Study of 61,000 Microsoft employees shows the shift to remote work has hurt communication and collaboration, threatening productivity and long-term innovation

GeekWire 2021-09-13 Todd Bishop

Context & Ripple Effects

A year after Microsoft formalized a policy letting managers approve permanent remote work, the company's own researchers ran the largest inside look at what that shift did to its organization: a study of 61,000 employees finding that collaboration networks tightened around close teams while cross-group communication thinned — the weak ties that carry new information between groups are exactly what eroded.

That finding set up the arc that followed: Microsoft reversed course with a full February 2022 reopening of its Washington facilities, and years later its FY2026 filings showed product R&D roles declining for a second straight year even as total headcount fell for the first time since 2016 — the long-term innovation risk the study flagged now visible in the company's own numbers.

First-order effects

Microsoft's own hybrid-policy architects now have internal evidence that the flexible arrangement they approved in October 2020 was degrading cross-team communication at scale, putting hard data behind the push to reopen offices.

Second-order effects

The study gives every large employer weighing return-to-office mandates a citable dataset from the most-wired workforce in tech — shifting the debate from preference surveys to measured collaboration loss.

Third-order effects

If siloed collaboration persists across the industry, firms that treat office density as an innovation input will pull ahead of fully distributed rivals — a structural split the later R&D headcount declines at Microsoft hint may already be registering.

The trend: Companies are increasingly arbitrating remote-versus-office policy with internal network analytics rather than employee preference alone.

