scheme for new industrial chemical product life cycles—requires that we identify representative sales and profit patterns (both their shape and timing), and numerous product relationships as well. This is important if one is going to provide adequate explanations of typical sales and profit patterns. The product life cycle itself provides a useful framework for noting changes in operating characteristics which affect the level of Performance. A related goal of this study seeks to evaluate empirically Specific financial and marketing concepts that are relevant to effective new product evaluations. The determination of the minimum apprOpriate M 8Actual documentation is given in Chapter IV, Method of Data Collection. 17 time horizon, i.e., the shortest period of time required in any analysis to reasonably predict success, illustrates one useful output of such research. Need for the Study Since the product life cycle concept was first introduced years ago, many have examined and re-examined its implications. Most often it is proposed as the basis for long-range planning and market strategy. Yet before the concept may be Operationally employed, many functional relationships between sales and time should be examined for various product classes. It may be conjectured that the sales and profit patterns and also the timing of the product life cycle vary on both a product and an industry basis. Yet few empirical studies have described existing patterns. No classification scheme exists for product life cycles in any product class. This research study seeks, then, to fill this void for new industrial chemical products. Executives often feel intuitively that adequate explanations are available for most individual product life cycles. A noticeable decline in a firm's sales may be attributed to competitive product introductions. A product may be de-emphasized or completely withdrawn if customers integrate backward. Such explanations can further our fundamental understanding of product behavior. And documentation of structural characteristics will serve that end as well as providing a firm basis for input in statistical analysis, which can determine the extent of predictive content within the collected data on new industrial chemical products. As its ultimate contribution, this study will use structural characteristics to try to identify product life cycle distributions which in turn can be used as inputs in a capital budgeting model for new product proposal evaluations. .. . ' 5 “won. . . . ‘ . “4. v--. . ... .- '~ - ' ‘ba ' ‘n. . ' a "" - ’-.: \ -,_ v ._ ._ ‘. .I CHAPTER II REVIEW OF SALIENT LITERATURE ON PRODUCT LIFE CYCLE THEORY Introduc t ion Actual efforts to deve10p mathematical expressions for demand patterns have been going on for decades. Because the ultimate results of product management often comprise some type of quantification, their interest may well surpass the simple exercise of fitting historical data to actual products. In measuring mathematical trends, any researcher looks for communal data suggesting an apprOpriate, representative trend. Ideally one examines the patterns of a product for phenomena which may be translated into possible shifts in future demand, if any, once all the developments which might affect the outcome are evaluated. In forecasting, any mathematical expression of demand defined over time ends up with a projection of demand some time in the future. So we should examine some of the mathematical functions commonly employed in business applications to better understand why the product life cycle described in the literature has such wideSpread appeal today. In terms of prediction considerations, the problems center on forecasting technological change and market acceptance. Linear Function Aggregate demand assuming the form Y = a + b(t) , expressed in common units, defines a linear trend, the simplest method for fitting historical data on an industry-wide basis over time t, where a and b are calculable coefficients. When the time span is short enough, this technique can be quite useful for demand forecasting. 19 FIGURE 2"l LINEAR MODEL DEMAND TIME Although the graphical representation for this model is unquestionably linear over time, it actually suggests a declining but positive growth in demand over the projected life Span of the industry, since the units in the base year are increasing through time. With marketing forces and technology influencing market deve10pment as much as they apparently do, demand fitted as a straight line for prolonged time periods may fail to represent actual behavior for long-range planning activities. Exponential and Logarithmic Functions In a number of business situations involving growth and decay the natural mathematical model assumes an eXponential function: for example, interest rates continuously compounded. And the literature often describes long-range projections of basic economic data for specific industrial sectors in terms of annual growth rates. A graph of the mathematical function Y =-‘ (Y ) (l + r) easily demonstrates such non- t-l linear growth, where r equals the calculated growth rate expressed as a fraction, as depicted in Figure 2-2. 20 FIGURE 2-2 EXPONENTIAL MODEL "a" @MAND r—J / / TIME Th . . e logarithmic form of the compound growth curve log Yt = 10g Yt- + log ( 1 1+ r) becomes linear when graphed on semi-log paper, which permits the easy display and use of data having an underlying exponential trend, i.e., ConStancy 0f relative growth. Any researcher reporting long term annual growth rates assumes that demand is an increasing function over time. The Partic . . ular Case Of Continuous compounding over kt years is given by the equation = kt . Yt Yoe . and takes the same general shape of Figure 2-2 above. Since a . n equation of this type follows an exponential pattern, it too aSSUme . S a Straight line on semi-log paper. 0 A prOdUCt having a sizable pent-up demand can experience high Initi a1 aCCeptance and rapid growth immediately following its market introd . ““10“: particularly if the communications program is sufficiently effeCt' . We and the firm has adequate production capabilities early 1“ the PTOduc t f“fiction ‘ life Cycle. This demand pattern, also an EXPonential . assumes the form Y = y (1 .. em) as depicted in Figure 2-3. t max \ \K \ /\ 21 FIGURE 2'3 EXPONENTIAL MODEL "s" @MAND / TIME The rate of increase for any product situation is determined by the term B. The above equation may be supplemented, if a decline phase extends beyond the maximum demand level, by using a rotated parabola function. 2 A second degree parabola fitted to data (given by Y = a + bt + bt ) can aSSume the demand pattern graphed in Figure 2-4. FIGURE 2-4 and DEGREE PARABOLA F‘ DEMAND "' ~\ \ TIME 22 Even though higher order polynomial equations may be calculated to follow virtually every variation in data, such patterns seldom represent long term trends; such an approach for long-range planning purposes would likely meet questionable success. Gompertz Curve Also applied in actuarial science, the Gompertz curve is one growth pattern purported to be typical of industrial deve10pment. This type of growth is found in industries whose product uses are directly related t0 the growth of population, assuming that purchasing power is no obstacle t b , where to product movement. The natural form of the equation is Y = ka- a and -b are constants between zero and one; its logarithmic form is 10g Y: 108 K 'I' (log a)-bt. The rate of growth over time is not constant, appearing instead as a decreasing quantity. The term log K is the 108<':11'ithm of the maximum value approached by the function as it becomes aSYmPtOtiC. Figure 2-5 approximates the pattern of the Gompertz curve. FIGURE 2'5 GOMPERTZ MODEL DEMAND ’ TIME 23 Because product movement is dynamic in nature, one expects demand to shift as the relative influences of demand determinants change. If a product demonstrates competitive strength, the growth phase may predominate in its life cycle; otherwise, its displacement by competitive products eventually imposes a declining pattern upon overall product demand. Many consumer and industrial product categories have experienced market displace- ment in the past, e.g., recent government statistics indicate decreased shipments for the following products during at least one reported year since 1955: canned fruits and vegetables, metal household furniture, primary zinc and explosives. Some products bounce back with renewed strength, while others sustain continued declines. Products displaying a satisfactory Gompertz trend historically seldom behave as the illustration indicates once the curve approaches the asymptote. The failure of this curve to describe eventual increasing or decreasing consumption as shifts in demand pinpoints its serious limitation for long-range forecasting. CO_nventional Product Life Cycle From our limited knowledge of published product behavior, we may describe a generalized aggregate industry product life CYCIe recognizing Six distinct stages, as shown in the next exhibit. Although the number 0f Stages presented in current references which describe the product life CYCle concept differs, the traditional pattern of industry demand ov ' ~ . . . . . er time is very similar to that represented in this section. \ Actual reference will be cited later in the text- DEFINITIONS OF TERMS USED IN FIGURE 2-6 Industry Sales: aggregate sales in dollars or standardized units of the product sold by all firms competing in the market in the time periods indicated. Industry Earnings; aggregate net profits after taxes in dollars or standardized units of the product resulting from its marketing by all firms competing in the market in the time periods indicated (assuming an acceptable, uniform treatment of all expensed items after the product's launching to derive the calculated figures). Research and DevelOpment Expenditures: aggregate costs in dollars or standardized units for applied research and deve10pment work done by all competing firms and directed toward the commercialization of the product. Accumulated Total Investments: accumulated investments through time in dollars or standardized units by all competing firms in new fixed assets, transferred fixed assets, rolled-back capital and working capital, \3 24 FIGURE 2'6 CONVENTIONAL PRODUCT LIFE CYCLE INDUSTRY SALES NDUSTRY \ W'\ K, I \< EARNINGS I I \RESEARCH 8 DEVELOPMENT EXPENDITURES I L \— I I ACCUMULATED / INVESTMENTS I I DISCOVERY PRODUCT MARKET -2 INTRODUCTION GROWTH MATURITY SATURATION DECLINE l r 7 T l I - I O I 2 3 4 5 6 7 8 9 TIME PERIOD 25 Although these sales and profit patterns were envisioned for consumer products, they may not necessarily describe industrial products; certainly industrial goods usually have longer cycles. DevelOped in order, these stages include: 1. Discovery Any product succeeding in the marketplace can be traced to its origin as an idea. DevelOping any product from the idea inception through field studies is usually expensive, but essential if the firm is to effectively evaluate its potential. Innovative firms must obviously risk committing large expenditures in research and deve10pment characteristic of this stage. Once the product proposal survives a number of systematic management reviews, a decision must be made on its commercialization. Will the product cash flows be sufficient to recoup the investments required to market the product and still allow an adequate return on these investments? If indications are promising, funds may be committed to initiate production capability. Late in the discovery stage, the innovative firm makes the initially modest but vital investment in plant and equipment for the production of the hypothetical product. The key factor in the discovery stage is to organize a product development program resulting in the desired product. 2 . Produc t Introduc t ion The product passing through a low volume introductory stage must create an awareness and garner subsequent demand. Consumer education may be required if the product has unique applications or physical prOperties, along with considerable expenditures in the promotional campaign for consumer products, if applicable. New industrial products must nearly always demonstrate technical superiority or cost reductions over competing products before users change their buying patterns. n. - - ".0... ‘ i -----... "I~t I ’ "...... -O -. I 1 °. 0 . u. ‘ I. ‘v . l -. ..fl I“ 26 3. Market Growth During the so-called take-off stage, the product records rapid increases in consumer acceptance as a result of the firm's initial efforts to promote the product and its purchase, with adequate distributors and reliability in performance the two crucial factors. ‘Where large potential markets develop, entries by competitors attracted by this potential often occur in the latter part of the growth stage. 4. Maturity An increasingly competitive environment forces the firm to alter its market strategy to check its declining profits. Various forms of nonprice competition can be introduced to insure umre effective coverage, or the firm may attempt to differentiate its product through quality changes. Repackaging may help make the product more appealing. Advertising programs can be altered to more effectively segment market types and to influence both other-brand purchasers and nonbuyers. By shortening distribution channels, the firm can gain greater control and broaden product exposure. 5. Saturation Prices weaken in the early part of this stage because the industry's capability to supply exceeds consumption. Though price is one important element, other operating variables are affected by competitive pressures; for example, new distribution channels are often needed to provide greater economy. This stage calls for an even greater emphasis on.developing more effective marketing programs. .‘.. .D~ .-..-.». o.q ..- I .--.~..u ' . ‘- .- ’ e. "e ... ‘”4- .._ \. . \ . ”'N ... v .. ‘e. r ‘--. .. ¥ "v .‘ 27 6. Decline Whenever demand declines, sales decline as well. Through product improvements and technology changes, old products become obsolete and are eventually displaced in the marketplace. Cost reductions forestall rapidly declining profits. The only alternative is to liquidate, consolidate, or diversify, and thus we might expect a trend toward concentration among producing firms. The sales and profit patterns in the product life cycle of a firm can differ significantly from those of the industry, depending on the firm's time of entry into the market and its effectiveness in capturing a dominant market share. If the firm was the innovator, the shape of its curve in the early stages would be the same as the industry as long as it constitutes the industry; thereafter patterns of sales and profits are shared among competing members. Even when industry profits are increasing, the innovator could suffer declining profits if his position has been weakened. Many sales and profit patterns are available for firms that enter after the market develOpS. But after the market becomes saturated, it is far more difficult for an entering firm to succeed. Product behavior is most likely to vary radically between firms for such reasons, regardless of industry trends. Even though a firm may produce a product having rapidly growing market acceptance, this gives no assurance of instant success: some products (and some firms) will fail. But all products have life cycles, though they may vary in length and magnitude. Eventually every marketed product will be displaced; and if business executives can remember this, they can prepare more readily for such transitions without their develOping into real crises. The analytical screening of new product proposals requires a multi- dimensional approach. No treatment in the literature adequately handles the selection process. So the search for methodologies to take into account all factors influencing new product performance continues. 28 Possible Applications of Product Life Cycle Theory Identifying specific life cycle patterns of sales and profit contributions may dramatically further efforts to model new product behavior. But before such possibilities can be explored, we should review suggested applications of product life cycle theory and their implications to product management. 0 As a management tool in launching a new product. Projections of the slope and duration of a product's life can be used in product planning to indicate possible marketing strategies that may be employed after market introduction to lengthen lead time over competition. 0 As an evaluation technique in directing research and development efforts. Any innovative firm having broad-based research and deve10pment capabilities can increase its new product success rate by examining the growth rates of various industries during several stages of market development. Research activities could then be limited to those product areas in which potential ideas can be transformed into marketable products at the time when rapid growths are expected. The timing of the research effort can easily affect eventual success, since it partially determines the timing of a new product's entry. 0 For screening a firm's existing product mix. It is difficult to measure analytically the effect of one product's sales on the sales of other products within a given product line. Yet sometimes individual Products become too costly to maintain on the market and should be eliminated. The product life cycle concept focuses on the relative profit contribution of any product in various stages of its deve10pment. 2 Theodore Levitt, "Exploit the Product Life Cycle," Harvard 3.9319888 Review, Vol. 43, No. 6 (November-December, 1965), p. 84. ' 3Philip Kotler, "Phasing Out Weak Products," Harvard Business States. Vol. 43, No. 2 (March-April, 1965), p. 107. V s. s .I a n .- g I ‘. ." . .\ " ' .... 'V. — , ._' _ V n. ‘ ._ . ... .;u . ,‘ . ~‘q - _. - ... ‘ c, . ._’ "n “-. . u ‘1 v.v I .l 29 e As a framework for extending the life cycle. The growth stage of a product may be stretched by any of the following strategies:4 1. Promoting more frequent usage among current users, 2. Developing more varied usage among current users, 3. Creating new users by market eXpansion, and 4. Finding new uses for the basic material. 0 As a key for deve10ping Optimum marketing programs. The stage of market development may indicate the type of marketing effort that a specific new product requires. In a rapidly growing industry where consumer acceptance is Spontaneous, for example, there is less need for firms to implement strong marketing programs. 0 As a basis for scheduling new product deve10pment programs. The profit cycle is important in timing research and development programs, eSpecially if the firm sets its primary objective as earnings growth. Often profits begin to decline before sales, which is when additions coming from newer product offerings should seek at least to match EXpeCted profit declines.6 o For modeling various facets of industrial dynamic processes. The stage of product deve10pment at the firm level affects order rates, number of unfilled orders, delivery delays, capacity changes, and their interactions . 7 0 As a means of selecting advertising strategy. One research study indicates that advertising is far more effective in the earlier stages of the cycle in terms of impact and productivity, concluding that, in the decline stage, advertising cannot significantly lengthen the life cycle.8 “— q 4 Theodore Levitt, 0p. cit., p. 89. 5 Frank J. Charvat and W. Tate Whitman, Marketing Management: A Qllglgzgitative Approach, (New York: Simmons-Boardman Publishing Corporation, ’ p. 130. 6 C. Wilson Randle, as reported in (no author), "5118888” Profit Cycle to Plan Product, R 6: D," Steel, Vol. 154.,N0. 20 (May 18: 1964): Po 31- 7 o . Ole C. Nord, Growth of a New Product; Effects of Capacity- mfiition Policies, (Cambridge, Mass.: M.I.T. Press, 1963), p- 3- 8 P C. Wilson Randle, as reported in (no author), "Key Ad Spending to lrofiI; gyde’ Not Sales Cycle: Wilson," Advertising Age, Vol. 35, No. 17 . o I - u no u... w -.u.. l \ . ... ..... .... -» ‘ “in. -~ .. . r. \ u ,_ ‘..' .". ' a..- , v . s..‘ . .... ~ 30 In summary, the product life cycle concept should provide a useful instrument in marketing research and financial planning. But it is little used in industrial situations because we have lacked empirical documentation of its value. This research has been designed to investigate the feasibility of modeling product life cycles for one class of products, new industrial chemicals. If specific characteristics common to classification types of life cycles could be isolated, they would promote our understanding of the complex interrelationships in new product behavior. Once we establish any stable system of product performance, we can make a sound case for using the product life cycle concept in forecasting demand relationships in a capital budgeting model for screening new product prOposals. “In ... .. r” - J..- 7I ‘ CA ~v... - --¢ v: ... nu. .. . \.. .. A . , u... . . . . _ .