"When teams are diverse, meaningful innovation is more likely to happen."
"We know intuitively that innovation goals are well served by cross-functional “SWAT” teams that are diverse in their membership. As Andy Zynga argued in an earlier post, diversity is a means to overcome the cognitive biases that prevent people from seeing new approaches or engaging them when found. But while this seems only logical, is there empirical evidence to support it? When such diversity is enforced can we expect it to produce results? How do we know “more is better”?
Stanford professor Lee Fleming and his colleagues have been working on these questions by looking for patterns in the teams behind patents. They find that higher-valued industrial innovation (by its nature also riskier) is more likely to arise when diverse teams are assembled of people with deep subject matter expertise in their areas. Other interesting findings in Fleming’s body of work include the observation of a bimodal distribution of outcomes for diverse teams (that is, a relatively high rate of failure and high rate of big successes, with not much middle ground); and the discovery that different kinds of communications networks foster different levels of diffusion of innovation. Fleming focuses on cross-pollination in the context of “big D” Development, which often involves recombination of existing knowledge to serve commercial goals.
Along similar lines, Ben Jones and colleagues at the Kellogg Business School of Northwestern University published a paper in Science last year focusing on diversity in the production of new knowledge, as reflected in the research literature. Looking for patterns across some 17.9 million papers indexed in Thomson Reuter’s Web of Science, they demonstrated that the most influential papers (most highly cited) were those that exhibited an intrusion of interdisciplinary information. They also found that groups were more likely to foster these intrusions than solo researchers. This is entirely consistent with Fleming’s findings for industry, and his attempts to dispel some of themythology around lone inventors. (One difference in the studies is that, thus far, Jones hasn’t observed the bimodal distribution that Fleming does; there is apparently no cluster of papers with abnormally low citations which also feature intrusions of outside knowledge.)
Taken together, the studies led by Fleming and Jones make a good case for assembling that SWAT team that can bring multiple disciplinary perspectives to bear on a problem. It isn’t always obvious how to do so, but we at NineSigma can point to an instructive example at AkzoNobel. AkzoNobel is a multi-national, multi-divisional manufacturer and distributor of coatings systems, or more simply put, paint. But paint is really not as simple as just paint; for example, coatings for automotive applications are very different from decorative finishes. Among AkzoNobel’s divisions are more and less conventional manufacturers of chemicals and polymers. Having grown by acquisition, the company has the typical silos, with organizational and geographic boundaries inhibiting the diffusion of knowledge…."