18 Journal of Marketing, January, 1970

Buyer Behavior and Related

Technological AdvancesJOHN A. HOWARD

The author offers his views regarding consumer behavior research, theprogress in this field, and the impact of developments on market re-searchers, marketing executives, company presidents, public policy makers,and basic researchers. In the author's opinion, the integration of thesedevelopments represents a major breakthrough in marketing with im-portant implications for the 1970s.

66TTOW MUCH more will I sell if I double my-^•^ advertising?" the manager asks. "How

should I analyze these facts?" the market researcherwonders. "Is the concept of a marketing informa-tion system a sensible one?" many people are asking."How do I explain to this Congressional committeethat advertising will not change basic eating hab-its?" a company president contemplates. "Shouldthe government provide consumer information pro-grams?" a Congressman poses to his colleagues."What is the process by which a buyer's attitudetoward the brand causes him to buy or not buy thebrand?" a university researcher muses. The so-lution to these and many other important marketingproblems are greatly influenced by the state of ourknowledge of buyer behavior.

The decade of the 1960s has been a time of rapidgrowth in consumer behavior research and practice.Drawing on the theoretical contributions and em-pirical evidence of behavioral scientists in otherfields of endeavor as well as their own creative work,marketing scholars such a.s Bass, Bauer, Ferber,Frank, Green, Kasgarjian, Kuehn, Massy, MyersNicosia, and Pessemier have made major contribu-tions to our knowledge of the processes of consumerbehavior.i Much of this work has already resultedin substantial improvements in marketing operations.

The quantity of facts and propositions on thebuying behavior of consumers accumulated by com-panies and researchers is enormous. However,many of these bits of information or propositionsstand alone and are used in the context of a specificdecision situation. What is lacking in this approachis the leverage to be gained from a comprehensivestructure which makes possible substantial predic-tions about consumer behavior. The need for sucha structure to guide research and to improve market-ing practice is obvious. It is this author's opinionthat efforts in the next decade will be directed atthe empirical validation and continuing refinementof such a theory.

The list is too extensive to document here; see J. A.Howard and J. N. Sheth, The Theory of Buyer Be-havior (New York: John Wiley & Sons, Inc., 1969),pp. 421-429.

Based upon the splendid work of the men citedabove, a comprehensive theory of buyer behavior hasbeen developed at the Columbia Business School.-This theory integrates relevant behavioral proposi-tions from learning theory, cognitive theory, andexploratory behavior theory, with other ideas fromthe behavioral sciences and systems analysis. Thetheory posits the nature, relationships, operatingcharacteristics, and flows that exist in the consumerdecision-making process. Given a knowledge ofexogenous variables, the theory relates outputs (be-havior) to inputs (intemal and external stimuli)through hypothetical constructs which describe theprocesses internal to the consumer.

Developments such as those noted result in anoperational technology' which will have major im-pact on the science and practice of marketing. Thistechnology represents a systems view of the market,a view which distinguishes between the controllableand uncontrollable that play upon the buyer, andwhich brings relevant factors into a quantitative,coherent, unified, operational picture.

The immediate importance of this technology' isconfined to the narrow topic of buyer behavior andto the interests of the market researcher. It hasalso far-reaching implications for the marketingmanager, developer of marketing information sys-tem, company president, public policy maker, anduniversity researcher.

For Company Market Researchers

Market researchers will collect less data but willderive more information from it. In applying theseideas they will contribute to the technology and tofitting the theory to their particular markets.

First, the question of which facts to collect willbe more systematically decided; i.e., those factsspecified by the theory. The researcher can be morediscriminating instead of collecting facts which hehopes will be useful.

Second, collecting facts from the same person attwo or more points in time—a panel—will be more

- Same reference as footnote 1, 458. pp.

Marketing in the 1970s 19

common. Panels have rarely been efficiently used.̂Third, more systematic data storage procedures

will be adopted. Casual observation indicates thatresearch departments' ways of storing facts havebeen a serious handicap to the extensive analysisessential for adequate diagnoses. There will be, forexample, greater accessibility to "marginals" whichare both important in themselves and an essentiallink between the questionnaire and the analyses thatfollow.

Fourth, statistical techniques which incorporatecausality and its increased explanatory- power will bemore widely applied. Multi-equation models willenable us to deal rigorously with the complexity thatcharacterizes consumer behavior in response to mar-keting effort.^ From each period's data, laggedcorrelation can obtain more specific information aboutthe nature of the relations'" so that in the nextperiod the multi-equation model can be more pre-cisely specified to the particular market. In thisway, the researcher can increase the precision ofhis methods from period to period. A goal of thistechnological development will be to get the "noise"level down; i.e., to reduce the unexplained variancein the dependent variable. Experience indicatesthat the crucially important variables of advertising,price, and selling effort are most subject to the"noise" problem, which limits linking them directlyor indirectly to purchasing behavior. This linking isessential if real progress is to be made. The theoryis deliberately designed to focus upon the marketingor controllable variables.