- ._ ‘ -. p s CHAPTER I I I REVIEW OF SALIENT FINANCIAL CONCEPTS IN EVALUATING NEW PRODUCT PROPOSALS Introduc tion Most recently management has come to subject more and more decisions affecting corporate performance to critical review. Proposals should be screened for conformity with corporate objectives, requiring some quantitative evaluation of the impact such expenditures may have on corporate performance. Judging from past trends, we can assume a continued growth of quantitative techniques in finance and marketing, both in short- and in long-range planning. Increasing Emphasis on CapgtgggBudgeting Capital budgeting techniques have principally been applied to the analysis of plant and equipment funds committed to major projects: they usually direct major corporate decisions on investment strategies. These basic techniques are process oriented, however, lacking specific product direction; and any attempt to expand their applicability to decisions regarding recognized market opportunities represents a new point of daparture. The demand estimates required as inputs in capital budgeting models are often slighted because of the relative difficulty of estimating demand. The chemical industry has, in fact, been peculiarly limited to the process orientation, since in many cases intermediate processing steps are required before a finished chemical product results, and it has been traditional to roll up manufacturing costs around the processes involved. Any required process may produce economically important COproducts, in which case no single chemical product could be aPPI‘Opriately evaluated by itself. 31 32 Rapid technological change often renders products or processes obsolete. It is unreasonable to expect any product to find a future market niche because of past performance. So, the emphasis should be on filling market needs as they become known. Establishing profit centers at the product level seems one of the best ways to achieve the proper orientation, yet such a scheme requires that various functional personnel such as researchers and marketers have a voice in product management. Applying capital budgeting techniques can clearly promote an understanding of product behavior and aid in the selection of investment strategies at all Operating levels. Costs and Investments Considered All applied resources, physical and human, have associated costs and, theoretically, these can be assigned on a direct or allocated basis to any given product or family of products. (Where c0products are produced in any given chemical process-assuming all have economic value—we can evaluate the impact both on the individual products and on the total.) Research and development costs incurred prior to the decision to build plant capacity for a chemical product are sunk costs, so costs prior to the first sales year are usually charged against the associated corporate or divisional accounts. Management decides whether or not to provide production capacity by projecting future income streams; in this, it is logical to exclude all previous sunk costs. Any further research and development costs during the life of the product could be treated as Part of the product evaluation process and amortized appropriately or, as the practice may become, expensed at the time of actual expenditure. Expensed items and investments are conventionally distinguished on the basis of longevity in use. Current legislative interpretations delivered by Federal Government regulatory agencies determine the kinds of expenditures that can be charged off as costs to expense, and those that must be capitalized and written off through asset depreciation. The major expense items within the chemical process industries include raw material, labor, energy, quality control, insurance, start-up costs, u ......u- u- .. "sum-I1 VII .- .. iv .UI i .. All . -.. a u... I. ‘ n‘ '--. -. “fizgn 4": .- ‘W... ' ..‘ 33 marketing, research and deve10pment, and general administrative charges. Investments, on the other hand, are basically applications of funds to a specific use over a longer time period. Since the products in the sample vary in the lengths of their investment and economic life expectancy, some attempt should be made to take these factors into account, both in accepting and rejecting new products, and in ranking established products by performance. Even though the data contributed by the reSponding companies were as accurate as possible, varying accounting practices among firms in treating all the many fixed and variable costs associated with product behavior, including depreciation policies and deve10pment expenditures, make it unreasonable to draw exact comparisons. Having made every effort to standardize the treatment of accounting information for communality in experience, this writer believes the product histories here covered do indicate actual behavior, but only in an approximate sense. Investment capital may be distinguished as fixed investments and working capital, with fixed capital conveniently divided into new, transferred, and carried-forward fixed investments. New fixed investments for plant (buildings and prOperty) and equipment were valued at actual outlay prices. Transferred fixed investments for plant and equipment, evaluated at replacement cost, cover existing fixed capital shifted to the production of the new product under consideration. And carriedoforward fixed investments include that existing fixed capital used in producing raw materials or chemical intermediates which are, in turn, used as inputs for chemical processes resulting in the production of the new Product under study: these investments were valued at actual outlay prices or replacement values, depending upon whether the equipment was new or transferred. Working capital investments include raw materials inventory, work in process, finished goods inventory and credit allowances. Actual working capital requirements vary radically with the nature of the Product and with established inventory and credit policies as well. -avv a .. .... .... . . ha... . .. ..., ." x. q I ~. ‘. ~- ._“‘ -. -.. a . ‘ag \ q... _- -., . s e r” .-.~ ,' , ,. ' ‘. “ ,‘v . .- ‘. 'o ,_ \. . '0 . .. 34 All initial product investments began one and one-half years before introduction, requiring 18 months on the average to establish a production capability within the chemical field. Whether or not the product was produced internally, first year sales marked time period one, with all subsequent changes in investments recorded on an annual basis to reflect increases or decreases in overall investment. Depreciation Policies Considered Non-cash depreciation directly affects the actual level of taxation in business operations. For various forms of accelerated depreciation (e.g., the sums of the years' digits method) act to reduce immediate tax payments and increase distant tax payments: their advantage lies only in the concept of the time value of money—a dollar today has more value than a dollar received any time in the future. Since depreciation figures affect net cash flows, they should be included in investment analysis. Again, widely divergent forms of depreciation calculations among the contributing firms studied necessitated our developing a standardized approach to render product inclusions comparable. Though any technique would be more or less arbitrary, the method presented below at least takes the capitalization process into account. Seeking to rank products against various performance criteria, we may use any technique so long as it is applied consistently in all product analyses. 1. A straight line depreciation policy is used, assuming a ten year economic life for all fixed assets. 2. Whenever additional investments are made after the first year, the nondepreciated balance is released at the end of the analysis. 3. If the product remained stable in the marketplace at the end of the last time period in the study, and the number of recorded time periods for sales is less than the maximum possible of ten, future sales and investment requirements will be estimated by extrapolation to the final period by using its predetermined polynomial or other apprOpriate mathematical function. II- 35 4. If the product had been withdrawn at any time from the marketplace, all depreciable fixed assets not yet depreciated, which supported the production of the product, are credited to its cash flows as released assets. Salient Financ ial Concepts The many evaluation techniques suggested for investment analysis variously affect the measurement of immediate and, more importantly, long-range performance. Although pure survival matters a great deal, of course, the accelerating pace of market development and technological change make the relating of actions with pre-established objectives more essential. Each of the techniques here presented, then, should help in analyzing the performance of products included in the research. As we have come to expect, investment accounting practices vary widely among major chemical manufacturers, sometimes. even within a given divisional operation. So each product investment schedule was considered individually in an attempt to put all investment within a common framework. Book or net investment figures were avoided as purely accounting conveniences irrelevant to measures of performance. 1. Cost of Capital Any discussion of evaluation techniques based on discounting Procedures should consider the cost. of capital concept, its meaning and calculation.1 Actual costs of capital for select firms over Specific time periods are measured as the first approximation of the discount factors that management must face in new product investment decisions. For many sources of funds can finance investment decisions, no one. financing source necessarily funding product decisions for large manufacturing firms. Usually, in fact, a Variety Of capital funds Support 1 One procedure for calculating the weighted cost of capital is Sgggested by J. Fred Weston, Managerial Finance, (New York: Holt, Rlnehart and Winston, Inc., 1962), pp. 226-249. , n - a .. ... ‘ ‘ i I ran».-. ... -‘" ‘r‘ .4 . . u. " ‘lu~-.n .. ... ‘ '<~ ...; ‘ . , '~. o, ‘ g ‘~-...‘. . I‘I . ‘ h" ‘a- .. . "w .1 '\ §_ ‘ I‘. v.‘ , - - 36 a composite of new product prOposals. Indeed, since major industrial chemical manufacturers typically face many new product decisions during any planning period, no single funding decision will grossly affect either the aggregate leverage position of the firm or its overall cost of capital. The marginal cost of capital constitutes the relevant discount factor in calculating present value measurements for proposed ventures. The average cost of capital is traditionally assumed to be constant over small nmrmmnmal changes of leverage that a firm employs at any given time, thus, the marginal cost of capital equals the value of the average cost of capital .2 ’ 3 For our research purposes, a firm‘s relevant cost of capital shall take a weighted average of the costs of each type of capital, from all financing sources, with the basic model as follows: KO=W1K1+W2 K2+W3 K3 where K01: overall capitalization rate of operating earnings, reflecting both business and financial uncertainties. W. =weight of a given class of fund i based on the market value of that source relative to the market values of all financing sources for the firm. K1 = capitalization rate of short term liabilities. K2 -- capitalization rate of long term debt. K3 = capitalization rate of equity capital. See Myron Gordon, The Investment, Financing and Valuation of the Corporation, (Homewood, 111.: R. D. Irwin, Inc., 1962). —_J An excellent, controversial discussion of the firm's financial structure and its effect on the cost of capital can be found in F. Modigliani and M. Miller, "The Cost of Capital, Corporation Finance, and the Theory of Investment," American Economic Review, Vol. 6.8, No. 3 (June, 1958), pp. 261-297. .... ... Iv va l. .0. u . ~- ... a .. ~. . .. . . ...“ ‘ “l D " D. ... ..‘ . . _ ‘ . O ‘ .- v “~.. ’. . ’ ‘ .... _ 'I» _ . ‘.§ 37 Largely a future oriented concept, the cost of capital should specifically consider what the firm is facing at the time the decision to reject or accept any new product prOposal is considered as well as any changes the firm may face in the future. Any quantitative measure of the cost of capital may theoretically range from a low, equal to the prime rate of money, to infinity. A. Cost of Current Liabilities Short term debt, one source of capital available to firms and having an associated cost, is seldom included in any determination of cost of capital, perhaps from its relative unimportance in many situations. But, since it affects the utilization of all resources of the firm, it does affect the cost of capital and should be included; taxes payable, wages payable and other noninterest-bearing current liabilities are generally excluded from the calculation. For our purposes, the applicable rate of interest used for any year is the rate of commercial paper (4 to 6 months) at current rates deflated by the apprOpriate tax rate. Interest payments are tax deductible, so the effective cost of interest-bearing debt. depends on the existing tax rate. B. Cost of Long Term Debt Dependent on the type of fund, long term debt is measured at current rates using market values and interest rates applicable to that type of financing for any given year. If market values are unobtainable, coupon or bank rates are substituted in the calculation. Again, all interest rates are deflated by the apprOpriate tax schedule. C. Cost of Equity Capital For our purposes, the capitalization rate associated with the equity portion for any period is based on market values of all types of stocks outstanding, assuming that current market prices reflect investors' evaluations of the firm's commitments and capabilities not only on an 38 immediate operating basis but for the distant future. Since retained earnings imply a cost to the firm, they are in fact regarded as a type of common stock investment; retained earnings contribute one source of funds and hence bear an Opportunity cost. From an investor's viewpoint, of course, the source of funding is independent of expected returns; any normal returns from the use of retained earnings justify their use. The rate used in assigning costs associated with preferred stock is the coupon rate, that is, the effective rate acceptable to preferred stockholders at the time of purchase. The apprOpriate discount rate for common stock probably should not be based on any current price-earnings relationship, especially for growth oriented firms (as many chemical manufacturers are). So, to provide a more realistic approximation to the true capitalization rate associated with common stock, the long term return on the common equity measurement is chosen because it reveals earnings growth yield on the common stock after risk valuation. It does not reflect the present situation, but measures eXpectation. The median return on common equity for the five year period after market introduction is actually used as the appropriate capitalization rate. In terms of our information requirements, all common equity measurements are known, so no projections are necessary. 2. byback Per iod The number of years required to recover the initial depreciable fixed investment in plant and equipment, defined as the payback period, is probably the most widely used technique in ranking investment decisions in industrial situations. The simplest payback period—a lumped investment and averaged incremental cash flows—appears algebraically as:5 4 This approach includes an approximation of the impact of growth provided by retained earnings. The problems encountered in measuring the cost of equity capital are discussed by E. Solomon, The Theory of Financial WEE: (New York: Columbia University Press, 1963), pp. 69-78. *— . 5The equations defining those financial techniques actually used in the analysis of collected data are noted with an asterisk. 0", b I: ‘v'. 't ‘-I ., _ \31 '4 1: : ‘ ... . . T' i = ' .— 313:! q 4 I A h. o.__ ‘-.I 39 (3.1) where PP = payback period in years IL = lumped investments required before market introduction n 2 GOP. J L— = mean operating cash flows (after taxes), i.e., n average net profits after taxes plus depreciation, from time period 0 through n. In nearly universal use whenever liquidity (rapid investment recovery) is a meaningful financial objective, this simplified technique does not involve entire economic lives and income streams, but only those parameters for the time period essential to recoup fixed investments. This time concept ignores the temporal patterns of cash flows, however, as well as possible important contributions beyond the calculated time period, and thus fails to measure profitability prOperly. lt seldcm accounts for working capital requirements as investment inputs. Furthermore, the payback period may not provide a good indicator of risk. The risk of not getting production started on schedule due to technical difficulties in equipment design would all but be eliminated after the production system 1s tested and proven operational. The risk associated with competitors lowering traditional price levels at the time of product introduction would be resolved based on actual behavior. The risk of incorrectly assessing the impact of advertising media programs on product acceptance can be significantly reduced once checked against actual sales records after the product is launched. In many similar situations, some risks are reduced after the market introduction stage. 80 the payback technique . ... ,..- 0v e . ...-- ”qu .v~- .... .-.,. 0- p- >: u b! . .. .. __' 'rh. ‘ “bu u um. . "N ~ 1.... -. ..,... ." a . '*'--.n.. 'n H" \ ..-... ..- -- . ' o . .... .. m ‘ s . § § . c I u I 40 will demonstrate a bias toward investments that have disproportionately large cash flows early in the product life cycle. The payback period concept may be redefined in situations where investments vary over time, especially important in chemical processes where investment totals are sensitive to varying levels of output because Of carry-forward investments associated with raw materials. Working capital is generally ignored, with investments usually defined as fixed capital only. Where multipurpose equipment is used in the noncontinuous production of chemical products, investments can be distributed equitably to all products on a time-in-use basis. A modification of the simple payback equation incorporates a change in the definition of the investment variable. IC PPT = *(3.2) n \j‘ OCF j = 0 n where PPT = payback period in years IC = maximum accumulated investment figures from time period 0 through n OCF. = yearly Operating cash flow in time period j (after taxes), i.e., (Sales - cash charges - depreciation) (l - TX) + depreciation, where TX = prevailing tax rate. ... , cocoa 41 3. Accounting Rate of Return This financial technique seeks to measure relative product profitability, though there is little apparent agreement about which of several alternate procedures is best. For products where investments only occur in time period -a, common definitions include: EAT I _ t = l ARR — x 100 (3.3) 1 l t = -a n X EAT J j = 1 n ARR = x 100 (3.4) I t = -a n E EAT J J' = 1 n (0.5) I | t = -a Where ARR = accounting rate of return EATj = earnings after taxes in time period j I t i -a = required fixed investment at the time of initial authorization in time period -a n = number of periods in the analysis 0“ Ital-- 2.... 5.... I. In ...}. .t. .....l c . ‘ "- a . , ."f‘ “-.-. Ian ..I.‘ *- _ a L. , ‘ h h ‘ ‘ ‘ . I a” 42 The first formula takes into consideration only book profits for the first period, ignoring the contribution of future income streams; and equations (3.4) and (3.5) ignore cash flows as well as the time value of money. In no case, here or below, are investment changes through time taken into account, again favoring products having relatively high initial earnings . For products having distributed fixed investments, the denominator could be altered to define the period of maximum investment, though this does not change its limitations. n E EAT. J J' = 1 n ARR = ‘ x 100 *(3.6) (0.5) 1 I t = C where ARR = accounting rate of return EATj = earnings after taxes in time period j n = number of time periods in the analysis I . t = c = maximum investment in time period c 4' Mm on Investment Analysts within the chemical industry often employ the rate 0f I‘Eturn investment calculation, a limited concept by definition in the Sense that it only measures a product's degree of Pmfitabiuty against it - , s Immanuel“ rate for a given time period. Its formula is shown typically as: ... - — I - .. ... ‘u — ...-- v . - I \- k__ .. e o.l u i u “u, u- I. — .._ .. a ‘- y . .‘ N -. \ u v I. ‘4 '~ ‘ a 1 u 1“. N | .u .- ., '~ ,‘:' a v -. 1 -. 