OPPORTUNITY RECOGNITION PROCESSES: A TAXONOMY AND OUTCOME IMPLICATIONS
Gaylen
N. Chandler, Utah State University and Jönköping International
Business School
Jonas Dahlqvist, Jönköping International Business School
Per Davidsson, Jönköping International Business School
ABSTRACT
INTRODUCTION
HOW OPPORTUNITY IS DETECTED
CLASSIFICATION
OF OPPORTUNITY DETECTION PROCESSES
EXPECTED OUTCOMES OF THE
SEARCH PROCESS
DISCUSSION
IMPLICATIONS AND CONTRIBUTIONS
CONTACT
REFERENCES
TABLE 1
TABLE 2
FIGURE 1
We create a taxonomy of opportunity recognition processes for new business initiatives being developed in 136 firms that registered first sales in Sweden in 1994. The three categories include proactive search, reactive search, and fortuitous discovery. Results indicate that initiatives discovered through a proactive search are implemented more rapidly than those discovered through reactive searches or fortuitously. However, as time passes, the advantages in implementation speed are reduced and in some cases nullified. This implies that the opportunity discovery process has an impact on the timing and speed of implementation, and may have measurable longer-term impacts on profits.
Shane and Venkataraman (2000) posit that entrepreneurship is the intersection between opportunities and enterprising individuals. The vast majority of published entrepreneurship research focuses on individuals and opportunity exploitation processes (Busenitz, et al. in press); there is little published research focusing on opportunity detection processes. An emerging body of research seeks to address this deficiency by focusing on the cognitive processes used by entrepreneurs to detect opportunity (e.g. Gaglio, 1997; Hills, Shrader, & Lumpkin, 1999; Singh, Hills, Hybels, and Lumpkin, 1999). The logic of this stream of literature is that individuals create schemas that represent cumulative experience, learning, feelings and meanings and that these schemas are used to help entrepreneurs focus on the most promising opportunities. While a focus on how entrepreneurs think may be enlightening, there is also a lack of research that revolves around the behavioral processes that might be employed during the opportunity recognition phase.
Thus, rather than analyzing how entrepreneurs think, we seek to better understand what they do. An understanding of how opportunity is detected is an important key for understanding the entrepreneurial process. Yet, there is little theory informing the process, and even less empirical work that identifies, compares, and contrasts opportunity detection processes. A notable exception is Shane (2000), who studied behavioral processes and showed that a unique combination of market and technological knowledge allowed people to detect opportunities without specifically searching for them. Although in Shane’s study opportunities were discovered fortuitously, Drucker (1998) argues that most innovations result from a conscious, purposeful search for opportunities. The process of opportunity detection, whether through fortuitous circumstance or purposeful search, may have implications for the subsequent development and exploitation of the opportunity. Thus, our research has two central questions: (1) Can a taxonomy of opportunity discovery processes be developed? (2) How does the opportunity discovery process influence subsequent gestation behaviors and performance outcomes?
We begin with a discussion of opportunity detection processes. We then use cluster analysis to create a taxonomy of opportunity discovery processes that include: (1) proactive search, (2) reactive search, and (3) fortuitous discovery. After developing the taxonomy, we evaluate whether the opportunity discovery process is linked to subsequent gestation behaviors and performance outcomes in a sample of 136 emerging business initiatives within 136 six-year-old firms.
Although an emerging body of research seeks to identify and codify the cognitive processes used by entrepreneurs in recognizing opportunity, there is little research in the entrepreneurship arena seeking to identify and codify the behavioral processes employed by entrepreneurs in recognizing opportunity. There is, however, some discussion in the economics literature regarding how opportunities are discovered.
Search for Opportunity
Drucker (1998) claims that most opportunities are discovered through a purposeful search. The theoretical foundation underlying the search for opportunity is found in neoclassical economics. Caplan (1999) summarizes the Stigler (1952) neoclassical economics approach that has been adopted in management by March and Simon (1958) in their work on search. From this perspective, opportunity detection is the outcome of a successful rational search process (Caplan, 1999) in which competing alternatives are sought out and evaluated. Savage’s (1954) expected utility theory specifies when a search is likely to occur. Decisions to search or not to search are contingent on the perceived utility of a successful search and the perceived probability of completing a successful search. If the perceived risk is too high given the expected payoff, the search for opportunity is not initiated. Because there are time and resource costs involved with search processes, entrepreneurs with more resources available will be likely to be involved in more extensive search processes. Uncertainty is recognized in each decision, but it is expressed in the form of known probability distributions (Lucas & Sargent, 1981); thus, failures that occur when incorrect decisions are made are expressed as a probability in a known function (Caplan, 1999). From the neoclassical perspective, such searches can be conducted only when relevant information about the technology and/or the market allows individuals to rationally define what they are seeking for (Savage, 1954).
