Whose and which problems get solved? Success factors in selective broadcast search: problem, people, or pervasiveness?

Problem solving is complex and often requires many attempts. Often, the problem owner cannot fix the problem while some others already have a solution or know how to come up with a solution. Lately, using the crowd as problem... [ view full abstract ]

Problem solving is complex and often requires many attempts. Often, the problem owner cannot fix the problem while some others already have a solution or know how to come up with a solution. Lately, using the crowd as problem solver has become popular. Problems are exposed to a large enough crowd. Some potential solvers may already have a solution to the problem or are capable to generate a solution. While crowd-based problem solving seems to be promising little is known about it, especially when it comes to the question which problems get solved and by whom. This study analyzes more than 4 years of data of a large multinational company’s internal problem solving community. During this time, problem owners requested answers to 1,039 problems (they sent so-called urgent requests). Surprisingly, more complex problems have a higher probability to be solved. Problems which are portrayed in simple way to target a broad audience are more likely to be solved, too. This indicates that solutions may stem from actors who are not necessarily the intended audience – and who draw on alternative knowledge and heuristics to solve the problems. Solvers are more likely to solve problems of problem owners who helped others in the past, not of those in high formal positions in the company. These re-sults indicate that community members do not display opportunistic behavior by pleasing the bosses or solving problems that are very simple. Further, this paper investigates a new concept called “selective broadcast search”. Traditional “broadcast search” sends out a problem to a high number of solvers and tries to encourage a high number of responses. In a company’s internal context, in which all solvers try to solve the problems during their working hours, this concept is likely to be extremely costly. Thus, this paper analyzes an alternative approach called “selective broadcast search” in a natural field experiment. In “selective broadcast search” key-words that describe the problem are matched with keywords that describe the interests of the solvers. Thus, the number of requests each potential solver receives can be reduced drastically. The change from “broadcast search” to “selective broadcast search” affects the number of re-sponses and the solving probability in the short run, but not in the long run. We argue that the underlying mechanism is that sending out too many messages leads to an information overload of a potential problem solver. The reduced number of messages each solver receives leads to an increased willingness to read and respond the mail. This mechanism offsets the reduced number of recipients of each message.