43 (EAT)j ROIj = x 100 *(3.7) I J where ROI = return on investment for time period 1 expressed as a percentage (EAT) = earnings after taxes in time period j I = accumulated investment in time period j Because cash flows vary widely throughout the life of a product, calculated ROI values may vary widely. As will be evidenced in the research findings, one cannot accurately project performance during the first several years in the case of industrial chemicals and expect it to be typical. Apparently one should really project demand and cost relationships far into the future when evaluating product prOposals; in which case the most representative return on investment figure would be its median. But other financial techniques yet to be covered are more exacting and informative. This measure obviously ignores the timing of cash flows, possibly a serious limitation. Such a subtle distinction may indeed be significant when one must examine a host of different attainable combinations in selecting a composite set of new product opportunities for market development based on profit Optimization. 5. Internal Rate of Return The internal rate of return, defined as that discount rate which will equate the discounted earnings cash flows and discounted investment cash flows over the product life cycle, can handle many of the limitations of other methods for evaluating product proposal work. It allows for Varying product lives, taking into account the time value of money; and unlike the present value calculation (to be covered later), it makes no assumption about the exact cost of capital. Where the initial investment is fixed in size, the expression appears asz 44 n (EAT + D)j I-a - Z O, and solve for i (3.8) j = o (1 +1)J where I-a == lumped investment made in time period -a D = depreciation charges (EAT + D)j = total cash inflow in time period j i = internal rate of return expressed as a fraction to the base 1 JED And where investment timings vary throughout the product life cycle from time period -a to n, the formula appears as: n FEAT + n) - 111 E: = 0, and solve for i (3.9) j = -a (l + i)J where i = internal rate of return expressed as a fraction to the base 1 [(EAT + D) - IJJ = net cash flow in time period j j S. n Depreciation figures are necessary inputs in calculating the internal rate of return for they are used to determine tax liabilities. Any formula that considers the time value of money, such as the internal rate of return, includes with investments all fixed and working capital, so the timing of cash flows should be a relevant consideration. Marked changes in yearly investment totals are likely to occur through .- ' ' n . W a “ us. a, . s ." h. ~n \ ‘O- "I , u’ I 'n. . c. . w. . Du. ‘ In. ‘~ 45 time—which is where mathematical complications may arise. In general, two rates of return are possible whenever net cash flows shift from a positive to a negative figure during the analysis, whether this shift is due to an overall lack of profitability when investment expenditures for plant expansion exceed earnings, or to the release of any substantial amount of nondepreciated assets at the end of the product life cycle.4 Theoretically, any product can be accepted if its calculated internal rate of return exceeds the cost of capital for the firm, though naturally the cost of capital must be determined to use such a break-off point. In practice, many decision makers choose break-off points somewhat higher than the cost of capital to reflect possible marketing and related risks, depending on the nature of the product. This technique may not allow proper rankings of alternate product proposals having unequal product lives: a product with a 107. rate of return may not be better than one with am 8‘7. rate of return. The internal rate of return method implicitly assumes that cash funds produced by any product are reinvested at the earned rate of that product. Occasionally this assumption is acceptable, as when a higher reinvestment rate pertains. But, in many situations, other discounted financial techniques evaluate and rank product proposals more effectively. All similar approaches still require estimates of performance over a significant portion of the product life history. 6. linesent Worth Method A promising discounted cash flow technique, the present worth method, explicitly handles the common situation facing chemical deve10pment decision makers where patterns of investment and earnings vary throughout the product life cycle. The earnings stream after taxes is discounted at an appropriate discount rate, and the investment stream similarly. Then if a Product's net present worth is positive, that is, the algebraic sum h. 4See John G. McLean, "How to Evaluate New Capital Investments," Harvard Business Review, Vol. 36, No. 6 (November-December, 1958), PP. 65'67 o «- ......., , , o n .. "...... ...n . a 'Iv «\v... nin- . ‘~I ...- n a— n n _ . - - 46 of its discounted earnings exceeding the sum of its discounted investment figures, the product prOposal is sound on the basis of financial data and may be accepted. This method also allows for ranking various new product proposals according to their calculated profitabilities as long as they have equal product lives. The net present value is expressed thus: n (EAT + D)j d Ij NPV = Z -————-—- - Z -———-—— *(3.10) j=0 (1+1)J j=~a (1+1)J where NPV = net present value I = incremental investment required in time 1 period j (EAT + D)j = total cash inflow in time period j i =' predetermined discount rate Most critical is the selection of the apprOpriate discount rate, though the cost of capital concept has been used. The exact meaning of any net present worth figure in dollars remains unknown. Yet deepite this confusion the present value method does attempt to systematically quantify the important parameters in evaluating performance. Problems arise when product lives differ. Taking the smallest Gomon length of all product lives in a given set of attainable Opportunities seems unreasonable for it favors higher net cash flows early in the product histories, overemphasizing liquidity. The simplest approach acceptable to the analyst involves converting present value totals to 47 annualized figures. The expected net profit contribution for any product on an annualized basis, in the present worth approach, would be simply: n EAT + D d I ( )j X j j=0 (1+1)j j=-a (1+1)j *(3.11) n where n = time horizon in years considered d s. n j s. n Associated financial and marketing risks can be accounted for in the evaluation process by adjusting the discount rate employed to arrive at present value figures. Product inclusions would be based on that combination of products which, over a given time horizon, maximize their sum of present values per year of market life, taking into account the usual resource limitation thwarting capital formation: this involves ranking yearly present values, and approximates an optimum selection of product combinations . 7. Equivalent Rate of Return The equivalent rate of return is based on calculated present values of net earnings over investments in a specific time period, and makes the same assumptions regarding product lives and net cash flows. But the performance of a given product can here be related more meaningfully as a percentage return figure, the type of expression commonly used among P90p1e assigned the reaponsibility of evaluating product Opportunities. Unlike the internal rate of return formula, this makes no assumption reSarding reinvestment rates. . I. 9.9' :P on» his I V ‘a ‘w I \ ,“1‘ I “N ‘9- . ‘ml ". ‘I ‘s \, ‘ “' 48 The formulation suggested by Herron is:5 n (EAT + D) d I Z J 5: J 1'0 (1+1)1 1..-, (1+i)j ERR = x 100 *(3.12) d I Z J j =3 -a (1 + i)1 where ERR = equivalent rate of return expressed as a percentage d 1 n j s. n This measure essentially defines the specific return made on cumulative discounted investments through time. The equivalent rate of return inadequately evaluates differing product lives. And since it cannot apprOpriately rank projects by absolute profit contributions even where product lives are identical, it has its drawbacks. Nevertheless, accept or reject decisions based on the sign of the equivalent rate of return are accurate: that is, as long as the equivalent rate of return is positive in value, accepting the product would enhance earnings growth. Higher positive equivalent rate of return figures reflect greater returns an asset utilization, but they may not necessarily indicate Optimum earnings per share results. 5David P. Herron, "Comparing Investment Evaluation Methods," Chemical Easineering, v01. 74, No. 3 (January 30, 1967). PP. 129-130. .. ...» - .o‘ {-0 " ' “h"h‘ Iv... - as. .4 a h a. a a” __ 49 8. Profitability Ratio If one assumes a reinvestment capability for the summed net cash flows at the predetermined discount rate for those products in the analysis having economic lives shorter than the maximum, then net cash flows may be discounted out to the time period covered by the product with the longest economic life. Where multi-purpose equipment has use elsewhere, the released capital outflow is discounted in the fashion described previously. The resultant net discounted cash flow stream is expressed as a ratio to the discounted capital outlay stream.6 The formula relating discounted net cash flows to discounted investment flows for all products having market lives of e periods, but less than n periods, is shown as: P- e d W (EAT + D) ' I e (EAT + D) d I ;0 j E; j Z i Z J + J- i- -a ' j j=0 (1+1)j j=~a (1+1)j (1+1) PI __ L (1 + 1)“ J " d I. I fl 2 J d Ij j= -—a (l + i)‘1 Z + Lj = -a (1 + i)J (1 + 1)n where PI = profitability ratio *(3'13) jEe d s. n e E, n 6Such methodology is adapted from Edgar A. Pessemier, New Product . ' ————__ Decmions, (New York: McGraw-Hill Book Company, 1966), pp. 77-78. ..., .- . .. wwuuuva at. . as .— U '- ' - I - . ' . — ...,_‘ |' u ..., 50 The ratio for products having market lives of n periods is merely the ratio of the conventionally calculated discounted net cash flows to discounted investment flows, using the formula: n (EAT + D)j d Ij Z - Z j-O (1+1)J j=-a (1+1)J PI = *(3.l4) d 11 X j"-' -a (l + i).1 where P1 = profitability ratio d f. n Positive ratios on profitability indexes indicate acceptable product proposals in terms of financial performance, with profitability being a relative function of asset size, after all nondepreciated assets are released. The larger the ratio, the higher the relative profit contribu- tion, in terms of more efficient utilization of capital equipment. 9. Performance Index It is quite possible to combine a number of these financial definitions and to devise analytical formulas tailoring product selection to specific corporate objectives. For example, management may desire to weigh equally the effects of present value and liquidity. One such performance index, PIN, takes into account both of these concepts in the form: .I I III ‘I’~ .lllv Nu‘ 51 n 2: (EAT + D - I)J e EAT + D d I ( ’3 J J - -- z - . j-o (1+1)J j--a (1+1)j (1+1)n PIN = d I.1 d I X j + 2: J j - -a (1 + 1) J "' ‘l‘ (1 + 1)n *(3.15) where PIN = performance index n = longest time horizon in the set of projects being evaluated d g n e :5 n J s n Here products are ranked not only according to their net cash flows but to the extent of total cash generated throughout their market lives as well. The equal weighting scheme in this case was obviously arbitrary and could be altered to suit individual needs. Summary No one financial technique for evaluating product proposals can be considered clearly superior; each has distinct advantages and disadvantages. It is most important to compare product proposals using may different and combined financial techniques for a better understanding of expected performances. Nevertheless, the annualized discounted present value model seems to be most suitable when one desires to rank proposals and select alternatives on the basis of only one relationship, net discounted cash flows. The empirical portion of this thesis will investigate the Similarities and dissimilarities among these described techniques; then, 99mm more appropriate recomendations can be made. CHAPTER IV METHOD OF DATA COLLECTION Selection of Industg The industry to be studied was selected partly according to its apparent level of emphasis on new product development activity: presumably management would respond more readily to this study if it was already stressing new product deve10pment as an industrial strategy. And expecially since contributing firms would have to retrieve the underlying data—an expensive, time consuming process-only those industries having adequate staff and informational capabilities could be expected to cooperate. Certainly the research and development branch of the chemical industry has always been heavily involved in new product development. Furthermore, wherever long-range planning received considerable attention, the capabilities required for internal coordination of this programmed study usually appeared as well. Once the study was underway, we found in fact that respondents Of the participating firms indeed felt that their efforts were contributing to a more comprehensive documentation of product behavior in their own field, generally concluding that such work was indeed worthwhile. As an expression of gratitude for their partici- pation, the respondents were offered a cepy Of the research findings. Definition of New Product Predictably, the firms in the industry differed over what constitutes a new product, since different people had different orientations and responsibilities to one or another specialized facet of Operation associated with new product work. Nevertheless, establishing a definition of a new product was necessary to allow consistent interpretations among responding firms. We shall define a new product as differentiable from existing Products on the basis of composition, structure, form or shape, including adaptive products introduced by firms who did not previously Offer them as an integral part of their product mix (where the product is actually new to the firm, though not to the marketplace). 52 cum ‘ s ...- U... . . o.- o“ u ‘ '- ,,......'~—~ -~ . v w i" . y. .... .. .-o' . a u u ..-. - p9 - av loin hears-a a .....p. . . .. c " .2' \ - ..h .... ... . ......_: | . .- . . , - ---. ...” -. . - ‘ua: v... V . s‘: . . .l. . 1 v a ‘l . ... a ..d‘ 13'“. - \‘s s, .s .l s.__ i . ‘1 I. . .‘ ‘. \ .~. .... _ a. . 53 All products added to existing product Offerings as a result Of acquisition, merger or combination are excluded, because in many of these cases, product programs continue to be implemented by the same basic line management, even after an integration Of capital structures has been completed. The products in question may be new to the parent company, but not to the elements within the company having operational control and primary reaponsibility over product deve10pment work or the marketplace. The fruits of purchased technology may, however, be included. Since we desire to investigate the progress Of only significant new products within the basic chemical industry, we have arbitrarily set a minimum annual sales level of $50,000 for this study. And any new product so qualified retains its classification for five years, another arbitrary limit insuring relative consistency in information, and allowing relative comparability of new product deve10pment experiences across all firms included in the sample. Any product included must have reached this minimum for the first time between the years 1955 and 1960. Most accounting data are considered for reporting purposes at the end of the fiscal year (the end of the calendar year for most major chemical firms). Again to render statistical comparison meaningful, a new product must have been introduced for five complete fiscal periods; any partially completed introductory period preceding the start of a complete fiscal period for a given product will be adjusted to reflect annualized patterns when analyzing life cycle data. Such a procedure allows for meaningful comparisons without introducing any built-in bias. As new products must be distinguishable from existing products on the basis Of composition, structure, form or shape, formulations are generally excluded unless they prove to have been unique either from a marketing or a technological standpoint. Internal consumption must be less than 107., for if large amounts of the product were captively consumed, the make or buy decision would be made primarily from cost considerations. I ....- Is- a a "‘ ' ‘th :--,_.__ p u ...-b .l s. 54 Technological breakthroughs in synthetic fibers and plastics have spurred unusual expansions into new fields within the chemical industry. Yet much underlying growth has stemmed from basic chemical products: 'K . . it must be emphasized that the dramatic growth rate of the industrial chemical industry has reflected and has been made possible by the large flow of new products. The volume of Older products has increased as our national economy has grown in size. But it has been the sales derived from the new organics, particularly synthetic materials, which has made the difference between an about average growth rate and one which is twice as rapid as that recorded by the national economy. However, a continuation Of past trends will depend upon the continued high rate Of innovation which has contributed so significantly to the brilliant record achieved by industrial chemicals."1 Thus it was that only industrial organic and inorganic chemicals were included in the defined population. The Standard Industrial Classification system devised by govern- mental sources was used to identify specific products, thus protecting the vital interests of contributing firms from disclosure Of confidential information. The major group, chemicals and allied products (SIC 28), includes three general classes: (1) basic chemicals, such as acids, alkalies, salts and organic chemicals; (2) chemical products used in upgrading processes, such as synthetic fibers, plastic materials, and pigments; and (3) finished chemical products for ultimate consumption, such as drugs, cosmetics and soaps, or for use in other industries, such as paints, fertilizers, and explosives. The SIC 281 group is a further subdivision of the chemicals and allied products group and includes industrial inorganic and organic chemicals. Any products included in this study, then, can be coded as follows: *— 1 Jules Backman, Chemicals in the National Economy, (Washington, D.C.: Manufacturing Chemists' Association, Inc., December, 1964), p. 3. Standard Industrial Classification Number 2812 2813 2814 2815 2816 2818 2819 55 Inclusions Alkalies and Chlorine: sodium hydroxide, potassium hydroxide, sodium carbonate, potassium carbonate, sodium bicarbonate, chlorine, and the like. Industrial Gases: gases in liquid, solid, and compressed forms, as acetylene, nitrogen, and hydrogen. Cyclic (Coal Tar) Crudes: coal tar crudes, coal tar acids, medium and heavy oil products as creosote Oil, naphthalene, anthracene, and their homologues (coal tar crudes produced in recovery ovens and petroleum refineries not included). Intermediate Coal Tar Products: cyclic organic intermediates, dyes, color lakes, and color toners. Industrial Inorganic Pigments: all inorganic pigments, as black, white, and color pigments. Industrial Organic Chemicals, N.E.C.: noncyclic acids, aldehydes, amines, solvents, polyhydric alcohols, synthetic perfume and flavoring materials, rubber processing chemicals, cyclic and acyclic plasticizers, synthetic tanning agents, and chemical warfare gases. Industrial Inorganic Chemicals, N.E.C.: salts, alums, calcium carbide, hydrogen peroxide, phosphate, sodium silicate, ammonia compounds, anhydrous ammonia, fertilizer materials as muriate and sulfate of potash, rare earth metals (alkali) and metal salts. ~ u... - .... 56 In summary, all new products included in this research design: 1. Must possess SIC Codes 281x. 2. Must have attained minimum sales volume of $50,000 annually for at least one year, reaching this volume for the first time between the years 1955 and 1960. 3. Must consume less than ten percent internally. 4. Must be differentiable from existing products in composition, structure, form or shape. 5. May not be a formulation. 6. Must be new to the firm, though not necessarily to the marketplace. Se ection of Population This dissertation takes major chemical manufacturers for its pomflation. Since an estimated 850 chemical firms sell over $50 million amnmlly,.a speculative listing in any event, each major chemical firm must be listed: 1. In the tOp 500 largest industrial firms (ranked by sales) as compiled by the editors of Fortune magazine. In the directory of companies filing annual reports under the Securities Exchange Act of 1934 as Of 1963 under SIC Codes 281 and 289.3’ 3. 