The Collaboration Toll of Remote Work

Researchers analyzed communication patterns among at-home Microsoft employees during the pandemic and found that collaboration suffers when everyone works remotely.

MIT Initiative on the Digital Economy


Photo: Tanatat for Adobe Stock

It may seem that Zoom has elegantly replaced other forms of business communication, but new research shows a different scenario.

As companies debate the impact of large-scale remote work, a new study of more than 61,000 U.S. Microsoft employees found that working from home (WFH) causes workers to become more siloed in how they communicate, engage in fewer real-time conversations, and spend fewer hours in meetings.

The study, published September 9 in the journal Nature Human Behaviour and co-authored by MIT IDE researcher David Holtz, made use of data gathered before and after Microsoft imposed a company-wide, WFH mandate in response to the COVID-19 pandemic.

The findings suggest that a full-time remote workforce may have a harder time acquiring and sharing new information — which could have long-term implications for productivity and innovation among information workers.

“Microsoft’s WFH mandate created a unique opportunity to identify the effects of company-wide remote work on how information workers communicate and collaborate,” said Holtz. He conducted the research as a doctoral intern at Microsoft while a PhD student at MIT Sloan, and co-wrote the paper with Microsoft colleagues Longqi Yang, Sonia Jaffe, Siddharth Suri, and seven others. The analysis was based on anonymized data describing the emails, instant messages, calls, meetings, and working hours of the overwhelming majority of Microsoft’s U.S. employees.

MIT IDE Content Director, Paula Klein, asked Holtz — an assistant professor at Berkeley Haas, faculty affiliate at the Berkeley Institute for Data Science and research affiliate at the MIT Initiative on the Digital Economy — to explain a few of the research findings in the following Q&A.

IDE: Do the changes you observe reflect the effects of remote work in general, or are they specific to what we are experiencing during the COVID-19 pandemic? Why focus on the collaboration aspects of remote work?

David Holtz: Historically, it’s been difficult to measure the causal effects of remote work because only specific types of workers (e.g., certain roles, performance levels, or teams) were permitted to work outside of the office. This all changed with COVID-19, when essentially overnight anyone who could work from home was forced to do so by their employer.

David Holtz

At the same time, during the early stages of COVID-19 folks were dealing with a lot of other factors, such as fear of contracting the virus, school closures, home childcare, and an inability to go to places like gyms, restaurants, and bars. These factors would not be present if people were working from home during more normal times. In this study, we set out to separate the effects of remote work itself from the effects of other COVID-19-related factors so that we could understand the effects of remote work more generally.

For many of us, there is seldom a full day of work that doesn’t include a lot of formal and informal communication with colleagues in the form of quick conversations, meetings, emails, instant messages, etc. Previous research across a number of fields has documented how critical collaboration and communication are to creativity, innovation, and productivity. Because the relationship between collaboration and innovation/productivity is so strong — especially among information workers — we thought it was important to measure how remote work causes collaboration patterns to change. Since it is difficult to measure productivity and innovation for information workers, and there isn’t wide agreement on how to do so correctly, we focused on outcomes that our dataset enabled us to measure well.

IDE: Many recent reports focus on the benefits of remote work for workers and employers such as flextime, lack of commuting, and worker availability. You also present the disadvantages — such as more siloed interactions, more asynchronous communication, and static communication patterns. What are some of the biggest takeaways/red flags for executives weighing WFH policies?

DH: Based on research studies that have come before ours, we argue that without intervention the effects of remote work we report may make it harder for workers to acquire and share new information, which could also impact downstream business outcomes. In addition to the benefits you mentioned, we think this potential pitfall is something firms should take into account when making decisions about if and how to implement firm-wide remote work.

The model we used to analyze our data also allowed us to separate the effects of firm-wide remote work into two different components: how your own collaboration patterns are affected when you work remotely, and how your collaboration patterns are affected when your collaborators are working remotely, too. In many cases, your collaboration patterns are affected as much by whether your collaborators are working remotely as by whether you yourself are working remotely.

This finding suggests that firms need to be thoughtful when adopting policies like hybrid work or mixed-mode work. For example, our results suggest that having teammates and collaborators in the office at the same time improves communication and information flow for both those in the office and for those who are remote. That said, future research is needed to truly understand how to best implement these types of policies.

IDE: What were the biggest surprises you found?

DH: Some of the biggest surprises come from the fact that the effects of remote work were not always what you’d expect based on collective experience during COVID-19. For instance, we found that remote work causes the amount of time folks spend in meetings to go down. This is counter to what many of us have experienced in the past year or so, when it has often felt like the Zoom meetings never end!

Our results suggest that this huge increase in the amount of time spent in meetings had more to do with the pandemic than with remote work.

Our findings really highlight how dangerous it can be to infer causal relationships from descriptive data, especially when something like COVID-19 is happening in the background.

IDE: While your sample size was large, you note that the study was limited in its time frame — three to six months — and long-term effects may be different than what you found. As WFH models become more normalized do you think collaboration might bounce back? What are some areas of further study?

DH: I’m hesitant to speculate about long-term effects; in the future, workers and companies alike may intervene (through policy and/or through technology) and counteract many of the effects we describe in the paper. However, one thing we note is that during the time period we study, remote workers were able to leverage their pre-existing work networks and connections. We expect that as more employees get hired into new jobs where they’ve never met their colleagues in person, or as more firms transition to a fully remote workforce, it won’t be possible for workers to rely on preexisting rapport that was established in the physical world.

There are many interesting ways that future research could build on what we’ve done in this study. For instance, it would be fascinating to conduct research with data from other firms, including some that are in different fields or based outside of the U.S. While we are reasonably confident that our results can generalize beyond the context we studied, there are no guarantees.

It would also be interesting to explicitly link the collaboration changes we observed to longer-run measures of productivity and/or innovation, both at the individual- and firm-level, to see if the downstream impacts that one might expect actually come to bear.



MIT Initiative on the Digital Economy

Addressing one of the most critical issues of our time: the impact of digital technology on businesses, the economy, and society.