The “classical” model of search is difficult to apply to actual search behavior on the firm level (Simon, 1957). The concepts of goal satisficing and bounded rationality open up a more realistic search model. Levinthal and March (1981) distinguish between search stimulated by failure, which is assumed to be more narrow and focused, and search encouraged by slack which is less controlled, broader, and with a wider variance of outcomes. Cyert & March (1963) refer to search stimulated by failure as “problemistic” search. They postulate that such a search will initially be directed towards the local environment and will continue until an acceptable alternative is detected. Problemistic search is distinct from search encouraged by slack because of the specificity of the problem. A more open-ended search may be undertaken for a variety of reasons but only at a cost. Since search processes have associated costs, organizations need to have slack, i.e. the firm needs to satisfice its stakeholders to such extent that they find it worthwhile to continue to contribute to the firm. From this, we can detail two major situations why firm would look for opportunities. In some cases, firms search for opportunities as a solution for specific problems expressed in terms of dissatisfaction with the performance of present activities. We refer to this type of search as reactive search. In other cases, firms search when they can. Their search activities may be initiated for a variety of reasons such as formal strategies, norms or sub-group interests, but a prerequisite is the existence of organizational slack. We refer to this type of search as proactive search.
The “Eureka Experience” Discovery of Opportunity
The discussion on search does not include the possibility that opportunities may be detected through fortuitous processes. The neoclassical assumptions associated with search are violated whenever decision-makers are unable to “look before they leap” (Savage 1954, p. 15). Therefore, in contrast to search processes, Kirzner (1997) proposes that most opportunities are discovered through fortuitous circumstances. In Kirzner’s view, neoclassical assumptions of complete knowledge are replaced with an assumption of ignorance about new products/services and how the market will react. Kirzner (1997, p. 72) states “What distinguishes discovery . . . from successful search . . . is that the former (unlike the latter) involves that surprise which accompanies the realization that one had overlooked something in fact readily available.” The ability to recognize an opportunity depends on the individual’s unique knowledge set with respect to technologies and markets. Because individuals are not omniscient, any given individual cannot recognize all entrepreneurial opportunities. This has lead to a conclusion that many discoveries do not spring from a successful search process, but rather somewhat fortuitously based on the unique knowledge base of the entrepreneur enables discovery when one is not searching (Fiet, 1996; Shane, 2000). The Kirznerian view holds that future opportunities are not likely to be discovered through theorems including relative frequencies or error-learning models.
Unlike neoclassical theory, Kirzner’s (1997) perspective allows for mistaken discovery. Some plans, as a result of initial entrepreneurial error, turn out to be mistaken. This provides a realistic explanation for individuals who think or hope an opportunity is available and charge forward heedlessly. Kirzner (1997) states that these errors tend to be systematically eliminated as market and technological experiences reveal the infeasibility of some courses of action and the hitherto unnoticed profitability of other courses of action. Flawed plans can be corrected over time through the responsiveness of entrepreneurs to their own errors, or the alertness and imaginativeness of other entrepreneurs (Cockburn et al., 2000; Harper, 1996). Thus, Kirzner’s model specifies a non-equilibrium starting point, does not require complete information, allows individual choice to change conditions, and allows for opportunity recognition errors.
Expected Outcomes of Opportunity Detection Processes
The type of opportunity detection process employed has implications for the exploitation process. In accordance with our definition, proactive searches for opportunities are conducted under conditions of relative resource slack. Notions of what constitutes slack have varied widely (see Bourgeois, 1981, and Lant, 1985 for comprehensive reviews). The conceptual and empirical difficulties with defining slack have been an obstacle to research on this topic. Overall, slack implies that the organization is not operating at complete capacity. In other words it has resources that are not utilized completely in the ongoing production process. This implies that the organization will have resources available for product/service development, marketing activities, and other events and processes that must occur before the product /service is ready for sale or delivery. Thus, we expect firms employing a proactive search process to complete gestation behaviors for a new product or service and achieve sales and profitability sooner than firms employing reactive search, or fortuitous discovery.