0n available Compustat tape Of Standard Statistics, Inc., at the Computer Center, Michigan State University, East Lansing, Michigan, under SIC Codes 2800 Chemicals, 2823 Synthetic Fibers, or 2899 Chemicals and Chemical Preparations (1966 edition). \ C 8The Fortune Directory Of the 500 Largest U.S. Industrial Gunmat1°ns’" Easiness Vol. 74, No. 2 (July 16, 1966), pp. 230-251. Fil' Securities and Exchange Commission, Directory Of Companies ITEHELAflBual Reborts with the Securities and Exchange CommiSSLOn . JLELJ$E¥Securitie§_§§2hange Act Of 1934, (Washington, D.C.: Office ofStatnMfical Studies, Division Of Trading and Markets, 1963), PP. 117-121. The Securities and Exchange Commission has chosen to include product 8 Conventionally listed as 282 under the 281 code. We thus see fiber .materials, Synthetic resins, synthetic rubber and other man-made 8 “KnudEd- Each industrial firm was placed in one selected category a teradetermination was made of its major line of business. Offering Plastic :— hh-a- 57 The final alphabetical tabulation of major chemical firms, hmxmforth identified as the pOpulation for the study, is as follows:5 10. ll. 12. 13. 14. Two corporations, American Enka and International Minerals and Firm Air Products and Chemicals, Incorporated Air Reduction Company, Inc. Allied Chemical Corporation American Cyanamid Company Celanese Corporation of America Chemetron Corporation Diamond Alkali Company The Dow Chemical Company B. I. duPont de Nemours & Company Eagle-Picher Company Ethyl Corporation FMC Corporation W. R. Grace and Company Hercules, Inc. Corporate Address P. 0. Box 538 Allentown, Pennsylvania 150 E. 42nd Street New York, New York 10017 61 Broadway New York, New York 10006 Wayne, New Jersey 07470 522 Fifth Avenue New York, New York 10016 201 E. 42nd Street New York, New York 10017 300 Union Commerce Building 9th and Euclid Cleveland, Ohio 44114 Midland, Michigan 48640 Wilmington, Delaware 19898 American Building Cincinnati, Ohio 45202 100 Park Avenue New York, New York 10017 633 Third Avenue New York, New York 10017 7 Hanover Square New York, New York 10005 Wilmington, Delaware 19898 Chmficals, were not included because they either had no SIC 281 Code pnmhmtion or had introduced no new products (as defined by the afore- mentioned criteria) within the time period Specified. 18100 5.. § a ... .0 7| 'I ‘ .1. k g 'l .- ; " qz'. ..‘ 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 58 Firm Hooker Chemical Corporation Koppers Company, Inc. Monsanto Chemical Company Olin.Mathieson Chemical Corporation Pennsalt Chemical Corp. Reichhold Chemicals, Inc. Rohm and Haas Company Stauffer Chemical Company Union Carbide Corporation Witco Chemical Company, Incorporated Wyandotte Chemicals Corp. Corporate Address Niagara Falls, New York 14302 Koppers Building Pittsburgh, Pennsylvania 15219 1800 North Lindbergh St. Louis, Missouri 63166 460 Park Avenue New York, New York 10022 3 Penn Center Philadelphia, Pennsylvania 19102 525 N. Broadway White Plains, New York 10603 Washington Square Philadelphia, Pennsylvania 19105 380 Madison Avenue New York, New York 10017 270 Park Avenue New York, New York 10017 277 Park Avenue New York, New York 10017 Box 111 Wyandotte, Michigan 48192 Eggformance Record of Major Chemical Companies The chemical industry as a whole has made substantial gains in physical output and dollar shipments during the last two decades. More progressive domestic chemical manufacturers have successfully broadened marketing participation in above average growth areas, particularly through integration movements, both forward toward final markets and backward toward sources of lower cost raw materials. Balancing the anthNIOf profitable upgraded chemical products against the support of commodity type chemicals helped generate above average gains, both endeavors requiring considerable funding in research and development. Naunmlly merger activities and other factors aided this historical 81'0Wth as we 11 . ... .. .. . 'r . u‘h ... \uv ... . .- V‘ '- (a. “a“ "nu N a.» ‘s- 59 Chemical companies with smaller asset and sales structures have typically been able to grow at above average rates during the 1955-66 period (see Tables 4-1 and 4-2). Reference is made here to the compounded rates of change, not generalizing with reSpect to absolute dollar changes for revenues. Ranking fourteenth in 1966 sales among the 25 major chemical companies, Ethyl Corporation has shown the most aggressive sales growth history during 1955-66 with an annual growth rate of 30.67., primarily due to its profitable bid for petroleum additive and industrial chemicals business.6 National concern over possible pollution hazards caused by gasoline emissions involving tetraethyl lead anti-knock additives has forced Ethyl to re-examine its long-range position in this field. The company continues to diversify through both internal development and acquisition to include biodegradable detergents, plastics and paper products. Since only three companies ranking in sales among the tOp ten in 1966 had above average sales growth rates in the 1955—66 period, it appears that large asset firms have not dominated long term relationships in rapidly growing market areas, although the evidence remains inconclusive. Other companies showing exceptional annual growth rates include Air Products and Chemicals, Inc., with major expansions in industrial gases (23.87.); Witco Chemical Company, enjoying a strengthened position as a Specialty chemicals producer through the acquisition and new product routes (20.07.); and Celanese Corporation, with primary emphasis on the marketing of synthetic fibers, organic chemicals, and plastic products (17.17.). Those chemical firms having above average growth of net sales over the 1955-66 period also tended to have higher earnings growth (see Tables 4-2 and 4-3). In fact the comparative rankings between these two performance characteristics did not begin to differ prior to the sixth rank position, the most notable exception being Air Reduction who ranked ninth in sales growth over the 1955-66 period but dropped to *- 6Prospectus, (Richmond. Virginia: Ethyl Corporation, October 7, 1964), pp. 7-14; Special Meeting of Stockholders, (Richmond, Virginia: Ethyl Corporation, June 27, 1967), pp. 16-18. u trontefimo n r u a . ”-.....- ‘ ..- \> \ ...... ‘-*u n-A- - . <—..~ “.... .- \ ,- ... . n...._ ._‘ -. ‘u.,‘ .. .. "n. -. '--' l y ... . "u. ‘ 'I~ ¢.: - -. n4. !‘ ... \. '\ ... ... - In. 1;. .“. ‘N'. ,‘ ,- u. ‘ ‘\_ o u r ._ ,l . , n e . s . ". ,s o .. ‘ A v. s a . eighteenth in changes of net profit to common stock over the same time period. With the emergence of strong demand for industrial gases, Airco has benefited from its regionalized production capabilities in develOping specific users. But its competitive cost position was disrupted in the long run because of a management decision to limit production to relatively small scale Operations. Although the company had a broad marketing and distribution network, it has recently had to switch to larger production units in order to establish a more favorable cost position. Such policies did limit earnings over the time period under study.7 There were two companies, Diamond Alkali Company (now Diamond Shamrock) and Hooker Chemical Corporation (acquisition completed by Occidental Petroleum), whose experiences on changes in earnings per share and net profit to common substantially differed once ranked (see Tables 4-4 and 4-5). Diamond Alkali significantly improved its relative position on a per share basis, while Hooker Chemical slipped in the rankings as its number of shares outstanding increased through time at a faster rate than that of other chemical firms relative to their earnings. Among these companies we note a tendency to have a relative dilution of earnings on a per share basis, i.e., the propensity to increase the number of shares outstanding was greater than that of overall earnings growth. Thus the annual growth rates of earnings per share typically fell below that of net profits to common stockholders, often due partly to securing additional capital through equity financing for needed plant expansions. The Rohm and Haas Company was the only exception, since its 8.772. annual growth rate on net profit to common was less than that on a per share basis. In this case, there was no real earnings dilution over the 1955'66 period. It is difficult to generalize about cash flow patterns of major Chemical companies over the 1955-66 period since interfirm accounting practices have varied so widely. For many firms, capital investment M 7"A New Airco Diversifies for Growth," Chemical and Engineering M. Vol. 41, No. 44 (October 14, 1963), pp. 44, 46. . ...n- u- a — u ' ..x “...-«.... \ u, . _ v.- "vo-..,;: . ~. ..; - ':-. “- to.» A: ' -~ .. ch”... .. U I: . y“ . u .h . .. ,- u.‘ r. a“. . u,‘ -.'.. . .\ ‘ 61 requirements have increased faster than the earnings record on a share basis over the long run, pinpointing their willingness to accept deferred cash benefits far in the future (see Tables 4-4 and 4-5). Part of the explanation may lie with changes in firms' depreciation policies during the time period under study. Since it was strongly expansionistic-minded relative to its current earnings position, Reichhold Chemical Corporation stood out, seeking in many ways to broaden its product mix and earnings base. When compared with the historical earnings patterns of other major chemical producers, Reichhold has managed to improve its relative ranking some twelve positions in the listing of growth rates for cash flow figures . Industry-wide relative market valuations of stock have fallen during the 1955-66 period (see Tables 4-4 and 4-6); in fact, 567. Of the companies found that price movements of their common stocks during this time period did not even keep pace with changes in their respective earnings records. For the long term investor, the chemical group has fallen into disfavor, the investment community apparently concluded that the long term growth potential of the industry generally warranted more conservative stock valuations. Mean price-earnings ratios of the major chemical firms under study between 1955 and 1966 ranged from a low Of 12.02 for Eagle-Picher Company to a high of 27.38 for The Dow Chemical Company (see Table 4-7). Diversified chemical companies probably attract capital on the basis of long term growth rather than on expectation of immediate earnings. Many companies have been accorded rather liberal price-earnings ratios. For example, Dow Chemical has a strong, broad position in basic chemical commodities, its large scale Operations being instrumental in maintaining favorable production economics. Dow Chemical also is a manufacturer Of many upgraded products, having diversified activities in animal and human health areas, plastics, metals and packaging. Its commitments in overseas activities are expanding rapidly to meet the needs of growing foreign markets.8 8 The Dow Chemical Company, 1965 Annual Report, pp. 2-7. -~v- I NC...» .... ”....v - .- r . ...-...- " ‘Iua... .' ’ a n "I'Ciuu\ '4 I b "-1. ll ‘ v Hu‘... .1. I‘m \ ‘ . I ‘H . ‘ "fi- . .n u ’D‘ . ‘ ' ‘1 ..h ‘9'. . 'I- ‘~-- . ‘ q \ . .‘b ! vv«. .,‘. O. ‘ a. w _ On-D . x ‘ ‘ -. ...“: .‘ -, .I I ... ;-. . h" I ..0 a . .. ..; 62 Record capital eXpansion programs budgeted by the chemical industry have required the use of funded debt. Many firms would simply not have been able to add new capacity in established product areas to meet the growing market needs if they were limited to internally generated cash sources. In terms of the pOpulation, seventeen firms out of the 25 had mean long term debt positions exceeding 207. of total capital (see Table 4-8). The favorable provisions of the Revenue Act Of 1964 were quite a stimulus to increased plant investment activities, with its allowance for a 7‘7. investment tax credit resulting in an effective reduction in corporate taxation rates. Not only did this measure have an immediate impact on profitability, but it provided worthwhile investment incentives for growth. Most firms now have considerably higher leverage capital structures than the long term positions of major chemical firms shown in Table 4-8 covering the 1955-66 period. Profits before taxes as a percent of sales for basic chemicals began to dip in 1958, recovering well the following year, though they have never since managed to surpass the 17.57. high of 1959.9 In recent years chemical firms have been unable to keep rising costs in line with a more stabilized price structure. Part of this deteriorating profitability is due to significant increases in capital expenditures and expensive start-up costs of new facilities. Return on common stockholders' equity has improved significantly for the chemical industry since 1961, though continued pressure on profit margins after 1965 has forced a reversal in this trend.10 During the 1955-66 period, the mean return on common stockholders' equity in the pOpulation ranged from a low of 6.07. for Koppers Company, Inc., to a high of 22.07. for Ethyl Corporation (see Table 4-9). KOppers Company, however, has depended fairly heavily on less profitable nonchemical fields for earnings, including highly cyclical engineering and construction activities. Coupled with less aggressive marketing efforts in chemicals and plastics, inefficient producing plants and price attrition account ——___ 9 Federal Trade Comission- Securities and Exchange Commission, Mlgsgterly Financial Report for Manufacturing Corporations, various issues ~1968. , 10 Ibid, various issues, l96l-l968. n»... '01. . u, \ ' ‘ ... .M ‘ We ‘ . 3' u“v" " o ’ ...‘_ . -._ ~ . . '.. \ w- '. - , .. ‘ . ‘v . < 63 for its poor showing on the return measure relative to other major chemical ll . manufacturers. 0n the other hand, Ethyl has managed to sustain a very aggressive expansion program, capitalizing on patent protected positions in a select number of profitable upgraded chemical products. Whenever investment Opportunities exist within any Operating firm, executive management must select capital sources if authorizations are to be granted. Within the defined population, six companies have had mean retention rates in excess of 607. (see Table 4-10). All except Reichhold Chemicals, Inc., had above average earnings growth, with the retention of earnings important in achieving these performance records in the 1955-66 period under study. The duPont situation is peculiar: since this company requires a high cut-Off return on investment rate before prOposed projects are acceptable to management, the number of new investment opportunities did not far outpace that which could be funded from current Operations. The company has paid out in dividends the value of its net income earned during the entire twelve year period, plus some more. Yet the company's earnings from its investment in General Motors common stock, before its forced divesture in January, 1965, were sizable. In actuality, the aggregate of preferred and common stock dividends paid from duPont sources was 71 percent of earnings for the ten year period 1957 through 1966.12 It is interesting to note that duPont had no long term debt Obligations Outstanding before 1965. 11 "KOppers Prepares for the Good Years," Chemical and Engineering 593.2: V01. 42, NO. 42 (September 28, 1964), pp. 33-37. 12 E. I. duPont de Nemours & Company, Annual Report for the Year 1966) p. 410 - ' v . u I 4 . I an: "Alu v ‘ Q I l. I H . n u .j. a . ... . . “-1.: . . in. ., s? -\ n I, 64 TABLE 4-1 PERFORMANCE PROFILE OF MAJOR CHEMICAL FIRMS IN STUDYI: 1966 (Millions of dollars) 2 Company Net Sales Net Income Ranking 8.1; duPont de Nemours & Company 3,185.1 389.1 1 Ikfion Carbide Corporation 2,224.0 231.0 2 Monsanto Company 1 ,611.9 112.4 3 The Dow Chemical Company 1,309.7 121.7 4 W. R. Grace & Company 1,278.7 58.6 5 Allied Chemical Corporation 1,245.6 89.2 6 (Min Mathieson Chemical Company 1,117.1 66.8 7 Celanese Corporation of America 1,019.9 66.7 8 FMC Corporation 1,009.7 62.9 9 American Cyanamid Company 952.6 94.4 10 Hercules, Inc. 601.0 52.3 11 Koppers Company, Inc. 430.0 11.4 12 Air Reduction Company, Inc. 423.0 28.8 13 Ethyl Corporation 384.2 29.2 14 Rohm & Haas Company 371.2 35.5 15 Stauffer Chemical Company 360.4 32.6 16 Hooker Chemical Company 284.1 25.8 17 Diamond Alkali Company 233.0 19.0 18 Chemetron Corporation 232.7 16.1 19 Pennsalt Chemicals Corporation 222.5 12.7 20 Witco Chemical Company 195.9 9.3 21 Eagle-Ficher Company 170.9 7.4 22 Air Products 8 Chemicals, Inc. 146.9 9.2 23 Reichhold Chemicals, Inc. 136.7 5.7 24 126.3 8.0 25 Wyandotte Chemicals Corporation *- 1Source: Basic data listed on Compustat magnetic tapes compiled lw'Standard Statistics Company, Incorporated, New York, New York, and nude available to Michigan State University. 2 Ranked by net sales in descending order. _—___ 65 TABLE 4-2 PERFORMANCE OF MAJOR CHEMICAL FIRMS IN STUDY ON NET SALESI: 1955-1966 Annual Growth Ranking Correlation Rate Of Net in Des- Coefficient Sales As cending Throu h Company - A Percent2 Order Time Air Products & Chemicals, Inc. 23.81 2 0.9544 Air Reduction Company, Inc. 9.77 9 0.9776 Allied Chemical Corporation 6.41 17 0.9577 American Cyanamid Company 6.37 18 0.9771 Celanese Corporation of America 17.06 4 0.9350 Chemetron Corporation 5.45 19 0.8526 Diamond Alkali Company 6.47 16 0.9610 The Dow Chemical Company 7.53 15 0.9954 E.]L duPont de Nemours & Company 4.99 23 0.9571 Eagle-Richer Company 3.15 25 0.7043 Ethyl Corporation 30.57 1 0.9509 FMC Corporation 13.49 5 0.9646 w. R. Grace a Company 9.24 I 11 0.9133 Hercules, Inc. 10-48 7 0.9797 Hooker Chemical Corporation 9.69 10 0.9848 KOppers Company, Inc. 3.58 24 0.7148 Mbnsanto Chemical Company 12.25 6 0.9776 (Min Mathieson Chemical Company 5.28 21 0.9440 Pennsalt Chemicals Corporation 10.08 8 0.9572 Reichhold Chemicals, Inc. 8.55 12 0,9667 Rohm & Haas Company 8.44 13 0.9908 Stauffer Chemical Company 8.39 14 0.9698 lhuon Carbide Corporation 5.18 22 0.9654 lfitco Chemical Company 19.96 3 0.9719 Wyandotte Chemicals Corporation 5.42 20 0.9554 —-__ 1Source: Basic data listed on Compustat magnetic tapes compiled 1W'Standard Statistics Company, Incorporated, New York, New York, and made available to Michigan State University. 2 . Based on the lepe of the regre581on line that best fits the logarithms of the data and indicates the compounded annual change for the statistic. 3 Indicates the goodness of fit for the data and the association 0f the actual and nreroran values through time. 66 TABLE 4-3 PERFORMANCE OF MAJOR CHEMICAL FIRMS IN STUDY ON NET PROFIT T0 COMMONl: 1955-1966 Annual Growth Rate of Net Ranking Correlation Profit to in Des- Coefficient Common As cending Throu h Company A Percent2 Order Time Air Products & Chemicals, Inc. 24.76 2 0.9526 Air Reduction Company, Inc. 6.41 18 0.8060 Allied Chemical Corporation 7.13 17 0.8278 American Cyanamid Company' 8.39 14 0.9332 Celanese Corporation of America 18.90 4 0.9466 Chemetron Corporation 8.16 15 0.4584 Diamond Alkali Company 6.34 19 0.7621 The Dow Chemical Company' 7.32 16 0-8657 E.1L duPont de Nemours & Company 0.92 25 0.3392 Eagle-Picher Company 1.63 23 0.1749 Ethyl Corporation 34.49 1 0-8893 FMC Corporation 14.95 5 0.9641 W. R. Grace 6. Company 12.09 8 0.8014 Hercules, Inc. 10.22 9 0.9631 H00ker Chemical Corporation 8.75 11 0-9179 KOppers Company, Inc" 0,97 24 0.1274 Monsanto Chemical Company 13.01 7 0-9480 (Min Mathieson Chemical Company 4.50 20 0-3723 Iknnsalt Chemicals Corporation 13.44 6 0-9723 Reichhold Chemicals Corporation 4.44 22 0.1577 Rohm a Haas company 8.72 12 0-9190 Stauffer Chemical Company 8.40 13 0'9215 Union Carbide Corporation. 4.46 21 0’7959 Witco Chemical Company 22.65 3 0'9935 Wyandotte Chemicals Corporation 9.85 10 0.5872 l b Source: Basic data listed on Compustat magne y Standard StatiStics Company, Incorporated, New Yor made available to Michigan State University. 2 1 Based on the slope of the regression lin Ogarithms 0f the data and indicates the compoun the statistic, 3 1-4.3--.. _ . .0 C _ -C jlllllt'nrnn run nnnnnncfl n‘r tic tapes compiled k, New York, and e that best fits the ded annual change for {Lil- Fnr fhra data and thn “minio- \\ u '1 -- . n I _ P u ...... . . . ... _ . ‘ . I u I- s. “ ~ I. \_~~ ,' A '- - V 1. '4 ‘¢ 0 r. 1 1.. u a 1 u . . ... . '- 67 TABLE 4-4 PERFORMANCE OF MAJOR CHEMICAL FIRMS IN STUDY :ON EARNINGS PER SHAREl: 1955-1966 Annual Growth Ranking Correlation Rate Of Earnings in Des- Coefficient Per Share 2 cending Through Company As A Percent Order Time Air Products 5: Chemicals, Inc. 14.82 2 0.8933 Air Reduction Company, Inc. 1.99 22 0.3768 Allied Chemical Corporation 3.47 20 0.6475 American Cyanamid Company 7.13 11 0.9078 (blanese Corporation of America 12.69 5 0.9621 Chemetron Corporation 4.77 17 0.2843 Ifiamond Alkali Company 5.55 14 0.7120 The Dow Chemical Company 6.09 13 0.7927 E.]L duPont de Nemours & Company 0.80 24 0.2992 Eagle-Picher Company 1.57 23 0.1666 Ethy1 Corporation 27.11 1 0.8598 FMC Corporation 13.26 4 0-9519 W. R. Grace 8: Company 6.38 12 0.6865 Hercules, Inc. 8.33 10 0.9506 Hooker Chemical Corporation 5.25 l6 0-8102 K0Ppers Company, Inc. 0.53 25 0.0674 Mbnsanto Chemical Company' 10.04 7 0 9357 (Min Mathieson Chemical Company 4.47 18 0-3687 Iknnsalp Chemicals Corporation 11.02 6 0-9684 Reichhold Chemicals Corporation 3.26 21 0.1152 Rohm & Haas Company 8.97 9 0'9208 Stauffer Chemical Company 5.41 15 0-3343 Ihficn Carbide Corporation 4.29 19 0'7794 Witco Chemical Company 13.36 3 0-9355 Wyandotte Chemicals Corporation 9.30 8 0'5755 c tapes compiled Source: Basic data listed on Compustat magneti and t’Y’Standard Statistics Company, Incorporated, New York. New York, made available to Michigan State University. Based on the slope of the regression line that best fits tlfler logarithms of the data and indicates the compounded annual change 0 the statistic. “ 3Indicates the goodness of fit for the data and the association of the aetual and predicted values through time. _ _. t, - —-- 1 - h .. fl “‘ up . ‘ ' 41 '1 .' b . ‘ . _" . . ...: . ... ‘r ..V|A ‘ . ‘1 H" i y . . .9, "". 