Organizational

 

The most important variables affecting the use of research are organizational structure (formalization and centralization), technical quality, surprise, actionability, and researchermanager interaction.

 Organizational structure variables have a particularly large effect on re search use.

 The degree of centralization and formalization (or lack there of) seems to outweigh life cycle ma turity, research purpose (confirmatory and exploratory), and to an even greater extent the report attributes in its influence on what and how much market research in formation is used.

 The more decentralized and less for malized firms are more likely to make greater (and perhaps better) use of the research they subcontract than differently structured companies.

Although our study was not designed to explore these possibilities, more decentralized companies con ceivably allow lowerlevel managers to take greater part in research activities, thus ensuring their commitment to the results of such activities.

 Also, the greater the in volvement (and hence interaction) of managers with re searchers, the more the former know what to expect and what not to expect and the less surprised they are likely to be by the final results.

 This interaction also provides researchers with insights about how they can best struc ture the contents of the final report in terms of the sev eral report characteristics studied.

 Less formalized structure would permit managers to be cognizant of the "business realities" facing their firms.

 These concerns could be communicated to researchers as a basis for making recommendations which are within the resource capabilities of the firm.

 We believe our findings about organizational structure are very significant in at least two related respects.

 First, in terms of management or organizational behavior, they are an empirical indication of the major impact of the internal environment of firms on marketing decision making.

 This aspect of marketing management is little researched.

 Second, our findings reflect the impact of organizational structure variables on the consumption behavior of organizations.

 The need to study the impact of structural variables on purchase behavior has been noted previously although little studied (Bonoma, Zalt man, and Johnston 1977) in the buyer behavior litera ture.

 These findings suggest that postpurchase or con sumption behavior is affected strongly by structural variables as well.

 This possibility merits much further exploration by researchers concerned with industrial or organizational buying behavior.

 Another factor found to be very important is surprise.

 Surprise is an important moderating variable between the research requested and produced and its use.

 Surprise can be thought of as a reality test which helps managers decide whether or not to use certain research results.

 If the results are intended to reinforce prior feelings (or information), a high degree of surprise would lessen ac ceptance of a research report's findings.

 The same would be true even in the case of explora tory research but to a lesser extent.

 Some degree of sur prise would be tolerated, particularly if the research de sign also had a partly confirmatory purpose (and the surprise did not occur there).

 Information is needed to reduce uncertainty, not to increase it.

 Even positive sur prises imply a change in status quo which in turn may create problems.

 These are not considerations that man agers weigh lightly.

 Surprise, then, is used as one arbiter of what information is accepted and used and what in formation is questioned or discarded.

 Surprise is especially significant in explaining the im pact of report attributes on use.

 The quality attributes in particular have 71% (for form quality) and 38% (for con tent quality) of their total effects explained by indirect effects via surprise.

 High quality enhances use partly by lowering the level of surprise which ordinarily inhibits use.

 Surprise also influences report attribute ranking.

 Al though actionability and political acceptability were ex pected to be the highest rated attributes, one reversal did occur: content quality was the highest rated attribute.

 An explanation for this discrepancy is given in the discus sion of the value of the surprise variable.

 If managers use surprise as a criterion in their decision to use or not use research findings, the magnitude of the surprise has implications for the behavior of the manager.

 Imagine a situation in which predominantly confirmatory re search was desired by managers and unanticipated find ings were produced by researchers.

 Managers must find a basis for criticizing the findings.

 Because they cannot comfortably tell researchers that they do not like findings which do not coincide with expectations, managers can select one (or several) of the research report attributes to censure.

 Content quality may be selected for several reasons.

 Managers often do not inform external researchers about how decisions will be made on the basis of the information provided.

 Therefore the actionability attrib ute may not be readily selected as a basis for criticism.

 Additionally the political acceptability of research results involves the managers themselves and their colleagues.

 Researchers are usually not well informed about the po litical environment within client firms (although other data indicate they would like to be thus informed; Desh pande 1979b).

 Hence political acceptability as a basis for criticism is not easy to employ.

 The way the results are presented (in terms of the form of the report) might seem too trivial an aspect to criticize.

 The final attribute, and the one most likely to be chosen, is the technical quality ("content quality") of the research done.

 It is generally not difficult to find one or more technical lim itations in any significant research undertaking.

 One can simply give special prominence to these limitations when the results are surprising.

 It is notable that data from the 16 interviews strongly suggest that managers are much more inclined to pay attention to research methodology and to do so critically when the research results are surprising.

 When the results are largely con sistent with expectations, little attention is given to the research methodology section of the report.

 Finally, managers do not appear to believe that the maturity of a product or service in its life cycle has a great impact on the use of research relating to that prod uct or service.

 Even the indirect effects of maturity on research use do not contribute very much.

following implications.

 1.

 The structure of an organization should be examined carefully to detect any inhibitory effects on research use.

 To the extent that an organization does or could make frequent use of research, an alternative organizational design (i.

e.

, decentralized responsibility) which en 25

 hances research use might be considered.

 Redesign may be temporary for a specific research project.

 For exam ple, research in the organizational behavior literature suggests that certain organizational structures may facil itate the initiation of new projects (i.

e., a decision to conduct a major research study) and yet inhibit the im plementation of those projects (i.

e., taking action based on the research) and vice versa (Duncan 1972; Hage 1980).

 Thus an organization which is highly centralized (which may facilitate the decision to conduct a major research project) may have difficulty in implementing results of the study.

 Centralization appears to inhibit re search use (see Tables 5 and 6).

 Thus, for purposes of implementing the research, a highly centralized firm may wish to decentralize decision making temporarily in the marketing area, at least during the implementation phase.

 Similar "switching rules" (Duncan 1972; Zaltman and Duncan 1977) can be developed for other relevant or ganizational structure features or variables.

 2.

 Managers should provide researchers with more infor mation about the decisions to be made on the basis of the research they produce.

 3.

 Personal interaction between managers and researchers is very important in creating trust in the researcher and consequently in the results of the research.

 The quality of personal interaction affects manager perceptions of the overall quality of the report itself.

 4.

 Providing a research agency feedback about the use/ nonuse made of the research is especially important if that agency is expected to have a continuing relationship with the firm.

 5.

 Researchers who favor an exploratory style of research should be especially sensitive to managers' tendencies to want confirmatory research containing few surprises.

 One respondent used the term "comfort zone" to refer to a manager's tolerance for counterintuitive and/or un expected findings.

 The larger the comfort zone, the more tolerant and accepting a manager is of such information.

 Special efforts will be necessary to widen comfort zones among managers if the results of exploratory research are to be accepted.

 Managers themselves might try to iden tify how they would react to unanticipated results in re search they commission.

 Comfort zones might be wid ened by managers and researchers generating a large array of possible research outcomes prior to the conduct of the research (Deshpande and Zaltman 1981).

 Use of a dialectic procedure, as has been employed by Xerox Corporation among others, might generate a range of outcomes which would make any particular outcome ap pear less surprising (Mitroff, Kilmann, and Barabba 1979).

 This approach could be facilitated by generating deliberately skewed responses to research questions for use by managers in a brainstorming mode prior to the research.

 Such responses are likely to include surprises and hence sensitize managers to the possibility of un anticipated research outcomes.

 The task of broadening comfort zones appears to have received little attention in management contexts and certainly is worthy of further inquiry