Firms employing a reactive search process are faced with resource constraints. They are searching because they have no slack. Firms with too little slack may discover an opportunity. Yet because of severely constrained resources they are less able to apply resources to product/service development and other gestation activities that must occur before the new initiative can achieve sales and profitability. Thus, we expect firms employing reactive search processes to be less likely to complete the gestation processes necessary to launch a new product/service and also to be slower in achieving sales and profitability.
Opportunities discovered fortuitously are likely to be more diverse than those discovered through search processes. Because they are more diverse, we also expect them to be exploited through experimental effectuation processes (Sarasvathy, 2001). Hence, we expect slower commitment of resources and more variation in exploitation outcomes.
CLASSIFICATION OF OPPORTUNITY DETECTION PROCESSES
The current study is guided by the basic theoretical principles outlined in the search and fortuitous discovery processes discussed above and informed by a limited amount of reported empirical work. It initially seeks to develop a taxonomy of detection processes. After deriving the taxonomy of opportunity discovery processes we seek to determine the extent to which the opportunity detection process employed has an impact on subsequent gestation behaviors and outcomes.
Methods
The current study is a longitudinal panel study, which focuses on the opportunity recognition process employed by the directors of 121 firms, all founded in 1994, and that were initiating a significantly new product or service in December, 2000. In addition to the survey in December 2000, the companies were surveyed subsequently in June and December of 2001.
Sample
This study employs a unique and well-developed data set. The sample for this study was selected from a panel with an original size 7256 new firms. The sample was created through a stratified random sample of all new business registrations in Sweden during 1994. The sample size contained approximately 30% of the firms in the target population. Since its inception, the panel has been systematically surveyed in the years 1995, 1998, 2000 and 2001.
In order to find firms involved in new business initiatives, a two-stage screening procedure was used on the 4950 firms that were still in the panel by the end of 2000. As a first step, a mail questionnaire was sent out in October 2001. Usable information was obtained on 4692 firms, of which 500 asserted that they had at least one new business initiative in progress. In December 2001, the second step was to run a telephone survey with the 500 firms selected in the first step. The telephone survey allowed for a more elaborated questionnaire and with a more rigorous specification of what constitutes a new business initiative. As a consequence, the final sample retained for further analysis was reduced to 250 firms. This final sample of firms received a mail questionnaire in June 2001. Usable answers were obtained from 136 firms resulting in an effective 54% response rate. Figure 1 describes the multi-stage process used to obtain data from respondents.
We analyzed any possible response bias between responding and non-responding firms. First, the two groups were checked for equal distribution of firms between industries (one-digit level of the Nace Rev. 1 industry classification) using cross-tabulations with X 2-tests. No significant differences were found. Second, response and non-response cases were compared on sales figures 1995–1999, sales growth 1998–1999 and the number of employees in 2000, both globally and for each of three broad industry classes: manufacturing, distributive trades and services. Since variables in the sample were not normally distributed, non-parametric tests (Mann-Whitney U/Wilcoxon W) were used to analyze between-group differences. The only significant difference (p = 0.027) between the response and the non-response groups was the number of part-time employed men (10–35 hours/week) in the service sector. Given that more that fifty tests were performed, this result might be expected even from a randomized process.
Measures, Analytical Methods, and Results
Two sets of measures were used in the analysis. The first set, designated as “clustering variables” include 16 items that asked specific questions about the search and discovery process. The 16 specific items, included in Table 1, were used in a hierarchical clustering analysis to determine whether respondents could be grouped into meaningful clusters based on their responses to search and discovery process questions. Cluster analysis is an exploratory technique that groups observations in a manner that maximizes between-group variation, and minimizes within-group variation. Complex patterns, not readily apparent through heuristic observation may emerge through careful empirical analysis. Identification of underlying relationships is one of the appropriate uses of cluster analysis (Everitt, 1980). An agglomerative hierarchical method using Ward’s (1963) criterion was employed in the analysis based on studies of clustering algorithms that show this method to be among the most reliable (Milligan, 1980). Data were standardized prior to the analysis (Milligan, 1980). There is no standard technique for determining the correct number of clusters (Green, 1978; Harrigan, 1985; Punj and Stewart, 1983). However, we were guided by theoretical factors. A priori, as described in the theory section of this paper, we had identified three possible processes: (1) proactive search (2) reactive search and (3) fortuitous discovery. The three cluster solution includes three distinct groups with mean values of clustering variables for each group distributed in such a way that cluster membership substantially conforms with the descriptions of the theoretical processes. The specific items, mean values, and ANOVA F-tests showing between group differences are included in Table 1.
