Coevolutionary Lock-In in External Search

While external search allows organizations to source diverse ideas from people outside the organization, it often generates a narrow set of nondiverse ideas. We theorize that this stems from an interplay between organizations' idea selection and the external generation of ideas: an organization selects ideas shared by external contributors, and the external contributors, who strive to see their ideas selected, use the prior selection to infer what kind of ideas the organization is looking for, and to respond. Contributors whose ideas are misaligned with the organization's selection tend to stop submitting ideas (i.e., self-selection) or adjust the ideas they submit so that they correspond (i.e., self-adjustment), resulting in a less diverse pool of ideas. Our central hypothesis is that the more consistent organizations are in their selection, the stronger the coevolutionary lock-in: organizations with greater selection consistency receive future ideas with lower content variety. We find support for these predictions by combining large-scale network analysis and natural language processing across a large number of organizations that use crowdsourcing. Our findings suggest a reconceptualization of external search as a two-way street: organizations are not simply passive receivers of ideas but send signals that shape the pool of ideas that externals share.