1.. . . v N. u '4'. . V i 4 . . n . 68 TABLE 4-5 PERFORMANCE OF MAJOR CHEMI AL FIRMS IN STUDY ON CASH FLOW PER SHARE : 1955-1966 I 4 Growth Rate of Cash Flow Ranking Correlation Per Share in Des- Coefficient Per Annum As cending Throu h Company A Percent2 Order Time Air Products 8 Chemicals, Inc. 20.88 2 0.9712 Air Reduction Company, Inc. 2.31 23 0.5824 Allied Chemical Corporation 3.61 21 0.8419 American Cyanamid Company 5.10 16 0.9506 Celanese Corporation of America 10.54 6 0.9446 Chemetron Corporation 4.61 17 0.5045 Diamond Alkali Company 5.35 14 0.8838 The Dow Chemical Company 4.38 19 0.8944 E. I. duPont de Nemours & Company 2.91 22 0.3343 Eagle-Picher Company 1.62 24 0.2604 Ethyl Corporation 27.35 1 0.9032 FMC Corporation 12.98 4 0.9606 1L R. Grace & Company 5.91 11 0.9478 Hercules, Inc. 8.23 8 0.9738 Hooker Chemical Corporatirni 5.29 15 0.8952 KOppers Company, Inc. 0.52 25 0-1740 ansanto Chemical Company 11.36 5 0-9784 (Min Mathieson Chemical Company 4.34 20 0°6991 Ihnnsalt Chemicals Corporation 5.44 13 0-9804 Reichhold Chemicals Corporation 8.17 9 0°7875 Rflhm & Haas Company 9.26 7 0°9794 Stauffer Chemical Company 5.45 12 0.9215 Ihucn Carbide Corporation 4.60 18 0.9240 lfitco Chemical Company 14.84 3 0'9882 Wyandotte Chemicals Corporation 6.74 10 0-8530 piled 1Source: Basic data listed on Compustat magnetic tapes comand lW'Standard Statistics Company, Incorporated, New York, New York, made available to Michigan State University. e that best fits the Ba ession lin sed on the lepe of the regr ded annual change for logarithms of the data and indicates the compoun the statistic. 3 if Fnr than data and tbs .Wtim 7-1:- a -0 C -C .C I lln1nnrnn *hn han‘nnflfl n1" 1 . . p ' -- -b.- I I. In. ...._‘_.._ ‘ .. ... ..-. c , ‘ “he“. n... . . F n... ' A, . ..-.H h' .‘w -. vu- -... ‘ ,.,_ . ~.. _ v- , Veg. 'u “I - ~ 7 ’s a e ‘. ' Iv... 'a.\ .. ow I... . . -v. .‘ \ -. \ ‘.'-. - -‘a ' .\ .._ . ... .v . 1 ,. .- I . -. . I‘- 69 TABLE 4-6 PERFORMANCE OF MAJOR CHEMICAL FIRMS IN STUDY ON AVERAGE MARKET PRICEl: 1955-1966 0R OTHERwISE INDICATED Growth Rate of Ranking Correlation Average Market in Des- Coefficient Price Per Annum cending Throu h Company As A Percent2 Order Time Air Products & Chemicals, Inc. 16.41 4 0.8922 Air Reduction Company, Inc. 3.17 17 0.4764 Allied Chemical Corporation 0.15 24 0.0490 American Cyanamid Company 7.57 11 0.8880 Celanese Corporation of America 18.56 3 0.9436 Chemetron Corporation 2.78 18 0.3346 Diamond Alkali Company 4.95 14 0.8016 The Dow Chemical Company 2.17 19 0.4385 3.]; duPont de Nemours & Company 1.02 22 0.3389 Eagle-Picher Company' 4.37 15 0.8338 Ednd.Corporation 32.65 1 0.9321 FMC Corporation 18.59 2 0-9837 1% R. Grace & Company 10.41 5 0-8653 lmrcules, Inc. 8.41 8 0.9035 Hooker Chemical Corporation 1.97 20 0-4910 KOppers Company, Inc. 0.52 23 0.1087 Monsanto Chemical Company 8.39 9 0.8752 Olin Mathieson Chemical Company -1.07 25 -0.2384 Pennsalt Chemicals Corporation 10.37 5 03633 Reichhold Chemicals Corporation 7.924 10 N-C- ROhm 5! Haas Company 6.65 17- 0'7667 Stauffer Chemical Company 4.32 16 0'5428 Union Carbide Corporation 1.21 21 0.3942 Witco Chemical Company 8.975 7 N'C' Motte Chemicals Corporation 5.914 13 N'C' N.C. = Not Calculated Source: Basic data listed on Compustat magnetic tapes compiled by Statldard Statistics Company, Incorporated, New York, New York, and made available to Michigan State University. 1 Based on the slape of the regression line that best fits the th 0&ndrhm3 Of the data and indicates the compounded annual change for e Statistic. 3 . Indicates the goodness of fit for the data and the associatlon of the Zetual and predicted values through time. Calculated in time nae-ind 1056-1966- \. LJJU 1 'F 4.. "~Fv» .. -. ’ --- .- a '----. .I . -~ . . . I ‘~ ..I _ . . . A .. . . .‘-L' , . .... .. \“- . .. 70 TABLE 4- 7 PERFORMANCE OF MAJOR CHEMICAL FIRMS IN STUDY ON PRICE-EARNINGS RATIOI: 1956-1966 OR OTHERWISE INDICATED Ranking in Mean Price-Earnings Descending Company Ratio (Times)2 Order Air Products & Chemicals, Inc. 22.94 4 Air Reduction Company, Inc. 16.58 17 Allied Chemical Corporation 19.96 10 American Cyanamid Company 18.95 12 Celanese Corporation of America 12.43 23 Chemetron Corporation 15.80 19 Diamond Alkali Company 15.05 20 The Dow Chemical Company 27.38 1 E. I. duPont de Nemours & Company 24.65 3 Eagle-Picher Company 12.02 25 Ethyl Corporation 12.03 24 FMC Corporation 17.56 16 W. R. Grace & Company 16.36 13 Hercules, Inc. 22-66 Hooker Chemical Corporation 20.69 8 Koppers Company, Inc. 13.78 22 Monsanto Chemical Company 20.30 9 Olin Mathieson Chemical Company 17.99 15 Pennsalt Chemicals Corporation 21.97 7 Reichhold Chemicals Corporation 18.933 13 Rohm 5: Haas Company 26.17 2 Stauffer Chemical Company 19-23 11 Union Carbide Corporation 22.20 6 Witco Chemical Company 14.874 21 Wyandotte Chemicals Corporation 18.53 14 Source; Basic data listed on Compustat magnetic tapes compiled Edmond.“ StetiStics Company, Incorporated, New York, New York, and e available to Michigan State University. b Price'fiiarnings ratio defined as adjusted average price divided d Y adjusted earnings per share (adjustments are made for stock splits an stock diVidends), re Median selected to eliminate distortion caused by abnormal earnings come: in at least one year of Operation. . r - A. n. .3 p a. Inc-eh. ' . . n ', a " b‘n.‘ as. u 1% -- \.' '. ._. ._ ‘1. r ,. k. ‘l ... \ 9 u ,_ . . ~\ u. . ‘\ ‘. u.» \ 71 TABLE 4-8 PERFORMANCE OF MAJOR CHEMICAL FIRMS IN STUDY ON LONG TERM DEBT T0 TOTAL CAPITALl: 1956-1966 Mean Long Term Debt Ranking in to Total Capigai As A Descending Company Percent Order Air Products & Chemicals, Inc. 47.05 2 Air Reduction Company, Inc. 31.20 8 Allied Chemical Corporation 29.73 12 American Cyanamid Company 16.73 18 Celanese Corporation of America 34.87 5 Chemetron Corporation 32.17 7 Diamond Alkali Company 25.73 14 The Dow Chemical Company 25.11 15 E. I. duPont de Nemours & Company 0.36 25 Eagle-Picher Company 23.80 16 Ethyl Corporation 53-65 1 FMC Corporation 30.07 11 W. R. Grace 8: Company 41.29 4 Hercules, Inc. - 2.36 23 Hooker Chemical Corporation 32-63 6 K0PPers Company, Inc. 16.18 20 Monsanto Chemical Company 30.74 10 Olin Mathieson Chemical Company 42.56 3 Pennsalt Chemicals Corporation 21.17 17 Reichhold Chemicals Corporation 26-85 13 Rohm & Haas Company 1-22 24 Stauffer Chemical Company 16-49 19 Union Carbide Corporation 31-17 9 Witco Chemical Corporation 16.15 21 Wyandotte Chemicals Corporation 5-76 22 ‘ . Source; Basic data listed on Compustat magnetic tapes com- filled by Standard Statistics Company, Incorporated, New York, New ork, and made available to Michigan State University. 2 f. Long term debt defined as debt obligations due beyond one 18°31 Period with purchase obligations and liabilities to offers excluded as well as subsidiary preferred stock. 3 The value of total capital was found by summing the values of 10% term debt, preferred stock valuation, and common eQU1tY- 4 .. The long term debt to equity valuation position is one per \\ .. .. ._- .‘H- .....‘Uii 1 ...” ' a" - “1 u U . V . ‘ ... . i l ' 1. h. -‘. l t.'r 7‘. ’7 72 TABLE 4-9 PERFORMANCE OF MAJOR CHEMICAL F RMS IN STUDY ON SELECTED PROFITABILITY RATIOS : 1956-1966 Ranking of Return on Mean Return Mean Return Common Stock on Total on Common Equity in Capital As Stock Equity Descending Company A Percent ’ As A Percent Order Air Products & Chemicals, Inc. 9.87 13.65 8 Air Reduction Company, Inc. 9.43 11.67 15 Allied Chemical Corporation 9.28 11.52 17 American Cyanamid Company 11.79 13.42 9 (blanese Corporation of America 7.94 12.50 12 Chemetron Corporation 9.05 10.43 18 Diamond Alkali Company 9.46 12.58 11 The Dow Chemical Company 10.43 12.22 14 3.1; duPont de Nemours & Company 17.29 18.83 2 Eagle-Picher Company 8.14 9.80 19 EthYI Corporation 10.40 21.99 1 FMC Corporation 10.59 13-82 7 W- R. Grace 6: Company 7.90 8.73 22 Hercules, Inc. 14.40 15.40 4 Hooker Chemical Corporation 10.78 14.42 KDPPeIS Company, Inc. 5.54 5.97 25 Monsanto Chemical Company 9.34 11.57 16 Olin Mathieson Chemical Company 7.33 9-51 20 Ihnnsalt Chemicals Corporation 8.30 9-41 21 ReiCthld Chemicals Corporation 7.37 7.92 23 Rohm & Haas Company 14.56 15.08 5 Stauffer Chemical Company 11.67 13.36 10 ‘_ (Continued) ll | u u I.-. ' . a , .~ .. a .\ 9..., ..<'\ -. u.. v u r u n .5,- .‘ . h -. ‘c 73 TABLE 4-9 PERFORMANCE OF MAJOR CHEMICAL FIRMS IN STUDY ON SELECTED PROFITABILITY RATIOSI: 1956-1966 (Continued) Ranking of Return on Mean Return Mean Return Common Stock on Total on Common Equity in Capital As Stock Equity Descending Company A Percentza4 As A Percent Order Ikficn Carbide Corporation 12.32 16.25 3 Witco Chemical Corporation 11.02 12.23 13 Wyandotte Chemicals Corporation 6.39 6.57 24 E 1 . Source: Basic data listed on Compustat magnetic tapes compiled by SUHMard Statistics Company, Incorporated, New York, New York, and made available to Michigan State University. Return on capital for any year is defined as net income and fixed Cbarges divided by average annual total capital, i.e., (net incomet + fixed chargest) E (total capitalt 1 + total capitalt) ‘ 2 f Return on common stock equity is defined as net income minus pre- err“dividends divided by average annual common stock equity, i.e., (net incomet 3 preferred dividendst) k (common stock equity k t 1 + common stock equityt) 2 t The Value of total capital was found by summing the values Of long ehHCbbt: Preferred stock valuation, and common equity. . w _,.. .. u ‘9 9Ayu'a». h... v . vau- “-6.1. n '.-_‘ '. n “"‘i u. .... a .. 74 TABLE 4-10 PERFORMANCE OF MAJOR CHEMICAL FIRMS IN STUDY ON RETENTION RATESl: 1955-1966 Mean Retention Rate of Ranking in Net Earnings on Common Descending Company Stock As A Percent Order Air Products & Chemicals , Inc. 89.13 1 Air Reduction Company, Inc. 35.47 22 Allied Chemical Corporation 31.84 24 American Cyanamid Company 36.68 21 Cklanese Corporation of America 56.61 7 Chemetron Corporation 37.43 19 Diamond Alkali Company' 48.08 13 The Dow Chemical Company 42.31 17 E.lfi duPont de Nemours & Company -5.01 25 Eagle-Ficher Company 39.41 18 EthyI Corporation 82.58 2 FMC Corporation 60.74 6 W. R. Grace & Company 49.51 12 Hercules, Inc. 52.80 9 Hooker Chemical Corporation 42.45 16 K0ppers Company, Inc. 36.81 20 Monsanto Chemical Company 55.14 8 (Min Mathieson Chemical Company 44.61 14 Itnnsalt Chemicals Corporation 44.58 15 Reichhold Chemicals Corporation 78.482 4 Rohm & Haas Company 82°12 3 Stauffer Chemical Company 50-72 11 Ikfion Carbide Corporation 32.30 23 Witco Chemical Company 64.20 5 Wyandotte Chemicals Corporation 51-38 10 1 Source: Basic data listed on Compustat magnetic tapes compiled by Standard Statistics Company, Incorporated, New York, New York, and made aVailable to Michigan State University. 2 d‘ . Median selected to eliminate distortion caused by an abn ludend Payout in at least one year of Operation. ormal nun pr. '. I”. V '- -..L' .. '4 ‘- i ‘7 u, . l '3‘. .. w . ... . In. . bl . .. . I. ’ .. \ ~ ‘I— .p» 1 ‘ ‘ .‘~; ‘ . p . II CHAPTER V FUNDAMENTAL POSTULATES BEHIND THE RESEARCH EFFORT Any research study of this SCOpe requires a thorough investigation into numerous traditions directly influencing the research findings, into those postulates incorporated into the fabric of this thesis on the basis of widespread practice and not assessed empirically. Postulate NO. 1 New products are basic to the growth of any firm. The chemical industry's fine growth during the present decade has been largely based on technological progress, unfulfilled demand (running ahead Of supply capabilities), and favorable economic conditions, with one key strategy employed to achieve these results being a strong emphasis on new product development activity. Chapter II has already discussed the product life cycle concept. Certainly product cycles in any given industry have specific patterns for both sales and profits. This thesis, therefore, seeks common patterns relating these two factors within one Product class, industrial chemicals. Chemical firms have adOpted many planning orientations with varying levels of success: e.g., projections based on historical sales records of specific territories may be made to indicate future performance, 01‘ divisional performance records may be set against one another to illustrate comparative results. Yet firms stressing anything but a PrOdUCt orientation to business planning may lack the product data necessary to make those strategic decisions seriously altering eventual outcomes. Obviously one needs technological research and new product development activity. But management still must ask the right PrOdUCt questions, get the right information, and then act on that information- , . ‘3 d\ d‘ov- .‘0-‘fi’fl '......3 ...d —c- -.~ '1 p ...-c .... l ' r n -- ._.. v r- .. ......u» . n u. p --~ ‘ .4 I “" on r -.. ..‘, ,1 s - ...‘ \ I o“ 76 Sales and profit patterns assume finite shapes for any product, a fundamental concept in business development. Various linear and curvilinear models have been proposed to depict the product life cycle. Choice depends on the variables considered relevant and their patterns of influence. If revenue and profit are examined over time, these might vary as depicted by the model for new products posited by BoozoAllen and Hamilton, Inc., and shown in Figure 5-1. The two cycles (sales and profits) are analyzed separately. As sales begin to grow substantially in the growth phase, profitability also begins to demonstrate significant growth, so long as there is an unfilled demand for the product. Recognizing this demand, management provides the necessary production facilities which, along with economies of scale reflected in higher levels of plant utilization, generate increasing unit profit margins. But as sales growth tapers Off in the maturity stage, both aggregate and unit profit margins decline under the pressure of competitive forces (including price erosion). As Obsolescence eventually sets in, both aggregate profits and sales decline, and either the buyer need for the product may disappear or substitutions may win out on the basis of product improvement and the tactical superiority of competitors. mLIHU‘J-oOtl llfiAlV NHI~HV>O Hui Ksld OhmR‘MN - All. saves ...: :- 77 .2: :2:on w :23. £25 .EoEtoaoo cucoomom Essences: will .. 88m mzfluwo zofi._._m3.24..on 2.458 on omommz :noma 828mm 3m: ..azoEooa m230> mun—4m ] /I\ 63,; 33sz e mPODQOmQ ...—O m.._0>o mu... o_m....-9 ‘- \ u ; 6». "1H I h “4...“... p - uh , ; Acoofionh‘. 1 " nvl >- "ll-Ii: .& .. .. P' .... ... . v .. Q . ' ...“- .... ... 'cn § .5... ... ... 4-u . - .... , 78 Marketing dynamics enhance the importance of the product life cycle concept in preparing for possible eventualities. Most viable business organizations are growth-conscious. And given this corporate objective, the product life cycle concept supports assigning new product development activities high priorities: as Figure 5-1 demonstrates, any firm that is going to sustain growth while one existing product faces decreasing aggregate profit contributions must introduce additional products to fill the developing void. This reasoning applies even to well-diversified firms in which any one product may contribute a relatively small percentage of the total earnings. The Booz.Allen and Hamilton study showed that sales growth for major industries stems from two sources: existing products and new products. Growth from new products ranged between 467. and 1007. for major industries, with a mean of 75%.1 Obviously product planning is the key to achieving growth expectations, especially with new products. Thus we may assume that aggregate profits grow only as a continuing stream Of new products gain market acceptance. Evidence in current literature substantiates the importance of new products to growth for manufacturing firms, as shown in Figure 5-2 for eleven selected industries. 1 . Booz-Allen and Hamilton, Inc., Op. Cit., p. 6. _ _ . klmm- l ”mauv- OI_F BOVMKO mkflldl‘mu AHHILIHVHUA‘XNIU A..V~r mh HU‘JQHV~\A§ gl‘z .kK.V ZKVhlh ‘HMLV‘MV‘N ZKILV Ii. 0| Its: '1'. h 79 Joom 23> 262 .353 EX m 5:4 .88 e. octooom 3 .22 2so cs. m ....e ..mom. .8. .8223: m 5:4 I!!! 11 $8005 :02 co 2353282 :2... 62:60: co_mm_EEoo one... .23»... ..ll 83:8 . . 84.38 m .8552. 33% 80 n. 80 m. 086 806 000.». 0 253783 $382.. 532. escapee. om e0 8380:. 86m accustom l e0 8882. mgom 368th ”U o._. 2269a :62 2262a mazmim l” I” l“_ l l H ..mmpm w 202. mmmqa wmgm: maommwmzoz 83o m 28 .mzoem 32:9 H mmqkmxs. >mmzio--\ 82 Note particularly the relative difficulty in estimating the variables of savings, development cost and completion time. Seven of the projects involved were reviewed more than once after the projects began, and revised estimates made. Actual estimates were then closer to results and dispersions lessened, which shows that a learning process existed with this type of work, and that errors in estimation were reduced signifi- cantly as projects approached completion. Hence there is value in documenting the predictive capabilities in estimating demand for specific new and existing products. Tull and Rutemiller recently studied the relationships of actual and predicted sales and profits in new product introductions, based on . . 4 . data collected from sixty-three new product case histories. In addition to other models, they solved the general regression model of the form: Forecasti = C(‘Ffl (Actuali) 2 (5.1) . _ 1/2 Variancei - [2’ +6 (Actuali) ] (5-2) With the forecasts as the dependent variable and actual sales or profit data as the independent variable, they adOpted a maximum likelihood approach in estimating the regression coefficients (a, £7, 7, and 6), assuming heteroscedasticity, i.e., that each pOpulation does not have the same variance. The authors found that the variance model (5.2) cited here had the highest likelihood of occurrence and concluded that variance in forecasting both sales and profits increased at a decreasing rate as a function of the level of actual product behavior. Though they detected no bias in the case of profit forecasting, sales revenue forecasts for new products tended to be optimistic. ‘ 4 . Donald S. Tull and Herbert G. Rutemiller, "A Note on the Relation- shlP of Actual and Predicted Sales and Profits in New-Product Introductions," The Jaurnal Of Business, V01. 41, NO. 3 (JUIY, 1968), pp. 385'387. ‘9 L~n‘ ille- '- "A. and u. c )4 Ni. 83 Postulate NO. 3 The cost of capital concept aids in effectively evaluating the degree of financial uncertainty associated with new product decisions. The cost of capital, a significant issue facing the decision maker in his investment strategy, implies the effective utilization of capital assets, and can take into consideration the leverage position Of the firm and the expectations of investors. The cost of capital is that dis- cmnu:rate applied to estimated streams of Operating and investment cash flows which reflect risks in project evaluation. Theorists have traditionally limited the concept to financial risks facing the firm, arguing that all lnminess riSks eventually influence the financial structure of the firm. Such a definition suggests, of course, that all projects with positive net present values (i.e., discounted Operating cash flows greater than investment cash flows required to undertake the projects) are desirable. How does one select the prOper discount rate? Theoretical costs of capital range from the rate associated with riskless investments (the pure interest rate, if not higher) to near infinity. Though the cost of capital is fixed at any Specific moment in time, it changes with changes in factors affecting business outcomes and subsequent performance. Basically the cost of capital is a "futurity" concept. Any attempt to determine the cost Of capital empirically should rest on the expected capital structure as a composite of past decisions, eSpecially since most large investment decisions seriously affect future debt requirements, stockholders' expectations of earnings growth, and the changes in risk associated with new and existing business ventures (particularly the product mix) . ., a. ‘ fl rite-u ’1 a; V P“ v» £14. y: '3 :1”: .....u e... 'F ... r . -. n... ,‘ . v- " ”W hen... . "-0 ~. \ ,9. 7". V. I ‘ O ..-_\ ‘ we ' v. .: v. >\. p: “'1: 'Vo. ’ ‘n r. . ‘n — . ‘ a t '4 ’ I ' a u u.