The empirical clusters are generally supportive of the theoretical development earlier in this paper. However, there does not appear to be a pure “Kirznerian” group. The Proactive Search cluster includes elements of both search and recognition of opportunity due to unique knowledge skills and abilities. The Reactive Search cluster is distinguished by responses indicating that respondents felt forced to search for new opportunities because of poor performance. The Fortuitous Discovery cluster is distinguished by the significantly highest scores on the unexpectedness of the event when the entrepreneurs were not searching for a new opportunity, with relatively low scores on other dimensions.
EXPECTED OUTCOMES OF THE SEARCH PROCESS
Data were collected at three points in time (December 2000, June 2001, and December 2001) in order to evaluate whether the opportunity evaluation process has an impact on subsequent gestation behaviors and performance. These measures included both quantitative and qualitative indicators of progression towards exploitation of the new business initiative. At all three survey waves we asked:
In addition we asked respondents to report on the stage of development of the new product/service (1) idea or concept, (2) initial development, (3) tested on customers, (4) ready for sale. After initial data were gathered, we resurveyed respondents in June 2001. In June, we repeated the above questions as indicators of progress in the initiative. In December, 2001, respondents were contacted again. We asked the questions enumerated above a third time. We also asked them to report the total sales and profits for the company during the past year and the proportion of sales and profits contributed by the new business initiative. Initiative sales and initiative profits were calculated by multiplying the total sales/profits by the proportion of sales/profits. The performance questions posed in December are self-reported objective measures of performance. Chandler and Hanks (1993) showed objective measures of performance to be reliable when self-reported. Results are reported in Table 2 .
The results displayed in Table 2 indicate that initiatives based on opportunities discovered through proactive search tend to be developed more quickly. However by the end of the first year the differences are becoming smaller. Of particular interest is the finding that by the end of the first year, opportunities detected through fortuitous discovery have closed the initial gap in sales, and on average are more profitable (although not significantly so) than those detected through search processes.
The cluster analysis results indicate that new business initiatives can be clustered into a meaningful taxonomy system consistent with theory. It appears that new opportunities are discovered through systematic proactive and reactive searches, as well as fortuitously. It also appears that the opportunity discovery process carries over fairly strongly into the initial phases of the implementation process.
The June, 2001 results indicate that new business initiatives discovered through a proactive search process tended to be implemented faster than those discovered through reactive searches or fortuitous discoveries. In companies where initiatives were discovered through proactive search processes they discussed the new product or service with customers sooner, they were more likely to acquire new machines or equipment sooner, give the initiative its own budget sooner, they had people assigned to work on the initiative sooner, and they completed a business plan sooner. This indicates that business ideas discovered through a proactive search process are implemented faster than those discovered through reactive searches or fortuitously.
In contrast, the December, 2001 results indicate that the initial results are short-lived. After the passage of six months more time, the results indicate that the differences have narrowed. For example, the stage of development of the initiative has progressed slightly further for those initiatives discovered through a proactive search, but the gap seems to be closing. Also significant, is the finding that initiatives discovered through a reactive search are less likely to be given their own budgets by the end of the first year. Of particular interest is the finding that initiatives discovered through a proactive search, although implemented faster, do not appear to be more profitable than those discovered through other means at the end of the first year of operations. In particular, a few new initiatives that appear to have the highest profit potential fall in the fortuitous discover cluster. Although it is too early in the process to tell, it may be that opportunities readily discoverable through search processes have inherently less potential than those discovered fortuitously.
IMPLICATIONS AND CONTRIBUTIONS
This is an on-going research project with additional rounds of data collection scheduled. However, our initial results indicate that the opportunity detection process does have implications for the initial implementation processes, and may have implications for longer-term performance. The general pattern of results seems to indicate that the types of opportunities discovered through proactive search processes tend to be faster to implement, but perhaps less enduring than those discovered fortuitously. These trends will be verified through ongoing data collection and analysis.
