This article tests the insights and predictions of venture success as offered by reporters and experts in Inc. magazine, to the predictions generated from an analysis of data from a venture screening questionnaire. The venture screening questionnaire, consisting of 85 items covering four broad categories: (1) Individual Characteristics; (2) Entrepreneurial Behaviors; (3) Strategy; and (4) Environment, was wed to evaluate 27 "Anatomy of a Start-up" articles from Inc. magazine. The creation of the questionnaire was guided by the following premises: Individual Characteristics. We hypothesized that the chances of venture survival would be improved if (2) entrepreneurs had substantial knowledge and ability at the beginning of the start-up story; (2) entrepreneurs gained knowledge and ability during the start-up process; and (3) entrepreneurs continued to demonstrate substantial knowledge and ability at the end of the start-up story. Entrepreneurial Behaviors. We hypothesized that entrepreneurs who expended more effort in any of the following activities would be in new ventures that survived compared to entrepreneurs who expended less effort: Finding and Refining the Opportunity-comprised of 9 different activities, such as, defining the purpose of the business, planning analyzing competitors; Acquiring Resources and Help-comprised of 15 different activities, such as, finding investors, getting advice from lawyers, getting a loan, acquiring technical expertise; Operating the Business-comprised of 5 different activities, such as, dealing with distributors, managing the day to day operations of the business; Identifying and Selling to Customers-comprised of 15 different activities, such as, identifying specific customers to sell to, selling to customers, managing sales channels; Outside of the Business Issues-comprised of 4 different activities, such as, dealing with family problems, spouse, and friends. Strategy and Environment. The strategy and environment variables were characteristics requiring comparisons of the relative performance of new firms vis-a-vis other competitors and their industry characteristics, much like the questions used in PIMS research: first to entry, degree of innovation, rate of industry growth, size of market, relative price, and relative quality. There were 28 questions in this section of the instrument. We hypothesized that niche oriented strategies and high growth environments might be strategy and environmental characteristics common to startups that survived. In total, there were 85 questions that comprised the venture screening questionnaire. New Venture Survival. The measure of new venture survival for this study was a determination of whether the new venture described in each Inc. magazine article (Longsworth 1991) was still in operation as of January 1995. This date is nearly 4 years after the last case study that we analyzed was published (September 1990), and nearly 7 years after the first case study was published (February 1988). We were able to determine that of the 27 new ventures profiled in the "Anatomy of a Startup" series published in Longsworth (1991), 17 of these ventures were still in operation. A discriminant analysis was performed that resulted in seven variables that correctly classified 85% of the cases into new venture survivors or non-survivors. New ventures that survived were more likely to have: (1) entrepreneurs who gained knowledge and ability during the founding process; who devoted greater efforts to (2) dealing with suppliers; (3) analyzing potential new entrants and who (4) devoted less time to determining the identity of the business; businesses that had (5) "fundable" resource requirements (6) focused on products or services that were designed or produced to order; and (7) were in high growth industries. The classification accuracy of the the model was much better than industry experts (55% correct), competitors (55% correct), venture capitalists and financiers (40% correct), and customers (38% correct). Even though the discriminant analysis was better able to predict venture survival or non-survival compared to the experts, there are significant limitations to the reliability and validity of this one particular model, and the data set used. The primary value of this exercise involves making obvious the variables that observers use to make judgments about predicting venture success. One of the frustrations we experienced in analyzing the expert's predictions was our inability to glean consistent and general "rules of thumb" about new venture success from their observations. We conclude by discussing the value of academic research on new venture success predictors vis-ci-vis other avenues of inquiry and expertise:popular journalism and practice. (C) 1998 Elsevier Science Inc.