~ . ‘1. __ 84 Certainly the minimum acceptable rate of return for any project would be the actual aggregate cost Of capital of the firm, regardless of the business risks in any new product situation. Nonfinancial risks are seldom measured explicitly in the cost of capital calculation because they pertain uniquely to specific investment decisions. The decisions facing management today are the very ones that would largely affect the financial performance of the firm in the future. Since stockholders by and large lack any inside information on intimate details, one cannot expect their judgment to suitably reflect changes in company Operations in the short run. To offset this peculiar position, many decision makers, in evaluating individual product proposals, select a somewhat higher discount rate than the cost of capital to reflect the risks in expected cash flows; this tends to reduce actual calculated present values for proposed products than would Otherwise occur. Chapter III has already presented the exact model used for calculating the cost of capital in this dissertation. A concept with certain limitations, it does not adequately analyze projects with varying levels of risk nor venture interdependencies. But even if the cost of capital concept does not give ideal information in risk analysis, it remains a crucial variable in evaluating new product results. and. “ ""1”, .‘ AI 5| v e,. .N‘H {I “..., "A. ‘ \n a: s" " _ . .4 h .‘l D ‘ i .. 'Hu.‘ ..z‘ r ‘ I ‘I 3., .3“- '. ~ \ ‘v .‘I CHAPTER VI POSSIBLE INFLUENCES DETERMINING NEW PRODUCT BEHAVIOR In order to approach field investigative work methodically, a number of hypotheses were formulated as the general framework upon which to build the study. This chapter discusses Operating influences originally thought to be important in describing new product behavior. Whenever statistical tests would be meaningful in the analysis of data, the statement delineating the particular postulated influence is presented as a testable hypothesis in alternate form. We shall first discuss several general hypotheses concerning the product life structure of new industrial chemical products, and add a premise considering the time horizon appropriate to use in evaluating new product proposals. Then, after presenting the classification schemes used as indicators of performance, we turn to describing factors thought to relate to the performance of new industrial chemical products. General Hypothesis No. 1 No single representative nth order polynomial function best describes the sales patterns of new industrial chemical products. A function which is the sum of a finite numberof monomial terms, i.e., of the form cxn where c is a constant and n is zero or a positive integer, is called a polynomial of x. Extensions of the simple linear form y = a + bx, employing higher powers of x, give polynomial expressions of the type y='---'a-+-bx-§-cx2+dx3+...-i-nrxn where y ="- the dependent variable a,b,c,d m = constant terms n = the nth order of the polynomial expression x = the dependent variable 85 ' C - 1 c ".1... . . A .- »..- .y | 03‘ 86 Here we have a polynomial in one variable: y is a function of x alone. Ihia relationship of this type, a specific value of y is given by the sum (fifa finite number of terms, each of which consists of a specific power of xrmfltiplied by a constant. This form will be applied in the research snub“ with sales as the dependent variable and time the independent variable. We shall attempt to identify the specific types of sales patterns, iue., the general shape and timing, for new industrial chemical products :umluded in the sample in order to provide one classification scheme for lustorical product behavior. The implications of such an identification inns discussed in Chapter II. As long as any nth order explains variation hithe dependent variable above that accounted for by the remainder of the onkns and the overall mean of the dependent variable, it will be retained hithe polynomial expression. The significance level used for any standard ernnrof the beta weight associated with a given order, i.e., a decision cndterion to retain or delete any nth order expression, is set at the .OSlevel. Otherwise, orders will be drOpped one at a time (with new kmst squares equations calculated on the remaining variables), starting wiflithe variable contributing the least towards variance reduction until fluzsignificance criterion is met on an individual variable basis. General Hypothesis No. 2 The profit cycle for new industrial chemical products does not typically fit a declining exponential curve. 1 . The class of functions suggested by Bertram Schoner for incluSion intus developed stochastic model for the selection of research and devel0pment projects was of the type y = abx where y== the dependent variable a,b = constants X‘= the independent variable having a negative sign nuaind“3try for which he constructed his model sells highly technical 1 IHDdUCtS to other industries, not to the consumer. Certainly new chemica K Bertram Schoner, op. cit., p. 78. N.- J L \ o .- '.s a: ‘ ‘ l .5... . I N.C..» a ‘ cl. QII Y L t l . . v «M. y. fit. V I ”a ' “.5 'r1 'U ,4.. 87 guoducts in the chemical Specialty class should qualify for inclusion 1umer1fis classification scheme. For this is where the bulk of the emphasis on research and deve10pment work has been in the recent past within Huechemical industry. We particularly desire not only to pinpoint general shapes of profit data through time, but relate these experiences with those of sales histories. General Hypothesis No. 3 The profit cycle of new industrial chemical products does not typically descend while the sales curve is still rising in the maturity phase of the product life cycle. The timing of the profit cycle is another research issue facing a crUfical test. As previously stated, many researchers have suggested Hut the profit cycle reaches its peak before the sales cycle, implying thatpmaking of profits signals future weakness. We need to determine anextensive leading profit cycles are part of the product eXperiences ofindustrial chemical products. General Premise No. l The time horizon necessary to consider sales and profit contributions for new industrial chemical products exceeds five years beyond product introduction. The full economic life of a product would be an ideal time horizon fin murfinancial and marketing analysis of a new product proposal. Yet favindustrial chemical firms attempt to use the economic life quantitatively as unatime horizon. For among other drawbacks, the element of uncertainty isfklt to be an increasing function of time. Naturally, executives hope tmnzmany new products will remain viable for decades, yet analysts are Presently unable to predict with reasonable accuracy what that period of thmzactually is. In many cases where highly profitable new products remain viable, it serves no useful purpose to extend an analysis that far in the future. - I NIy . . ~ - A... . I --\ V» a» . ,_ _ "K A V t“.'-\.\ & 1'... "b ‘- -“t A . - A. ,~. ...~ ‘ ‘1 Lv .‘I ‘ h 'I d-. \.- u . . t l‘ «,4 . V ‘ c. .. ‘fi .‘ u 'A ‘A 5 ~ ‘A I u .. ,, . .1 a .. ~- 0 . '1 n .i ’ '1 ‘. . .x .. 88 Recent literature often seems to suggest that the obsolescence ratecflfexisting products is high and continually increasing. If this istnue, any attempt to project cash flows beyond the period during which dmeproduct is actually displaced in the market may fault the decision hiregard to commercialization. Emphasizing the management of working capital and liquidity may ohflxuct the selection of many potential new product prOposals. Under smflia policy, some minimum cumulative contribution to profits is imposed aszibudgetary constraint for a set of projects. This is not to say that mayn'industrial chemical manufacturers face such a prescriptive policy wflfldn.the context of individual product evaluation; it only becomes a snyuficant factor when various new product proposals are consolidated atthe corporate level, especially when retained earnings fail to supply Huzfunds required for its Operation, thus creating a dependence on outside financing. ngsification Schemes Used to Describe Performance Identified The following classification schemes will be used as indicators ofpmrformance in the evaluation of new product behavior: Total Sales: annual and aggregate Total Profits: annual and aggregate Rate of Growth (Decay) of Sales: annual Rate of Growth (Decay) of Profits: annual Rate of Growth (Decay) of Losses: annual Timing of Sales: years Timing of Profits: years OOVC\UI4-\wN:—‘ 0. O 0. Payback Period for Recovering Plant and Capital Equipment Expenditures: years 9. Accounting Rate of Return: as a percentage 10. Return on Investment: as a percentage 11. Discounted Present Value Sum: annual. aggregate, and critical turning point 12. Eli' -. : I). PIC. 13. Str'. .. . v V g -a::a:.es L. :emie: a: sea tc 15c 12.122295 i. , ..t ease 0: :h‘; I}; tI‘Mb‘. ‘ ‘2‘ p 2.5323: I. 1‘ I. _ nu: are l o._ . ‘ ‘05:, n ‘ ‘.‘ F 9 u .s ‘A .,_ I. I - w, v “a I“ ‘.. Q... ... b u . u‘.‘ .‘.’ ‘ ”a I m.“ 'I \ fl ‘~.. ._. . "._ "b u“ v .- n . ..‘ _“ ‘1‘ h. u .i' 'n‘ ' 4 l . «_ ‘Q 89 12. Equivalent Rate of Return: as a percentage 13. Profitability Ratio: as a ratio 14. Performance Index: as a ratio 15. Structure of Sales Patterns: distribution and timing Variables (8) through (14) are defined separately in Chapter III; the remainder are discussed with the research findings disclosed in Chapter VIII. Possible Factors Relating to the Performance of New Industrial Chemical Products Screening based on statistically determined probabilities will seek to isolate the factors which relate to performance. Potential influences on new product behavior are grouped into four broad categories for ease of treatment: those associated with market structure, buyer behavior, product characteristics, and related intrafirm experiences. ”Standardized testing procedures will be used. Of course, few of the relevant factors can be related in any simple fashion. These relation- ships are usually highly complex and obscure, which is why previous efforts to model product behavior have met with such limited success. If a stated hypothesis is supported at an appropriate confidence level, we may infer a conclusion. But on the. other hand, however, failure to support a hypothesis does not preclude any relationship, defined or otherwise; evidence is simply lacking. This apparently subtle difference becomes important when the research process seeks to reduce the number of variables involved in product structures to a manageable number. Since many factors were either judgmental or subjective, respondents for individual product histories were selected on the basis of their product knowledge or their access to those possessing such knowledge. The compilation of data involved the efforts of many different executives and analysts having diverse responsibilities. One can only hope that, in this process, the right questions were directed to the right people. .l' q . ,rwtaw L nu - v»..-L si. ». . . ... - *;~- " .\s..-:‘ . . , . ‘ .N "M‘ ... . -‘ ‘ x, v “a. . D . vux‘ . . . . \" ' ,4 a ’I‘; 4 I.’ 's,_ ‘ . " u r—v‘v ... .‘ . i. 'Y. r. 's n.‘ ' v 90 1. Market Structure A. Degree of Patent Protection Performance should correlate with the effect of any patent in pnflecting the product from external competition. B. Demand Trends in Derived Demand Situations Performance of chemical intermediates should relate to the rehnjve growth (decay) rates of Specific product markets upon which the (knived demands depend. Ignoring the cyclic, delayed effects of inventory balmums within distribution channels, two forces essentially determine commxfity product demand at the market level: demand for finished products thattnilize the chemical product in some fashion in their manufacture amithe rate of materials diSplacement via technological change. The fmner the growth rates of finished products and the more the tendency ofthe product to diSplace others, the faster will the product be pulled Huough the distribution channel (assuming adequate marketing coverage). Hus variable should be measurable in the performance records of the products under study. C. Duplication Difficulties by Competitors Patent rights and technological obstacles often thwart the attempts ofanw competitor to make important inroads in marketing an identical Prmhmt. Yet the difficulty of overcoming these barriers should be mnelated to achieved performance results. D. Extent of Capacity_Utilization at Industry Level Levels of plant utilization must relate directly to demand levels onhrwhen all producing firms are operating at full capacities; then Pnfiitability is determined by various managerial actions and process effufiencies. Unit product costs usually reflect changes in plant ......2-.' ‘In - a: up.o.-I§¢V‘O: 3:31 A 1 ; .10“. 0'... . . 154.. u L A 5 . " w... u» u I). ...: In . . w“ A. -~’ - ' .t-C "0" m. I...“ v »- i. '0 5 LT; ‘\ 1..” 'o. ‘v u 1] I s‘_ “.3"- ,._: ... N. ‘ \ n‘ \ _-. ... . ..' . 91 utilization; and they in turn directly affect the profit contribution of any given product. Those products with abnormally high idle capacity percentages on an industry basis should show price attrition to such an extent that it could damage existing average profit positions. This is most true when the demand for a product is inelastic, as long as relative price changes do not cause a shift in demand toward the product. E. Import Patterns Since imports are not sizable, particularly for new chemical specialties, the performance of new industrial chemical products is probably unrelated to the percentage of total imports to domestic sales of all producing firms. F. Market Share Where firms have commanding market shares, we expect to find generally better performance because of their controlling positions over critical decisions, such as prices, existing capacities and sources of supply. G. Market Trends The timing of new product introductions in relation to the capacity to supply demand appears to be a critical Operating factor. One expects superior performance from those chemical products in which the specific markets were experiencing strong increases in growth rates, assuming the firms have capacities sufficient to fill market requirements. H. Minimum Corporate Asset Size of Competitors Reggired to Compete Effectively Since the capability to commercialize many new industrial chemical Products often involves large research and deve10pment and plant and equip- ment expenditures, some chemical processors may not become competitors because of the lack of needed capital or adequate financing. Thus, those q . ,.-\'-ynan‘_~-‘ _. ..Zt...a... s nI-isy- ' . «to-...» (‘0 (I) f—fi :43 f1) m... .. ' v...,'. .- . ' "‘ ...... q | HI . Luv.“ a..‘_\ 4,.” . w“ .,_r . -9 . a .i 2 6‘. . \‘p “a: fi‘ ", .‘z ‘, q. ‘o '\ ‘x w a ‘\ ~ . x" \ . 92 few firms having the capability should be in a better position to offer the product in the marketplace more profitably, especially in early periods of commercialization when demand may be somewhat limited. This consideration reinforces the barriers of entry stemming from the technology required. I. Number of Consuming Industries The development of additional end-uses, one marketing strategy for expanding the total market for industrial chemicals, should generate greater returns . J. Number of Significant Competitors Performance should correlate inversely with the number of competitors. Competition does not necessarily have to take the form of price attrition, although this form has prevailed historically. If a high level of promotional activity is the only alternative Open to get product recognition, this expensive route may necessitate a relative loss in performance. The relationship hypothesized should even hold true for vary- ing demand levels, for products with high attainable sales potentials will probably attract other entrants. K. SuppLxL Characteristics of Factors Since we assume a direct relationship between raw material purchasing experiences and the prices of raw materials through time, profitability should decrease with shortages and difficulties of access to existing supplies of raw materials. 2. Met Behavior A. Degree of Backward Integration Large volume purchasers of Specific chemical products pose the constant threat of backward integration (if they have the technical and legal capabilities), for they seek greater economies in operation. Chemical manufacturers facing such a threat should experience an accelerated rate of be ex; 1 ‘ I G O in g 90 cl ‘ . "v.4. uni. b u M“. u." l - C . W"“*~ a! ' . ‘ ..., . ., ‘ ’ “-‘...‘5.‘ u.‘ ._\ ‘ O.) " .- ‘.‘ v 4 r‘~ .u‘ Ky ‘\ ‘ .‘ ,‘fi -,. _ n‘ H - ... ‘ ‘ v- .. " F H.“ ‘: Iv, 93 obsolescence which limits their expected returns, particularly when these moves are quite pronounced. B. Degree of Required Deliberation One expects no particular correlation between performance and the extent that the initial purchase requires deliberation. C. Dependence on Field Contact Work Performance and the extent to which the final sale depends on personal company contacts should be correlated in most instances. Personal contacts by the field representative may be necessary to generate the interest required to complete a sale. And the technical and informational SUpport which so many new industrial chemical products require may be supplied only by the company representative, often to the point that lost Opportunities with inadequate field coverage are costly to the firm. If this is true, then, we should find higher performance in products requiring personal contacts by field representatives, assuming adequate field coverage has been provided. The incremental costs of field coverage for any new chemical decrease if the product is either related to other product Offerings presently being detailed in some way or the consuming markets overlap those being detailed for other product lines. D. Effect of Industrial Advertisig on Source Selection Even though one eXpects improved performance with effective industrial advertising, such effects are possible over any range of performance records. The critical issue is determining the Optimum level 0f advertising expenditure. EXperimental advertising programs may be valu- able for firms who desire to better utilize their advertising resources. B. Effect of Product Quality on Source Selection Although the quality of new chemicals is an important consideration in selecting product sources, we have reason to believe that performance Will not be a correlate of product quality. ... - r.,_. . I r» ‘.I ~“‘ —— ...u u. ‘ . ‘n—nv 5". P A1 “a.“ C .. i 7 ln ... ..t ~ ~. .. I »' ..-.r , r -... \ "‘~ "' ~k,_ 2., . . 1‘ \; bun, ‘la. 5 .. .‘I .. ‘ ~., ‘ u; 'b 94 F. Effect on Sales of Related Products Since performance records used in this study do not reflect changes ingnofit contributions of related products, no relationship should appear between performance and the effect of a new product on the sales of related products. G. Extent of User LaboratogypEvaluation Performance shOuld vary directly with the percentage of potential umns that evaluate a new chemical product in their own laboratories using mqmrimental quantities. H. Level of Product Loyalgy Where their product loyalty is high, one generally eXpects higher remnnm from industrial chemical products (other factors being held constant) beanme buyer behavior is more stable and predictable. The producer would mqmct to receive the increase in purchase orders from buyers when their inmnzrequirements increase. Where these patterns are lacking, however, thepnOducer must compete just as vigorously for growth from these buyers as well as from potential buyers. I. Number of Contacts Required by Marketing and Technical Personnel Contact work presents a distinct challenge in product promotion shme typical industrial marketing situations are field intensive. The fluent of field commitment should correlate directly with most performance measures . J. Number of Product Sources An industrial buyer often finds it convenient to have more than omacontinuous source of supply for industrial chemical products. For ,. . "' to- ’ .n. .... an, uni-'Is'!’ . y' A a...) I.--‘ I bV. ...,- bus (II 1 ‘I-IOIA ...”. u 4 a 4-5 - .. F a. 't z} .... ., -. " ~ ...“ ....; 'au gas It) 95 this insures a continual supply when a producer may be unable to meet the buyer's needs, and it places constant pressure on individual producers to provide technical support and other needed services. On the other hand, such behavior stimulates price fluctuation through competitive action. Multiple sources thus tend to depress returns for those products that operate within this type of competitive situation. K. Number of Purchasers We expect superior performance from products with large numbers of existing purchasers . L. Number of Annual Purchases by Buyers Since purchasing patterns typically reflect inventory policies as well as extent of material usage by buyers, we cannot expect performance to vary on the basis of purchasing records, although large existing demand levels often necessitate high frequency purchase behavior. M. Recpgnition Experiences of Product Advantages bLUsers Any new industrial chemical product has advantages which may be more or less easily recognized by industrial users. Relative ease of recognition can spur product evaluation and eventual acceptance, and thus should relate directly to performance. N. Repptation and Image of Manufacturer A manufacturer's established reputation and image are thought to influence a user in selecting his source of supply, thus directly affecting product acceptance. 