Companies struggle with decisions about the composition of teams of knowledge workers and how to deploy those teams productively. The experience of the semiconductor industry, which has trouble getting most new products to market on time, is instructive: our research on more than 2,000 R&D projects at more than 75 companies finds that leaders underestimate how the dynamics of teams affect the output of R&D.
Using a proprietary database that measures semiconductor-development efforts in a consistent fashion,1 we examined productivity across a number of company sites. Increasing the size of R&D teams, we found, actually diminishes productivity. So does expanding the number of development sites. In the auto and wireless markets, for instance, R&D output decreased significantly as the size of project teams rose. Output also falls when companies try to manage design teams across multiple work sites, a path R&D managers often choose when they can’t achieve critical mass at a single location. Expanding from one site to three can lead to up to a 20 percent drop in productivity (exhibit). R&D efforts—and perhaps, by extension, other knowledge-work clusters—seem to have natural limits. Adding people beyond those limits diminishes returns. Organizations that manage complexity in an effective way at a single site may lose their grasp when far-flung teams take on complex tasks.
For more, download “By the numbers: R&D productivity in the semiconductor industry” (PDF–318KB), from the Autumn 2014 issue of McKinsey on Semiconductors.