Discussion

@emollick Ethan Mollick on x
A giant study at Microsoft adds to the growing picture of the sudden shift to remote work: the problem is not with existing work & teams; it is the future of organizations. People stop creating new bonds & become more siloed. Bad for innovation & careers. https://www.nature.com/.…
@jonerp Jon Reed on x
Interesting data but I strongly disagree with framing this study as a “remote work experiment” at scale. This is an extreme pandemic work adaptation, big difference. It is not a true test of the art of the possible with more flexible workstyles. (1/x) https://twitter.com/...
@jonerp Jon Reed on x
Example: if we weren't bound by pandemic concerns, the remote work downsides cited here could be mitigated by local meetups, regional sessions, hybrid office days and 1:1 mentor sessions. We haven't begun to test what we've learned in non-pandemic circumstances (3/x) https://twit…
@rwhitcomb Richard Whitcomb on x
Joined my current company as fully remote 3 years ago so almost half remote and half COVID remote. Not being able to fly in and spend a week meeting f2f with folks every few months does make me feel much more siloed. What we are doing isn't really normal remote. https://twitter.c…
@profkunze Florian Kunze on x
Impressive study of 61,182 Microsoft employees @NatureHumBehav shows that remote work leads to less social integration, less synchronous communication, and more silos in organizations. Clear evidence for a hybrid instead of a remote-only workplace https://www.nature.com/... https…
@rakyll @rakyll on x
I'd love to see a research on how pandemic impacted innovation and new initiates in tech companies before reading too much into this. During pandemic, keeping the lights on has been a greater focus. https://twitter.com/...
@zackkanter Zack Kanter on x
Good. Cross-team “collaboration” implies cross-team dependencies; silos are a sign of decreased coupling. Distributed work leads to cleaner interfaces. https://twitter.com/...
@x0rg Raffaele Di Fazio on x
Look I'm not a scientist, but doing a study on the impact of remote work in the time we are all forced to go through a collective trauma doesn't sound like the best way to get unbiased results. https://twitter.com/...
@jonerp Jon Reed on x
The reason this gets at my craw is that without the considerations I noted, companies could use this data to justify a knee-jerk return to centralized office work. You can make this cultural mashup work, and be more more inclusive, with flexible work. (4/4) https://twitter.com/..…
@josh_bersin Josh Bersin on x
New Microsoft Research found a lot of psychological challenges with Hybrid Work, https://www.geekwire.com/... and new @Microsoft Teams features to respond. Pretty cool stuff. https://www.youtube.com/...
@copyconstruct Cindy Sridharan on x
When I asked the same question pre-pandemic, a lot of people assured me it's nothing to worry about, it's no big deal, and everyone who does remote “right” won't have this problem. Now large scale studies prove this very point. https://twitter.com/... https://twitter.com/...
@lazerwalker Emilia on x
A study about remote work during a mass trauma event, when no in-person travel is possible, is not a reasonable study of “remote work”. My remote-prior-to-2020 MS team saw those same trends. Because we were all burnt out from working during a fucking pandemic. https://twitter.com…
@dcherring Dennis Herring on x
and the Death of Cities was declared *premature*... https://twitter.com/...
@wolfejosh Josh Wolfe on x
Siloes bad. https://twitter.com/...
@dkthomp Derek Thompson on x
Study of 61,000 Microsoft employees in 2020 finds remote work caused “collaboration to be more static and siloed...a decrease in synchronous communication ... and [made it harder for] employees to acquire and share new information across the network.” https://www.nature.com/...
@mahemoff @mahemoff on x
Remote work doesn't have to make people more siloed. You can have the best of both worlds, but need to reinvent process for the medium. Lift-and-shift was excusable in a sudden pandemic, but long term any org that doesn't adapt will crash hard. https://twitter.com/...
@eaton Actually on x
Fascinating paper breaking down changes in work cadence and technology use over the course of the pandemic. https://www.nature.com/...
@sandeepssrin Sandeep Srinivasa on x
“emails, calendars, IMs, calls of 61,182 US Microsoft employees over first 6 months of 2020 to estimate causal effects of firm-wide remote work. Our results show remote work caused the collaboration network of workers to become more static and siloed” https://www.nature.com/...
@amychanta Amy Chantasirivisal on x
I hate clickbait headlines like this. In the actual paper, they list out a slew of possible confounding factors. This is a study of ONE company that doesn't even consider the company culture and practices that put productivity and innovation at risk 🙄 https://www.geekwire.com/...
@jangles Neville Hobson on x
The research findings show the importance of #flexibility and #communication between managers and employees in #remotework and #hybridwork settings https://twitter.com/...
@mayhem4markets @mayhem4markets on x
Shots fired against WFH: “Study finds that $MSFT's shift to remote work has hurt communication and collaboration among business groups, threatening employee productivity and innovation.” 👀 My friend @benitoz mentioned this as a concern a while ago. https://www.geekwire.com/...
@akianastasiou Aki Anastasiou on x
Study of Microsoft employees shows how remote work puts productivity and innovation at risk https://www.geekwire.com/... remote work has hurt communication and collaboration among different business groups inside the company, threatening employee productivity and long-term innova…
@carnage4life Dare Obasanjo on x
Remote work creates silos. Talking to people outside your core group now has to be intentional versus catching them in the cafeteria or in the hallway after a meeting. I had a recent Zoom call with a peer where we realized we would have talked months earlier if in the office. htt…
@xamat Xavier Amatriain on x
Great work by Teevan et al. but too many interpretations like this one take the easy “remote work is bad”. How about “we spent years optimizing organizations for non-remote, and we could not react quickly enough when forced to go remote”? https://twitter.com/...
@markghuneim Mark Ghuneim on x
First study of size I have come across, contradicts all the annadotal working so much better from home stories. https://twitter.com/...
@calebwatney Caleb Watney on x
“Our results show that firm-wide remote work caused the collaboration network of workers to become more static and siloed, with fewer bridges between disparate parts.” https://www.nature.com/...
@krzyzanowskim Marcin Krzyzanowski on x
I beg your pardon? If anything, asynchronous communication makes new information sharing easier, not harder. Also, why would anyone look at pandemic WFH and make it equal to remote work? It's not. It was a different experience. https://www.nature.com/... https://twitter.com/...
@jespow Jesse Powell on x
As someone who has led remote-first companies for 20 years, I can say that it works and 2020 was different. I'm sure it was hard for unprepared companies and people to abruptly make the switch to remote, especially with kids stuck at home. Ignore noob data, study the veterans. ht…
@lawrencecoburn @lawrencecoburn on x
This is worth a read end to end: the first hard data I've seen on the hidden costs of remote work. Fragmentation, loss of trust, ossification of networks, leading to worsening communications and collaboration. This is the most exciting problem in enterprise tech today. https://tw…
@aleresnik Ale Resnik on x
“Our results suggest that shifting to firm-wide remote work caused the collaboration network to become more heavily siloed—with fewer ties that cut across formal business units or bridge structural holes in Microsoft's informal collaboration network—” https://twitter.com/...
@bryanrbeal Bryan Beal on x
What we've been doing over the past 18 months is not remote work. It is lockdown work. Normal Remote work includes regular get together's with your team. No pandemic lockdown study is an accurate reflection of remote work https://twitter.com/...
@kmcneel Kate on x
The trick is to build in time for relationship building and networking and chatting. I've worked remotely for 20+ years. I still talk regularly with people I have worked directly with for more than a decade. Almost always results in new ideas or connections or info. https://twitt…
@provisionalidea James Rosen-Birch on x
Nothing about this study is surprising. Orgs designed for in-person work are structured fundamentally differently from orgs designed for remote work. Every org designer has been shouting this since the pandemic started: you can't just put your org on Zoom and call it remote. http…
@_svs_ Svs on x
Orgs need to build interfaces between teams that do not depend on ‘ties’ or ‘time’. The work takes center stage. ‘Fun’ at office is about enjoying the work, not spending time with cute people or distractions. https://twitter.com/...
@seyitaylor St. on x
it's obvious that organisations as currently structured were designed for the office. to function well, remote orgs can't be “office orgs, but on Zoom”. will take a few iterations and generations to nail. https://twitter.com/...
@work_life_you Nancy Hey on x
Nice analysis of bridging and bonding capital changes at work with shift to remote work. Likely echoes non work relationships. Need to see if this holds after pandemic restrictions as reducing loose bridging interactions *was the whole point* of the restrictions. https://twitter.…
@somesheep @somesheep on x
I would argue that Microsoft is an exceptionally ill equipped sample to make any assumptions about how remote work in general will impact such things. Likely those are dysfunctions of Microsoft's own culture being amplified. https://twitter.com/...

Chronicles