0. Igne to Educate the User The average time required for marketing personnel to communicate Product characteristics to potential buyers varies with the extent of g...‘-v"' 5’ E ,, A ino'vonvfih“ . .«u,--~- aov' r“ . Bu. “Hi-u- — ,. . . a -p , a Ins-«... . . .. ... 9‘ A a».’u-. . ' I... " ‘n-u ‘ I a. u At. .‘ 2; ‘-. ""' in". . 'u ‘-.‘ 7.. ‘1 u, . . , ,' A. A "v. ', ._ :1, 96 deliberation and the complexity of the evaluation process; it should be inversely correlated with performance. P. Trade Relations With Users Valuable information on user eXpectations for future requirements and other pertinent data can often be obtained from users and incorporated in product plans. The ease of access to this information, better where trade relations are established and favorable, can enhance performance. 3. Educt Characteristics A. Cyclical Patterns We eXpect cyclical patterns to adversely affect the performance of new industrial chemical products. B. Degree of Marketing Innovativeness The selection of unique distribution channels, the alteration of existing advertising strategies, or a change in normal technical support i11UStrate a type of marketing innovation which may improve the market POSition of the firm. Such factors are expected to affect performance levels decisively. C- Degree of Matching Between Technological Characteristics and Market Requirements Performance should be directly related to the degree to which technj-Cal Characteristics of a new product match the market requirements- One EXpects superior performance from chemical products where these two factors are extremely well matched, assuming that the product's full POtential was estimated accurately and that timely programs were develOped t0 Effectively meet the marketing Opportunities- ' V p :5!‘ II ~V' — D - ,v .t'llhl‘ """V-ehu 97 D. End-Use Patterns All levels of performance are possible in most end-use classifica- tions. E. Level of Technological Innovativeness Technological innovation in new chemical products (whether process or production oriented) should generate above average performance, which in turn reinforces continued emphasis on technological improvements. F. Price Movements Price declines are warranted in those cases where induced demand shifts result in net gains, even though these effects are often 188899- G. Product Differentiation Strategy Many have questioned the value of product proliferation within the marketplace for products serving the same general end-uses. While Some insist such a move represents economic waste, others argue that it Serves to thwart product declines and to Stimulate primary demand as well. I“ this study, we can assess the net effect of such activities, where present, on the behavior of new industrial chemical produCtS- H' Eeamh and DevelOpment Harnessing Experiences Performance is probably independent of the degree of difficulty of harnessing the research and deve10pment program in commercializing any new industrial chemical product. 1- Seasonal Patterns Where the demand for a new industrial chemical product is highly Seasonal, inventory requirements and production scheduling become more ....-“" '¥ ...A an F .u ' ...-i" "' 6" Ill) I., ’ - c-nfi fl .\| ;n._ a.“ . L \ 98 diffnnflt to plan. If single-purpose equipment is used which cannot be adapletO any other chemical process, production inefficiencies usually result. J. Specificipy of Use The more Specific its uses, the less likely a new industrial dumncal product is to find wide industrial application. K. Standard Industrial Classification (SIC) Code Attainable performance levels of Specific products are Often Ummght to be directly correlated with the overall economic behavior of theh:respective broad classification groups. L. Technical Service Requirement Chemical products vary widely on how much technical service support isrmeded in the field, according to the nature of the product and the degreeof technical sophistication acquired by users. Still, one expects noparticular pattern between performance and the amount of technical sendce support required to insure user satisfaction. M. Trends in Cross Margins Gross margin trends Often signal changes in profitability, although itifi the absolute level of the earnings stream that affects performance mOSt directly. N. Type of Product Recovering economic byproducts and c0products may generate shflfificant revenue. Still, there is no reason to believe that chemical Prmmmms having byproducts or c0products outperform those without, and no nflationship should exist between the type of product and performance. . -r ..an.‘"' ’ c '- ‘ .. .-~“““‘L . .— ‘1 VI ‘ - I 1. -C so» .\ ; ...u .. - wu—y n» v- .t l 1'. r-v", 99 Anytunmmtive evaluation technique should, of course, take into account fluaeconomic contributions of byproducts and c0products. 0. Type of Product Demand As long as the manufacturer estimated market potentials adequately andckmigned the new product program accordingly, we should expect a mnmrior return, regardless of the product's purported demand type. 4. Related Intrafirm Experiences A. Effectiveness of Product Flows in the Distribution Channel If product movement through distribution channels is irregular muiburdensome, both the level and profitability of Operations should be seriously limited. B. EXport Patterns Foreign markets in many cases may provide excellent Opportunities “Dutilize any existing idle capacity on a profitable basis. C. Extent of Plant Utilization at Firm Level Since unit production costs change markedly with changes in plant lHilization,*we normally expect better performance with Optimum levels ofpflant utilization. D- Lutensigy of the Selling Effort In cases where profitable demand may not be fully tapped, marketing canI’I'OVide broader market coverage in the field by altering the sales-call ratlOS for given prospects. ~ r .. ......c-»..4 ‘ , , - ..."... vv- .q PL 5...... ' ‘ ‘11.. 0.... was t u v u... _ ' ‘ a . ‘I-A-~.-u5. c 9 P bu u ,_-; ‘-v I u._..‘ . I II ‘0 ‘ ,. ~': A a . - ‘x~s n, . ...- -~.,_: 100 E. Investment Patterns Since both marketing and financial risks are generally unrelated tozhwestment levels on a product basis, we would not expect performance tolmacorrelated with the size of the investment commitment in fixed assets and working capital, with the exception of absolute sales and pnfifit data. A firm's investments in one single product are usually not suffuflently large to represent a significant portion of its total capital stock. F. Length and Number of Production Runs Performance relates directly to production run experiences, since Inofitability is thought to be sensitive to effective production planning. G. Leyel of Clarity of Product Demand Performance should vary directly with the degree of clarity of demand forearmw industrial chemical product, since those firms which can size up mumntial market demands well can also gauge their various operations more effixmively. Any sizing-up process would necessitate a fundamental under- Samming of what comprises current and future needs of industrial users. H. Leyel of Research and Development Expenditures Products with high research and development requirements represent greater risk8. and must generally have higher expected returns ‘30 Commflmate the firms for bearing those risks- 1. Licensing Experiences Although licensing arrangements are profitable ventures themselves, esmufiAIIY in securing foreign Operations, there is no reason to expect thelikElihood of licensing arrangements to differ with performance. u‘l ..~ — -! .. _, .c' lOl Licensing arrangements represent the sale of technology. Usual accounting practices assign these proceeds to corporate accounts only, not reallocating them on a product basis. J. Management Evaluation of Relative Success The performance of an industrial chemical may be evaluated through time using many criteria. Some argue that product performance should be measured against pre-established marketing Objectives to determine progress toward these goals. In matching actual outcomes with the subjective evaluations of performance, we may infer which measures of performance are used by respondents in determining the relative success of new products. K. Marketing_Costs Economies of scale associated with higher output levels should allow for the Spread of marketing costs on a unit basis over a larger base, which should then reflect in the performance records of new products. L. Merger Activities Since the analysis does not consider profit contributions on an hmtitutional basis (it is product oriented), performance should not correlate with merger activities. M. Mode of Production Any marked underutilization of unique plant and equipment in the Production of an exclusive chemical product should produce inefficiencies limiting profitability. On the other hand, products produced in common production facilities give management a great deal more flexibility in scheduling production more efficientlY- w- .l' u. _ . H... I “hrs. ,. mv- o--..‘o. . r u v .... ......L.. WA».. ' sAO. .. '_ a “ "-nsd»... V u 1" pa. , I. v _, at. .~“~I 'I-l u, ., '.. "H .:~ “.14- f, ‘5 ‘1 p’ .I ‘ i a u. ‘4 _ "a u ‘. . 4. '. '- .‘ . 102 N. Orientation of Research and DevelOpment Program Research-based products are said to be developed as a result of pure scientific investigation, with end-use applications detected through numerous screening programs. Many compounds coming out of research labora- tories have unique physical characteristics, but little or no commercial value. But marketing-based products are developed through applied scientific investigation only after particular user requirements become known. Superior performance, then, Should come with marketing-based products, which reflect the efficiency gained by knowing in advance the existence of specific product demand. 0. Output of Research Activities Performance should relate to the number of products uncovered and commercialized, each usually absorbing its fair share of the originally budgeted research and development costs associated with the research project. Then, if joint production is possible, further economies of scale are usually realizable. P. Product Concept Acceptance Performance should be directly related to the marketer's general ability to communicate the characteristics and benefits of the product t0 industrial users. Q. Product Improvement Efforts Once a new chemical is marketed, the firm can collect evidence to evaluate its performance characteristics against its end-use applications. If deficiencies appear in the investigation, product improvement programs can be scheduled whenever economically justifiable. Presumably those Products having product improvements recorded at least once in the Product life cycle under study will show a lower rate of product acceptance b‘P-fore the product changes are incorporated. Future sales behavior is ...n'! ' g - ...-Viv“ ... p u be. ..: .3.. ...c ........... . r r' ES Wort-but. . u . ......b... . A I F ‘ "~li§bo cl. y) “N ..a: u..." u ’ v~ ."u-t: ‘ ‘ n u... .. ‘t.. ...: a." w“ 103 affected by the ability of the firm to overcome the technical problems. The fact that problems did arise, however, should in itself limit future growth seriously, regardless of the actions taken to correct performance deficiencies. In other words, the sales structure can be effectively related to the degree to which a product does meet desired performance characteristics . R. Promotional Media Strategy Industrial products differ in the extent to which they are tailored to specific markets. Advertising should be focused on those media in which the resources of the firm are utilized Optimally in reaching stated objectives and intended markets; horizontal mass media are unlikely to provide the most efficient approach. S. Promotional Outlgy Trends Advertising can help create interest in investigational evaluations and application studies, although promotional programs are seldom critical in the success of new chemical introductions. If promotional outlays are important in demand creation, however, promotional outlay trends should relate directly to performance. T. Source of Product Discovery Many new products can still be very profitable to a particular firm even if they are mere duplications of existing chemical products. U. Suitability of Marketing Capabilities Marketers generally find that an experienced, viable selling organization is prerequisite to a successful new product deve10pment PrOgram. It is damaging to undertake the marketing of a new product Without the inmnediate capability to do so. Integrating a large number of new personnel, for example, might disrupt uniform market deve10pment until their overall productivity reaches normal levels. !. u r —-— . . n. “W”. J5‘toU-rm 5 ‘ " 'IID " '1‘ S ...... u. . .. 5.54., . ‘- I‘v'dai u. (‘0 (v: 1.. ‘ 2191“?” “‘“sti. yo I l u .4‘ ‘I . '. R. v -. _ '. m o « 104 V. Supply Capabilities in DevelOpmental Sampling Programs Developmental sampling programs can be useful in gaining product acceptance and subsequent usage. Most potential buyers of industrial chemicals work closely with field representatives in evaluating user requirements. But the firm's capability to supply significant amounts of product to meet estimated user needs at the time samples are distributed is probably an important factor in overcoming any resistance in having the new product evaluated. W. Technological Specialty Egperiences It seems reasonable that chemical manufacturers would concentrate on research and development programs in selective areas in which they have existing eXpertise. But since major chemical manufacturers so limit themselves generally, performance should not vary with technological Specialty requirements. X. Type of Distribution Channel Used Major chemical manufacturers have most often chosen direct sale to users, allowing distributors to handle small lot purchases. Frequently used distributors are probably complementary to the traditional channel and economically justifiable. Early in market deve10pment, many firms attempt to develop the markets directly for they can thus better control the destiny of new product programs. Product management generally feels that the technical and sales support needed must be sensitive to market needs, even when information on potential end usage is incomplete. On the average, we expect superior performance from those firms using the combined direct and indirect sales approach, as long as they handle the profitable large volume business on a direct basis. 4.. £1.5‘ .uC- .... ... Out-Avon» -u...‘._., I “nut.“ . “"‘v-I- . '9! .4 r- '~o...4..\_ 7 L. .IH'ii.‘ " . a’, m... c._ ‘ 1 1 105 Y. Type of Fixed Capital Employed Given fixed levels of capital investment and noncontinuous production, we eXpect superior performance from more desirable forms of capital equipment, such as all multipurpose equipment, where capital expenditures may be allocated to the various products being produced on the basis of actual output or time in use. And greater production efficiencies result too, as long as changeover and start-up costs are reasonable. 2. Type of Marketing Representation Used The marketing function is Often handled by special marketing development peOple, at least in the introductory stage of a new chemical product. The technically trained peOple working on deve10pment programs are most qualified to handle the technical aspects of product usage, so many feel that they are in a unique position to handle the selling effort as well. Others argue, on the other hand, that the selling function can best be handled by the regular sales force, which knows and understands its assigned markets and business Opportunities within those markets; if technical problems arise, they contend, the regular sales force can easily handle them through referrals. We hypothesize that superior product performance relates to assigning the marketing function to the regular sales force. We have the Opportunity to evaluate these general approaches to market deve10pment and their outcomes using empirical support. l. y... 9.. an .. .-..-»...n w. . I. “.O . gt. CHAPTER VII METHODS EMPLOYED IN DATA ANALYSIS Introduction Various mathematical and statistical techniques will be employed in the analysis of collected field data. However, this effort involves more than the reporting of statistical data. Our inquiry is directed toward whatever Specific conclusions emerge from the analysis of empirical data, a quest leading ultimately toward a better understanding of new produc t behavior . Polynomial Determinat ion A methodological approach based on curve fitting has obvious advantages, expecially since this is an approximation problem involving a search for a function which can be defined on a number of preselected variables, but whose parameters are unknown. Approximation problems most often fit polynomial, trigonometric and eXponential curves. Polynomial and exponential expressions are typically involved in nonlinear systems. Because a polynomial expression can be generated for any arbitrary continuous function in a finite, closed interval, an approximation of any system, linear or nonlinear, based on a polynomial determination can best fit the empirical data associated with product behavior. The approximation criterion for generating parameters in the function y(g,t), defined on sales or profit data over time, is the minimization of the sum of squares of absolute deviations. The following depicts this: KEY? 107 A_ 2 v Y so + 31th + gztzq + . . . gktkq + e (7.1) A where Y = predicted dependent variable go, g1, . . . gk = estimation coefficients t = time in years, starting with the first year equal to the value one q = number of observations, 1,2, . . . N v E N - l, the vth order of polynomial e = disturbance The sum of the squared residuals is to be a minimum, i.e., N Z (Yq - Yq)2 is a minimum (7-2) q = 1 The least-squares parameters, g0, g1, . . . gk, are found by matrix algebra} We need make no distribution assumptions regarding the dependent variable and the independent variables while calculating the least-squares Coefficients, although the independent variables are assumed to be fixed Variables. The dependent variable is a normally distributed random variable With 8 mean based on the values of the independent variables for each Observation in the pOpulation. A constant variance is assumed to exist Over all observations in this model, with independence among observations. The amount of squared error accounted for by the independent variables is measured by the coefficient of multiple determination R2, whiCh is expressed statistically as follows: ‘\ 1 o Joh . See Carl F. Christ, Econometric Models and Methods, (New York. n Wiley & Sons, Inc., 1966), pp. 380-395. . i “"" ‘. , , w... ”h“ . I rm .V ““wos . u‘u. ‘01.. » 108 R = l - (7.3) An initial least-squares equation is obtained using N - 1 independent variables defined over time for sales or profit data representing an individual product situation. Through a general stepwise procedure, we delete from the equation each order of independent variables that is unable to account for the variation in the dependent variable above that accounted for by the overall mean of the dependent variable and the remainder of the independent var iables . The F statistic for each least-squares coefficient in any given equation is examined to test its significance level. The minimum level 0f significance accepted for inclusion is .05. In other words, a least- Scluares equation is calculated after each deletion until the significance PrObability of each ng statistic left in the equation exceeds the EStablished minimum. The F statistic for the regression weight gi is defined as beta weight "i" squared divided by the square of its correspond- ing standard error. The end result is the maximum-order polynomial expression for sales or profit data of a new product that is justified Statistically, and comparisons between product histories can be made on the basis 0f their determined polynomial expressions. No independent variable erroneously included in the regression equation will bias estimates of the Other parameters even if multicollinearity (a problem of high Covariance between a number of independent variables that are highly correlated) exists between the variable in question and one or more other independent variables. We must pay particular attention to any covariance beCause the sample size is quite small. qua-{:8 a“ :v‘vivb nu ~.,...-.. ... .g ’ 4‘ , "‘O'Imh. ... v y- . . w .... '3! en . r .. a .. 