Much has been written about the impact of remote work on collaboration and innovation in the workplace. Yet, few rigorous studies have attempted to document the actual impact on innovation. Now comes a fascinating new paper from David Atkin, M. Keith Chen & Anton Popov. They used an enormous amount of smartphone geolocation data to track face-to-face interactions. Their research sample includes over 51,000 employees in Silicon Valley in 2016 and 2017 (pre-pandemic). Here's an excerpt from their paper:
Our rich data on interactions allow us to open the black box of knowledge spillovers and isolate a particular channel: face-to-face meetings. To do so, we first link worker interactions—measured by the probability that a worker from one establishment “meets” a worker of another establishment by being in the same place at the same time with patent citations between their employers, an observable proxy for knowledge flows.
To calculate these meeting probabilities, we combine smartphone geolocation data with maps of building rooftops for all patenting firms in Silicon Valley, assigning workers to establishments based on where they spend a large fraction of their waking hours. To assign firm-level citations to establishments, we scrape citation data from recent patent applications and use the inventors’ hometowns coupled with the housing locations of workers to probabilistically assign citations across multi-establishment firms. The resulting dataset of establishment-to-establishment worker meetings and citations reveals a strong positive relationship between face-to-face interactions and knowledge flows, even after conditioning on rich controls for the physical distance between establishments.
Note the final sentence. Face-to-face interactions enhanced knowledge flows. They documented a strong positive effect of face-to-face interactions on patent citations. Here's more from the authors:
Implementing this approach, we find that face-to-face meetings significantly increase citations between establishments, with the strength of the effect twice the impact of physical distance on citations. Eliminating a quarter of face-to-face meetings in Silicon Valley would reduce the number of citations by approximately 8 percent..."
Naturally, some will argue that we have learned how to collaborate remotely during the pandemic, and have we have mastered a batch of technologies that promote virtual collaboration. Certainly, we have become much more effective at communicating and collaborating with others spread across remote geographic locations. Yet, remember that this study documents many informal, serendipitous face-to-face interactions among employees. Those interactions are much harder to replicate virtually. While many leaders are concerned about losing employees if they try to mandate a return to the office, they do have to consider this study's implications regarding remote work and innovation. Hopefully, more researech will follow, enabling us to gain a deeper understanding of the value of face-to-face interaction in the new product and new process development process.