Finally, in every analysis there is the task of in-terpretation and logical extension of the conclusionsbeyond the data. Systematic diagnosis is currentlyso severely handicapped that it is almost never done.Even if diagnosis were so highly valued that effortwas assigned to it in the face of horrendous timepressures, the appropriate data are not identified,the constructs are not validated, and they aretreated piecemeal instead of as parts of a system.If appropriate data were identified, adequate meansof collection would not be employed because of theexaggerated fear of contamination. If appropriatedata were collected, they would be stored in sucha way as to render them inaccessible for further use.Assuming these data were readily accessible, cur-

3 Winston L. Ring, "Adjustment for Measurement Ef-fects in Panel Data From Test Markets," paper pre-sented at 1968 AMA Fall Conference, Denver, Colo-rado.

*J. U. Farley and W. L. Ring, "An Empirical Test ofthe Howard-Sheth Model of Buyer Behavior," to ap-pear in the proceedings of the Third Annual Confer-ence on Buyer Behavior, Columbia University, May,1969.

5 T. V. O'Brien, "Information Sensitivity and the Se-quence of Psychological States in the Brand ChoiceProcess," unpublished doctoral dissertation, ColumbiaUniversity, 1969.

rent techniques of analysis would not be up to thesedemands. Finally, even if all prior conditions werenot met, diagnosis would probably be stopped be-cause by and large we have not had the necessarytheor>' nor developed the practical skills to use it.This description is a bit harsh but not much.

For Marketing Managers

"Marketing manager" refers to anyone who isconcerned not only with marketing, but also withthe total marketing picture irrespective of his levelin the corporate hierarchy. The technology willyield the marketing manager more information withless data because the system is more eflBcient. Thesystems view will lay before him a quantitativereview of the consequences of his marketingalternatives.

This integrated, causal picture of the influencesaffecting the buyer will provide a powerful leverto the marketing manager's understanding of themarket by giving him concepts which cut throughthe maze of complexity and which will encouragesharp diagnoses of marketing failures and successes.These diagnoses will be adding to his comprehensionof the market, enable the marketing manager tobuild a marketing philosophy which aids imaginativeinnovation and prevents him from being misled byminor aberrations of the market. More logical andelaborate control—performance measures—will bebuilt into his marketing procedures. His observa-tions of and conclusions about the meaning of thesemeasures will enormously strengthen the moresystematic diagnoses that follow.

Also, the premises underlying his .strategy willbe more explicit and articulated. These will notonly guide him but also will facilitate communica-tion. In relations with superiors, such as in de-fending his marketing plan and its supportingbudget, he will be better able to secure approvalbecause management will be in a better position tocomprehend and appreciate his arguments. Forthe same reason his relations with his peers willimprove. His directions to his subordinates will bemore explicit, better understood, and thereforebetter executed. Some of the differences in viewsabout the probable effectiveness of different planswill diminish. The better communication will im-prove relations with the market researcher. Thus,

• ABOUT THE AUTHOR. John A.Howard is professor of marketing inthe Graduate School of Business. Co-lumbia University. After being com-missioned by the Ford Foundation inl%0 to describe the state of knowledgein marketing, he then turned to theorydevelopment and more recently todirecting a large project devoted tothe empirical testing of buyer behaviortheory by cooperating with companiesin test markets.

20 Journal of Marketing, January, 1970

the manager can assume leadership in identifyingand designing the necessary studies instead of pas-sively reacting to proposals from the market re-search department. The manager can usually betterperform this role than the researcher because he ismore aware of the appropriate corporate goals andof broad market issues.

A variety of operations research and managementscience techniques lie fallow because of the lack ofbehavioral input. The technology will provide thisinput for some of these techniques which willsimplify the manager's decision making andplanning. In fact, with the behavioral relationsbetter understood, simpler optimizing techniqueswill suffice.

For Marketing Information SystemsThe new technology will facilitate the develop-

ment of a marketing information system. Such asystem typically implies a commitment to collectfacts on the same variables for several periods oftime. The theory will assist this task (1) byspecifying the facts, and (2) by suggesting waysof collecting facts. Moreover, the statistical tech-niques could form an essential part of such a system.

The comments made here, however, provide noanswer to the central question of whether the factsor derived information will "fit the manager'shead." An important but subtle phenomenon isthat each experienced executive has a "model" ofhis market which is not only unarticulated butusually totally subconscious. Yet his "model"strongly shapes which facts he considers relevantand how he interprets them. Until we have a betterunderstanding of what these "models" are and howthe manager processes his information, our ca-pacity to build useful information systems will belimited. Some evidence suggests that in a giventype of decision situation, managers use very similarinformation processing approaches*, but more con-clusive evidence is needed.

For Company Presidents

The technology has implications both internallyand externally for the president. Internally, theconcepts will make it easier for him to communicatewith his marketing personnel and because of thebetter communication his respect for them will beenhanced.