This research has the potential to make several contributions to the entrepreneurship literature. First, although the “opportunity” is thought to be an important component in the entrepreneurial process, there is little published research dealing with opportunity from a theoretical perspective, and a major dearth of research that focuses on the characteristics of opportunities and their discovery processes. The current research provides an empirical test of the outcomes of the opportunity discovery process in a real-time longitudinal study.
CONTACT: Gaylen N. Chandler, Dept. of Management and Human Resources, Utah State University, Logan, UT 84322-3555; (T) 435-797-2365; (F) 435-797-1091; chandler@b202.usu.edu
Busenitz, L., West, G. P. III, Shepherd, D., Nelson, T., Zacharakis, A. & Chandler, G. (Forthcoming). Entrepreneurship in emergence: Fifteen years of entrepreneurship research in management journals. Journal of Management.
Caplan, B. (1999). The Austrian search for realistic foundations. Southern Economic Journal, 65(4), 823–838.
Cockburn, I. M., Henderson, R. M., & Stern, S. (2000). Untangling the origins of competitive advantage. Strategic Management Journal, 21, 1123–1145.
Chandler, G. N. & Hanks, S. H. (1993). Measuring the Performance of Emerging Businesses: A Validation Study. Journal of Business Venturing,. 8(5), 391–408.
Cyert, R. M. & March, J. G. (1963). A behavioral theory of the firm. Englewood Cliffs, N.J.: Prentice-Hall.
Drucker, P. F. (1998). The Discipline of Innovation. Harvard Business Review Nov./Dec., 149–157.
Fiet, J. (1996). The informational basis of entrepreneurial discovery. Small Business Economics, 8, 419–430.
Gaglio, C. M. (1997). Opportunity identification: Review, critique and suggested research directions. I. J. Katz (Ed.), Advances in Entrepreneurship, Firm Emergence and Growth. 3, (pp. 139–202). Greenwhich, CT: JAI Press Inc.
Green , P. E. (1978). Analyzing Multivariate Data. Hinsdale, IL: Dryden Press.
Harrigan, K. R. (1985). An application of cluster analysis for strategic group analysis. Strategic Management Journal, 6, 55–73.
Hills G. E., Shrader, R. C., & Lumpkin, G. T. (1999). Opportunity recognition as a creative process. In P. D. Reynolds, W. D. Bygrave, S. Maingart, C. Mason, G. D. Meyer, H. Sapienza, & K. G. Shaver,(Eds),. Frontiers of Entrepreneurship Research. Wellesley, MA: Babson College.
Kirzner, I. (1997). Entrepreneurial discovery and the competitive market process: An Austrian approach. Journal of Economic Literature, 35, 60–85.
Lucas, R. & Sargent, T. J. (1981). Rational expectations and econometric practices. Minneapolis: University of Minnesota Press.
March, J. G. & Simon, H. W. (1958). Organizations. New York: John Wiley.
Milligan, G. W. (1980). An examination of the effect of six types of error perturbation of fifteen clustering algorithms. Psychometrika, 45, 325–345.
Punj, G. & Stewart, D. (1983). Cluster analysis in marketing research. Review and suggestions for application. Journal of Marketing Research, 20, 134–148.
Savage, L. (1954). The foundation of statistics. New York: John Wiley.
Schumpeter, J. (1934). The theory of economic development. Cambridge, MA: Harvard University Press.
Shackle, G. (1955). Uncertainty in economics. Cambridge: Cambridge University Press.
Shane, S. & Venkataraman, S. (2000). The promise of entrepreneurship as a field of research. Academy of Management Review, 25(1), 217–226.
Shane, S. (2000). Prior knowledge and the discovery of entrepreneurial opportunities. Organizational Science, 11(4), 448–469.
Simon, H. A. (1955). A behavioral model of rational choice. Quarterly Journal of Economics, 69, 99–118
Singh, R. P., Hills, G. E., Lumpkin, G. T., & Hybels, R. C. (1999). The entrepreneurial opportunity recognition process: Examining the role of self-perceived alertness and social networks. 1999 Academy of Management Meeting Best Paper Proceedings. Chicago, Illinois, August 10.
Stigler, G. J. (1952). The theory of price. New York: Macmillan.
Ward, J. H. (1963). Hierarchical grouping to optimize an objective function. Journal of the American Statistical Association, 58, 236–244.
2002
Babson College. All Rights Reserved. Last Updated March 2003.