Among the inputs that executives rely on to run competitive organizations, a particular type resists classification: ideas. Companies need a steady supply of ideas, and they often turn to external venues for help. Say, for instance, a company is developing an advertising campaign that needs art and text to convey a targeted message. If the company lacks the internal resources to produce content, it can use an outside venue—a crowdsourcing platform—to solicit ideas from artists and writers. Through the platform, the firm can collect ideas from a broad, diverse pool of people unencumbered by geography, bureaucracy, or complicated logistics. Indeed, the company can traverse the ideation process in a remarkably efficient way.
Research on ideation outcomes has, thus far, focused somewhat bluntly on the raw number of ideas produced in a given scenario, typically by means of an ideation contest. HKUST’s Tat Koon Koh and colleague argue that researchers can now take a more nuanced look at different ideation stages, pointing out that “there is a lack of attention on intermediate ideation outcomes” such as the number of ideas that are considered before a final idea is selected. To fill this research gap, the authors study how contest participants, known as “solvers,” make their way through the idea-generation steps. Their analysis reveals how solvers 1) scan for initial ideas, 2) select a set of promising ideas to pursue more deeply, and 3) ultimately select ideas worthy of submitting.
To investigate these dynamics, the researchers designed their own ideation contest, allowing them to observe, at a granular level, how contest participants behave. In the contest, solvers were asked to choose photographs to accompany articles published by a consumer-goods company. Some solvers were shown examples of photographs that might (or might not) have reflected the company’s tastes and sensibilities; other solvers were not shown any examples. Each solver’s progress was recorded, tracking the time spent in all three stages of the ideation process. The authors were particularly interested in measuring how solvers behaved in the earlier stages; that is, in how they allocated time between selecting initial photographs for consideration and identifying which photographs to analyze more thoroughly due to their potential to contribute positively to the company’s articles.
Among other discoveries, the experiment found that solvers who looked at example photographs ultimately generated fewer final ideas than those who saw no examples. In addition, say the researchers, “the quality of ideas [was] lower.”
By showing how solvers can be influenced when shown examples of potential ideas, the study makes a broad appeal to organizations who use crowdsourcing platforms. It makes sense, the authors say, to consider that looking at examples can potentially sap a solver’s commitment to generating new, original ideas. In the same vein, the researchers argue in favor of contests that are more sympathetic to contestants’ needs, whether by allowing more time for scanning and evaluating ideas or by better matching them with contests based on their ability, experience, and availability. “For example,” the authors note, “new contests with urgent deadlines may not be ideal for solvers who are heavily involved in existing contests.”