Equally, if not more important in this day ofconsumerism, company presidents will be betterequipped to explain and often justify their prac-tices to the public and to government agencies. In arecent Congressional hearing, the presidents of twolarge companies were asked by the chairman why

"J. A. Howard and W. M. Morgenroth, "Information-Processing Model of Executive Decision," Manage-ment Science, Vol. 31 (March, 1968), pp. 416-428, atpp. 425 and 426.

they did not use their advertising funds to changethe basic eating habits of the poverty stricken. Thehearing tran.script indicates that the presidents didnot answer the question. These very able menknew intuitively that such an approach was im-practical, but they lacked the concepts to adequatelyarticulate their experientially-derived beliefs tothose less versed in their particular domain of ex-pertise. This is a problem that all experts face,but those in a theory-based field can usually per-form far more effectively.

For Public Policy MakersConsumer interests have become a politically

relevant issue, and the many able Congressmenwho are directing their energies to formulatingand passing appropriate legislation will find thistechnology useful in developing a multi-faceted, ra-tional consumer policy. Concepts they need forcomprehending the buyer will enable them to com-municate more easily with the scientists in the field,with the philosophically-oriented individuals whorelate the scientists' facts to social values, and withthe special interest groups. The same technologybut with a somewhat different orientation in studydesign and interj^retation can pro\'ide specific datato guide their decisions. One of the questions iswhat kind of evidence will be accepted because therehas been a growing tendency to accept statisticalinference as evidence.

For University ResearchersIn a splendid recent review of a definitive book

in market segmentation, Ferber states: "It (thebook) suggests in analyzing consumer purchasebehavior we are still at the stage where the princi-pal difficulty is not how to answer the questionsbut rather how to formulate the questions andobtain meaningful data."^ The new technology im-plies that we are now ready to move beyond thisstage.

One of the major tasks is to acquire a fuller un-derstanding of the relations among the variablesin the system. O'Brien^ and others have shownthat this can be done. Some of the relations arenot only nonlinear but nonmonotonic over the rele-vant range. Hence, some of the relations may betoo subtle for field research and the laboratorywill play an essential role. The new technology toowill be facilitated because testing ideas that arepart of a system instead of an isolated hypothesisgives leverage in the development of new knowl-edge : control is easier to build in and interpreta-tion is much more productive.

T Robert Ferber in a review of William F. Massy,Ronald E. Frank, and Thomas M. Lodahl, PurchasingBehavior and Personal Attributes, Journal of Eco-nomic Literature, Vol. 7 (June, 1969), pp. 483-485,at p. 485.

« T. V. O'Brien, same reference as footnote 5.

Marketing in the 1970s 21

Another area of research will be directed towardreducing the high level of "noise" found in theresearch to date so as to improve prediction. Re-search on the relations among the variables willhelp by making it possible to specify the systemmore precisely. The timing of changes in the vari-ables deserves particular attention. Much of theresearch, however, will be methodological. Forexample, more refined measuring instruments andbetter ways of analyzing data to determine causalityare essential.

ConclusionLarge accretions of knowledge—often break-

through.s—seem to occur when different and oftenapparently unrelated strands of development beginto merge. The new technology represents one suchmerger, in fact, a set of two mergers. First, thereis the merger in psychology. Cognitive, leamingand motivation theories have begun to come to-gether. Second, two types of research in marketingthat were largely separate in the period 1955 to1965—the quantitative and behavioral approachesto buyer behavior—have since joined into a singlestream.

Another but different kind of merger appearsimminent—the manager and the market. Manage-

ment science and operations research have largelybeen concerned with normative models. However,these models made no pretense of representing themanager's thought processes. They were substi-tutes for these processes. One is beginning to ob-ser\e that some of the highly competent people as-sociated with this management science-operationsre.search point of view are attempting to buildformal models that are integrated into a manager'sthinking processes, which currently seem ratheridiosyncratic* This view will merge rapidly withthe new technology, and a most exciting period inmarketing should be the consequence. The newtechnology can replace many of the subjective esti-mates with objective data. Thus, because a man-ager's underlying premises will be more accurate,his decision process will be more systematic.

Finally, and for the sake of perspective, we mustremember that ultimately, as in all good science,the product of this merger will also carry the seedsof its own destruction and yield to still betterstructures than we are now capable of envisioning.

9 John D. C. Little, "Models and Managers: The Con-cept of a Decision Calculus," paper delivered beforethe symposium "Behavioral and Management Sciencein Marketing," University of Chicago, June 29 toJuly 1, 1969.

MARKETING MEMO

Marketing Ourselves to Ourselves . . .Companies take pains to project a sound external corporate image these days, but

few pay much attention to their in-house image with their own employees, accordingto Ernest Dichter.

Yet an internal corporate image is important and likely to become more so, saystbe bead of tbe Institute for Motivational Researcb. Among otber reasons: tbe tigbtlabor market, tbe restlessness of today's youtb. a need for bigb employee morale inindustries dealing directly witb tbe public, and tbe burgeoning mergers and acqui-sitions trend, wbicb often creates confusion among employees as to wbat tbey reallybelong to.

—"Tbougbt Starter: Pay Attention to^ Your Internal Corporate Image." Busi-

ness Management (April. 19691. p. 15.©1969 by Management PublisbingGroup, Inc. All rigbts reserved.