Brainstorming Beyond the Laboratory: Idea Generation Practices in Software Development Firms

Abstract

Idea generation has been a topic of creativity research for over half of a century. First proposed by Osborn, brainstorming has grown to be the most popular idea generation technique. Brainstorming is an idea generation technique that focuses on using quantity to breed quality. Laboratory studies show that brainstorming productivity can be greatly improved by reducing hindrances such as evaluation apprehension, production blocking, and social loafing. Based on these findings, researchers have attempted to build electronic tools that allow brainstorming participants to propose ideas to the group anonymously in parallel to assist the idea generation process. Although these electronic systems show great promises in laboratory conditions, they have failed to reach critical success in the field. Creativity research shows that there is a lack of understanding of how creativity is evaluated in the domain of software development. We conducted an observational study on Microsoft's idea generation practices. Based on the field study results, we designed a survey and deployed it at Microsoft and Google. Our findings suggest brainstorming is used very differently when compared to Osborn's original vision. Although software teams at Microsoft and Google have different idea cultures, they use idea generation to iteratively discover constraints that could be used to shape the solution boundary and verify feasibility of a proposed solution. Instead of trying to generate ideas in quantity, software teams use brainstorming meetings to encompass different possibilities and alternatives whereas impromptu idea discussions are used to explore an idea in greater depth. We propose an idea generation model based on the empirical results. We discuss design implications and propose future studies that could lead to better understandings and improvements of the idea generation process.