'u’ _ 5‘ X l n‘l ... ' Q VA "‘5 .A“ 109 The real purpose in this exercise is not to estimate parameters in the equation but instead to best fit data historically for classification purposes and to forecast the values of the dependent variable for predictive purposes. Forecasting procedures are not likely impeded so long as the joint distribution of the independent variables is not altered in the forecasting period, even though the separate influences of the independent variables may remain unknown. An Approach to New Product Evaluation There is no substitute for an analytical approach in evaluating new product decisions. The characteristic problem Of a new product not Producing a satisfactory return once introduced in the marketplace can be offset if adequate resources are allocated to a comprehensive study of the entire marketing situation. A decision to curtail a project based on conclusions drawn from unfavorable empirical evidence could conceivably save the firm from serious loss. A number of analytical approaches find use in the evaluation of new products. However, whatever specific approach is adOpted, an incorporation of uncertainty in the process appears to be mandatory. Figure 7-1 details the information required to quantify the impact of eXpected new product behavior. The next section contains an explanation 0f how uncertainty is introduced as an evaluative factor. SUMMARY TABLE FOR NEW PRODUCT EVALUATION FORM ITEM.DESCRIPTION Cumulative Operating Cash Flows Cumulative Investment Requirements Discounted Operating Cash Flow Sum Discounted Investment Sum Present Value Sum E i i dTV1B dTV1B dTV1B chs CALCULATION where i(19),, '2 where 1(25hn '2 where l<27)mx<19>m '2 <27>x<25> “here _]___ m ‘2 where .(28)m '2 IA IA IA lA |/\ Hllm 110 F1sum: 7-1 lllllllllllll lllllll lllll‘llll mouse .80: Anuan: lllllllllll lull I‘ll... mun—39$ 3.39533! “3595 3501.3: 39:5 £33.33»: a sonic-8. ICE ll! llllllllllli ldfihefl 38 :58 no :8 .9335 .5330th annex—km mauuaeum case—69$ lill 3.598 Tlllllllllll‘ll lllJ \“aj Luau-L 2300-3 a ...-o as: e nauu‘uus 03.63 ~38. Scour—«av : Santana 3:25.550: can: aqueous—3a Soap 355235 1.1.33 can: Eaton PCS 3:88.53 cox; 3:323; A 358533— 3: Sana-955 9.3.38 [on 9:05- u>5 any: I»: H3310 9:23 A-eanau “3:“..an on ASE-o 3:05.85 “:33 can: 28.: :30 nuao.‘ can .03 I03 :38 flue-3.: 320k can 9333825: .55 “duos 3:398 9.1533238 3:: :50 3.8 its.» 35 aqueomun has :5 3.00 air—son: can? ...-segue vac-Sana anon—h, 0.. h‘ T 32.8 :5 can... mastu< a. .025 an i 03.5 9.3:.» 03 a" .'u~ IKCL ZD~Lfl 90:20.... 18.: / 2. ...: “on 3 «3:5,: assessment .3: pianisitpln taunt ira— D u... - _ . ...-A'..1: . u ova “u: a 1.7.8 1.." .1 ...... _‘ ‘ u. 1.. l . a. b u ... ‘w .- 1“ 111 Program AnaLys is A computer program titled PROGRAM ANALYSIS is available for performing the mathematical calculations under uncertainty for all economic and associated data required in the evaluation of new products, written in Fortran IV for CDC 3600 computer equipment.2 One subroutine in the program calculates the apprOpriate discount factor to be used in deflating cash flow values during a specific time period. The formula for continuous discounting in an instant at the end of a given time period may be represented as: l (DF)i = rt (7.4) e i where (DF)i = discount factor in time period i expressed as a fraction e = 2.7182818 r = selected cost of capital or risk factor expressed as a decimal t. = time period i expressed in years The reference point is time period zero, which for most new product situations within the chemical industry was eighteen months after approval of a given capital eXpenditure. The underlying assumption for all estimates of parameters is that the associated estimated error terms are normally distributed. There is a random deviate generator within the program. Any estimated parameter is calculated such that a sequence of random numbers is generated which has a normal distribution with a mean equal to zero and a variance of: one. In other words, values for estimated parameters are randomly selected around their predetermined average values. A unit normal sampling —_¥ A listing of the source deck can be found in the appendix. .....~ . U. ‘ Mun“; _ .. ,, \'I '51 ‘3 . Mn.“ ...N «...». ] 112 distribution results after a large number of passes is made over a set of product data. A definite pattern for each performance statistic can then be delineated. If one wishes to evaluate new products on the basis of any other type of error distribution assumption, the program must first be altered to reflect this change. An empirical study of actual estimation processes should offer the most rigorous way of establishing a basis for selecting specific error distribution assumptions. This computer program may quite easily be employed as a deterministic model of new product performance by setting all variance terms equal to zero. This, in effect, implies perfect measurement and prediction. Parametric Versus Nonparametric Methods Since the statistical technique chosen may critically influence the outcome of research findings, any statistical test is selected on the basis of the sampling distribution of the statistic employed. Since the sampling distribution is a theoretical distribution, it can be characterized if its exact distribution is known. And once the sampling distribution of a statistic is known, we can make statements about the probability of occurrence of certain numerical values of that statistic. Parametric tests are based on Specific assumptions with regard to the type of pOpulation from which the sample is drawn. The parametric approach is apprOpriate if the variables studied are continuous and normally distributed. Unfortunately, there is no simple way to determine whether or not the collected field data in this study meet the assumptions of normality. And it is quite difficult to estimate the extent of error introduced if these conditions are not met, particularly with small samples as in the present case. u}... ...-U I ‘ .0511 “'I ' , ‘ a... ...-at.» , . , \.~. -lr' v,v |v‘ ....b. ny- ./ 113 Many research problems are encountered in the evaluation of new yxoductS‘whose solutions suggest the use of nonparametric techniques. Efiegel makes a clear and concise comparison between these two statistical approaches, suggesting under what conditions they are employed apprOpriately. "A parametric statistical test is a test whose model specifies certain conditions . . . about the parameters of the pOpulation from which the research sample was drawn. Since these conditions are not ordinarily tested, they are assumed to hold. The meaningfulness of the results of a parametric test depends on the validity of these assumptions. Parametric tests also require that the scores under analysis result from measurement in the strength of at least an interval scale. ”A nonparametric statistical test is a test whose model does not Specify conditions about the parameters of the population from which the sample was drawn. Certain assumptions are associated with most nonparametric statistical tests, i.e., that the observations are independent and that the variable under study has underlying continuity, but these assumptions are fewer and much weaker than those associated with parametric tests. Moreover, nonparametric tests do not require measurement so strong as that required for the parametric tests; most nonparametric tests apply to data in an ordinal scale, and some apply also to data in a nominal scale." Emulwhen samples are drawn from different pOpulations, nonparametric tests unibe applied to determine whether or not given pOpulation estimates differ on a statistical basis. A number of statistical tests will be used for inferential Inmposes in the analysis of new product behavior; the indicated scale Vfill determine the test to be chosen. In situations where any sumtistical model includes two or more types of scales, the test fluu:allows the usage of the lowest desirable scale in terms of power effluflencies will be used. The significance level reported for any k-sample test hithe research findings is the probability of the calculated value ofthe statistic or any larger value given no relationShip. In the case of rankoH mo.o sea on unmoflmflcmflm uoc mm3 manmwum> mcemwmamu mafia .oomHnumxumE mnu Bonn camwpnufls Goon mm: uoacoun onu no: no umnuona mflmhamcm as» :H popnaonfi munch no woman: onu mm vmcflwon 122 H a n a x x x maom.m mamm.o m oohNH u Nuue.a Nmmm.o OH comma : x x x x ¢NNQ.N Ream.o 0H comma n u x x x qum.N m~wm.o m comma u a a u s HmHO.N oom~.o e comma s u s HmoN.H NmNN.o a CONNH x x x x x mqmm.m Nmma.o HH ooHNH u u u u x x mooo.m mmom.o u OOONH s x x x x onm.~ aoqm.o 0H oomHH u u a x x x ammm.~ maem.o m cowHH . u n x x x mmq~.~ camm.o w OONHH : a x x x x Naom.m omma.o m oooHH u n u x x x x x nema.o momm.o w oomHH n s u x x x x x o-o.m qmmm.0 w ooefia u u a u x qum.~ mmoo.o m conga a u u s x cmmH.H «New.o m cowaa u u x x x x x comm.m momm.o o codaa : u u n x x ¢m~¢.N dqaw.o m oooHH u x moqm.~ mmm¢.o 0a coaofi u x x x x x mnwm.m Nq¢¢.o OH oomoH n s u n x mNmO.N quo.o m oesoH n x nmmm.m womw.o OH oooo~ u n s u n x x x x nmmo.m mmom.o o oomoH : x x x x x x wmmo.m mmaa.o OH oo¢o~ u x x x x x x x omaw.m qmmo.o OH oomoH u u n u x x x x HNHn.m mmmm.o n oouo~ : x x N mqu.m mmwa.o OH ooHoH OH m w m o m a m H owumwumum cofiumcMsumuon Amado» may nonaoz pmawmuom Hmaaochaom mo umpuo nu: coaumz oHnfiufiaz Hounmoaxm omou ncunusn mo ESEHaHS nonpoum ucoeowmmooo o muHmm Hmscnm EsameE wchH>Hp kn use» mum aH noopoun ao>Hw w you poumHaono mH wovcH uHmoun 059 H mHmozucmumn aH :Bosm mum moaHm> m>Humwoz "maoz -- -- H.s e.m s.m HH.HV H.o m.N N.N m.H oH ookNH w.NH o.HH m.mH N.HH e.sH m.N Hm.HHV Ao.wv He.sHv AN.HHV m oooNH e.o N.e N.e e.m o.w s.H AN.NV Aa.sv HH.ev Hm.wv a oomNH -- H.s s.e 0.0 o.m s.oH N.eH N.®H m.mH e.m m oosNH -- -- -- -- m.e N.H e.H o.s m.o e.o oH oomNH -- aH.NHv Am.Hv Hm.ov Ho.Hv AN.NV Am.HV Hm.Hv N.o Hm.oV oH OONNH m.mH m.mH H.m N.HH H.mH o.NH m.HH N.s s.H H.o s NooHNH -- -- -- m.o s.s m.H s.o m.o m.o Hw.NV a oooNH N.e m.N Am.ov N.m m.N N.N a.o m.o As.Hv HN.HV oH oosHH -- -- Hm.sv aN.ov Hs.eNv Ho.nmv Hm.mqv Am.msv AN.mNV AN.msV oH oowHH -- -- o.N o.m s.N N.mH m.oH N.s Hm.ov Am.Nv m ooNHH -- o.s aH.NV Ho.ov Hm.NV He.HV Am.ev m.o H.o N.o a ooeHH -- -- m.m H.s N.e N.q m.s o.e N.N s.H oH oomHH -- -- w.mH s.eH N.HH w.N e.m m.e m.N HN.oV oH oosHH -- -- -- m.m s.m s.m o.oH m.m o.N H.m oH oomHH -- -- -- H.sH N.NH m.oH H.oH m.N m.m N.m oH oONHH -- m.mH m.NH m.HH o.o e.m w.s e.m e.H Ho.Hv oH ooHHH -- -- -- N.oH H.kH o.oH N.m a.» a.m o.m o oooHH o.o o.o o.o o.o o.o o.o o.o Ho.ssv He.HmHv Ho.msHv H ooaoH o.o o.o o.o o.o m.s o.mH w.oH m.m Hm.cv Ho.Hv m oomoH -- -- -- N.HH N.oH m.m s.e o.s o.s H.H oH ooNoH o.o o.c o.o o.o H.m H.m N.a N.m N.HH N.n N ooooH -- -- -- -- H.o m.e s.m N.H N.N N.o N oomoH m.NH e.m H.N N.H m.H e.m N.m e.H H.o Am.Nv oH oosoH m.w H.m N.oH ¢.oH s.m s.m m.s o.N m.H m.H m oomoH -- -- -- Ho.ov He.oV a.o m.o H.c Am.ov He.ov oH OONoH H.m N.m e.m m.e m.HH m.s N.m An.ov Hm.ov He.HV o ooHoH 0H m m N e m e m N H sass nosasz umow mean @600 uusmoum Nlm WAMHAVH. mfiobnomm AmH no.0 men um unmonHdem no: we? oHanum> wanHmaou ones .momHnuoxumE oSu aoum camupnuHa oppose can now no nonuoc3 mHmNHmcm use aH popsHocH munch no woman: on» ma pochon 125 noon can u H s a u N N N nnaN.N moqw.o w CONNH n N N N nmwm.N mmmm.o oH oooNH . N N N N N N wmwn.m mumm.o 0H oomNH s n N N N N N owem.m Hamm.o m ocqu a n u n u « momm.~ m¢¢¢.o o ocmNH u a N N N mumm.N wowm.o m oomuH N N N N N N N N ome.m ammo.o HH ooHNH u i n u N N N N Noom.m mama.o n ooomH a N N N N nmmN.m oamw.o 0H ooaHH n n n N N qmmm.N ommm.o m oomHH u a u N N N mHmn.~ momw.o m oohHH n n « scam.u anH.o m oooHH u u u N N N N moHH.m comm.o w OOMHH a n u N N N N mNMH.m wmmm.o m ochH u u u u s Nmo¢.H nwmm.o n OOmHH u a n n N Omwm.~ qwom.o u oomHH u u N N mmmm.~ mmom.o m ooHHH a n a n N N N N N MHHm.m mmmm.o n oooHH a N N N mmmo.m mmmm.o 0H ooaoH s N N N N N Ncmm.m moam.o OH oomoH u s s s N mmmO.N Hmom.o m oesoH a N N N Nooo.m ome.o oH ooooH u n u u u N N N N nmmo.m emmm.o o coon 3 N N N N N N N N nmow.m mmmm.o OH ooqu 1 N N N N N N N mmow.m mamm.o oH oomoH n s a n s mnmm.H mewH.o n couoH n N N N mNHm.N owno.o oH ooHoH 0H m w m c m a m N H oHumHumum coHumaHEhouoa Annum» aHv wonaaz vonHmuom HmHBoamHom mo nopuo nu: GOmumz oHnHuHaz dunnonxm omou saHnusa mo H anchHz uoamoum ucoHonmooo r; ~-~L nu .s// ‘Mu a. ‘ \‘ h .u .. ”1.. A. 141 TABLE 8-15 PROFIT STRUCTURE 0N ANNUALIZED NET PROFITS (LOSSES) AFTER TAXESI Class Interval Frequency Percentage Between $-l98,150 and $23,466 9 33.33 Between $23,466 and $99,387 9 33.33 Between $99,387 and $1,159,600 9 33.33 27 100.0 Median $ 43,600.0 Mean $148,343.2 Standard Deviation $290,229.6 1Found by dividing the number of time periods covered in the analysis into the aggregate net profits (losses) after taxes results. TABLE 8-16 TIMING OF PROFIT CYCLE Minimum Number of Years Showing Actual Profit Results Frequency Percentage Zero 2 7.41 One 1 3.70 Three 1 3.70 Four 2 7.41 Six 7 25.93 Seven 8 29.63 Eight 2 7.41 Nine 2 7.41 Ten 2 7.41 27 100.0 Median 7'0 Mean 2.6 Standard Deviation ‘v‘v‘r re...“ ‘ "2..., n Canal a Ll 2352.216 C . . a6 V a‘.~‘ 5.- ~.-' 1- w», 142 Annualized net sales correlated highly with net profit results, tmving a calculated Spearman rank correlation value of 0.8468 (see Table 8-17). Firms tended to select as key targets for market development those industrial chemical products that were above average in profitability in terms of their absolute dollar contributions. Whenever a product had a persistent loss position early in its life history, management usually sought to improve the problem in order to reverse the situation quickly or withdrew that product from the marketplace. As one would expect, return on investment calculations correlated with profit figures. Since they were not time-discounted, the patterns of change in profit figures were more closely linked with return on investment data than those indicators discounting cash flows through time, e.g., the present value sum and the performance index. Even though investment patterns were associated with achieved profit structures, the decision to invest or divest was more than likely based on sales, not profit expectations. Evidence indicated that decision unkers may have overemphasized the sales contributions of new products in their development activities, at least in the stages prior to maturity. Other determinants of product behavior statistically related to the various levels of profitability included the degree of challenge facing the scientific community in supporting commercial development, the thrust of promotional outlays, effect of industrial advertising on manufacturer selection, and the type of fixed capital equipment employed in the production process. The nonparametric test results are summarized in Table 8-18. “ issearc‘: Code "...: (.1 ‘68 ac! I later 11\ 143 TABLE 8-17 SPEARMAN RANK CORRELATION TESTS OF ANNUALIZED PROFITS (LOSSES) AFTER TAXES (RESEARCH CODE VARIABLE 153) AGAINST SELECTED VARIABLES I..— Research Code Variable Rank Significance Number Variable Correlation (one-tail) 2 Accounting Rate of Return 0.6435 0.0001 4 Aggregate Net Profits (Losses) 0.9878 «40.0001 After Taxes 5 Rate of Growth of Net Sales -0.0446 0.4127 6 Rate of Growth of Net Profits -0.0570 0.4005 7 Rate of Growth of Net Losses 0-2000 0.4000 8 Timing of Sales Cycle 0.2253 0.1293 9 Timing of Profit Cycle 0.7356 <:0.0001 10 Payback Period -0.6740 0.0001 11 Equivalent Rate of Return 0.6026 0,0004 12 Profitability Ratio 0.5800 0.0008 13 Return on Investment 0.6062 0.0004 19 Cost of Capital -0.4573 0.0082 20 Market Share Statistics 0.1745 0.1920 21 Aggregate Research & DevelOpment 0.0177 0.4651 Expenditures 54 Median Yearly Incremental lnvest- 0.2772 0.0808 ments 118 Mean Cumulative Investment 0.6667 0,0001 Requirements 126 Extent of Plant Capacity Utilization -O.l735 0.1934 147 Annualized Discounted Present Value 0.4921 0.0046 Sum 149 Performance Index 0.5702 0.0010 152 Annualized Net Sales 0.8468 «c0,0001 154 Critical Turning Point for Present -0.6085 0.0004 Value Calculations 5031 ..4 144 TABLE 8-18 NONPARAMETRIC STATISTICAL TESTS OF ANNUALIZED NET PROFITS (LOSSES) AFTER TAXES (RESEARCH CODE VARIABLE 153) AGAINST SELECTED GROUP VARIABLES Research Code variable Variable Test Number (Number of k-Classes) Statistic Significance 73 Research and DevelOpment Har- Kiefer T 0.0344 nessing Experiences 4.2519 (5) 88 Outlay Trends for Product Kruskal-Wallis H 0.0149 Promotion 10.4853 (4) 90 Effect of Industrial Adver- Kruskal-Wallis H 0.0055 tising on Manufacturer 10.4179 Selection (3) 124 Type of Fixed Capital Employed Kruskal-Wallis H 0.0077 (3) 9.7431 ‘iesearch Code Tariable Ember h '3 L 5 145 TABLE 8-19 SPEARMAN RANK CORRELATION TESTS OF TIMING OF PROFIT CYCLE (RESEARCH CODE VARIABLE 9) AGAINST SELECTED VARIABLES Research Code Variable Rank Significance NUmbsr Variable Correlation (one-tail) 2 Accounting Rate of Return 0.7028 ‘=0.0001 5 Rate Of Growth of Net Sales 0.0197 0.4612 6 Rate of Growth of Net Profits 0.2456 0.1353 7 Rate of Growth of Net Losses -0.3l62 0.3419 8 Timing of Sales Cycle 0.4465 0.0098 10 Payback Period -0.7019 <=0.0001 11 Equivalent Rate of Return 0.7344 <=0.0001 12 Profitability Ratio 0.7238 -<0.0001 13 Return on Investment 0.6623 0.0001 19 Cost of Capital -0.3109 0.0572 20 Market Share Statistics 0.1674 0.2020 21 Aggregate Research & DevelOpment -0.1267 0.2644 Expenditures 118 Mean Cumulative Investment 0.3241 0-0495 Requirements 126 Extent of Plant Capacity Utilization -O.3791 0.0256 147 Annualized Discounted Present Value 0.5730 0.0009 Sum . _ 149 Performance Index 0.7128 ‘0.0001 152 Annualized Net Sales 0.5383 0.0019 153 Annualized Net Profits (Losses) 0.7356 ‘=0-0001 After Taxes 154 Critical Turning Point for Present -0.7034 ‘=0.0001 Value Calculations 1 :E 21:: 4'” i s a deal a sectzrs it I! 5 x1 146 4. Growth Characteristics Just over three-fourths of all the products reviewed had median 1mm sales growths greater than that of the national economy. And the same number of products exceeded the average annual growth rates of the chemical and allied products industry as well, one of the fastest growing sectors in the economy. Dynamic development in new products, then, offers the firms that can react sensibly to changes in demand and supply the rewards Of new business. TABLE 8-20 MEDIAN GROWTH RATE OF NET SALES Class Interval Frequency Percentage Between -53.6% and 15.1% 9 33.33 Between 15.1% and 24.8% 9 33.33 Between 24.8% and 56.4% 9 33.33 27 100.0 Median 19.1 Mean 17.4 Standard Deviation 21.2 1 M 147 TABLE 8-21 MEDIAN GROWTH RATE OF NET PROFITS Class Interval Frequency1 Percentage Between -100.0% and -1.0% 7 31.80 Between -l.0% and 31.8% 8 36.40 Between 31.8% and 56.7% 7 31.80 22 100.0 Median 14.6 Mean 9.5 Standard Deviation 35.5 1An indefinable situation existed between profit and loss data in one product history such that no meaningful disclosure of its growth rate was possible. TABLE 8-22 MEDIAN GROWTH RATE OF NET LOSSES1 Number Frequency2 Percentage 22.0 1 25.00 -11.l l 25.00 -4l.7 1 25.00 -76.3 1 25.00 4 100.0 Median -26.4 Mean -27.0 Standard Deviation 42.0 1Negative rates of growth of losses should be interpreted as reductions in loss positions through time. 2 . An indefinable situation existed between profit and loss data Zulone product history such that no meaningful disclosure of its growth rate was possible. EE’I‘ i251 vari; {na:‘:s a -. 3.1.411“. .1 . 35’. 01115.0 333211 ETC . v a” Iraq“ ' n‘v‘u‘. “3' ha nu... 5.1L ...,'_ -_ '1... tm ;'v‘.,.‘ ““33 ‘I. I I "IV- usuL“. 'h‘yo ... a I... 148 Few of the rank correlation tests of growth characteristics against those variables examined which were continuously scored revealed statistically significant results. Table 8-23 lists the variables involved and their test outcomes. A rank correlation value of 0.6217 between sales growth and profit growth strongly suggests that they are interrelated. As market deve10pment of a new chemical product resulted in above average revenue gains, Opportunities for notable profit eXpansions became enhanced. New chemical products with fairly Specific applications had no better performance records on either attained sales or profit levels. But the rate of profit growth did relate (see Table 8-24). A concerted effort at product promotion in Specific application areas can be more efficient than shotgun selling. The critical issue is identifying and communicating widlthe key buying influences of large consuming firms. Since many industrial sectors are concentrated regionally, the location of manu- facturing facilities in close proximity to user locations can realize immortant distribution cost savings. Rapid sales growth stimulated increased research and development work in.related areas which was partially defensive in character. For once a {xmition of marked penetration appeared in any product area, the firm was willing to support additional research in related areas in order to maintain this position, exploiting as well new discoveries that would add greater dePth to its product line. Marketing-based product programs tended to have higher profit growths