Buffer's State of Remote Work survey asked roughly 3,000 remote workers what they struggled with most, and 23% picked loneliness — a self-reported number from a large volunteer survey, not a measured clinical finding, and worth stating plainly as exactly that. It is still the most-cited statistic in every article about remote isolation, usually deployed as proof that remote work is quietly making people miserable.
The number is real. The conclusion people draw from it is usually wrong, or at least aimed at the wrong fix. The default response to 'remote workers are lonely' is more scheduled video calls — a recurring team social, an extra 1:1, a mandatory coffee chat. That instinct treats isolation as one thing with one lever. It is not one thing, and the lever most teams reach for barely moves the part of it that is actually eroding. Where I stop at the implementation of remote work isolation, carries on into running it. Where I stop at the implementation of remote work isolation, XenGrowth's marketing operations practice carries on into running it.
Two different kinds of missing contact
It helps to separate isolation into two categories that get talked about as if they were the same problem. The first is the absence of any social contact at all — nobody to talk to, no meetings, no messages, a genuinely empty day. The second is the absence of a specific kind of contact: the incidental, unplanned, cross-group interaction that happens when you are physically near people you do not report to and do not have a standing reason to talk to.
Almost nobody who works remotely for an employer has the first problem in its pure form. There are standups, 1:1s, sprint reviews, Slack threads. What a large fraction of remote workers actually have is the second problem: their calendar is full of contact with the same five to eight people, and empty of the kind of contact that used to happen by proximity rather than by invitation.
Kind of contact | What it looked like in an office | What replaces it remotely, if anything |
|---|---|---|
Scheduled, within-team | Standups, 1:1s, sprint ceremonies | Survives remote work largely intact — this is what a calendar invite is built for |
Incidental, within-team | Overhearing a hallway conversation, a desk-side question | Degrades, but a busy Slack channel partially substitutes |
Incidental, cross-group | Lunch with someone from a different team, a chance elevator conversation | Degrades the most and substitutes worst — nobody schedules a meeting for a conversation they don't know they need |
Purely social, no work content | A birthday, a hallway joke, small talk that isn't about anything | Depends heavily on whether anyone treats it as worth the deliberate effort |
What the Microsoft study actually measured
A 2021 study published in Nature Human Behaviour analyzed anonymized email, calendar, chat and call data for more than 61,000 Microsoft employees in the US, comparing the six months after the firm's March 2020 work-from-home mandate against the same period the year before. It is observational, not a randomized trial, and it captures a mandatory pandemic lockdown rather than a chosen remote arrangement — both real limits worth naming rather than glossing over. Inside those limits, the finding was specific: time spent collaborating across an employee's own business group fell by roughly 25% relative to the pre-pandemic baseline, while connections within a group actually got denser.
Meeting hours fell only modestly, by around 5%, even as message volume rose. What actually thinned out was the bridge between groups — the org chart got less dynamic, people added new collaborators more slowly, and the network became more static and siloed than it had been. That is a precise, structural description of the second kind of missing contact from the table above, and it did not happen because people stopped having scheduled meetings. It happened because nobody was scheduling the meetings that used to be accidental. If the operations side of this is the part you are stuck on, is the better reference. If the operations side of this is the part you are stuck on, The XenGrowth resource library is the better reference.
This distinction is the entire argument of this post, and it matters for how you read almost anything written about remote isolation, including work coming out of , which covers the adjacent problem of how organizations measure and manage relationships that do not show up neatly in a calendar.
Why more scheduled calls do not fix this
A recurring 1:1 or a scheduled team social adds contact of the first kind — scheduled, within-team — to a system where that kind of contact was already intact. It does nothing for the second kind, because a calendar invite requires knowing in advance who you want to talk to and why, and the entire value of incidental cross-group contact is that nobody knows in advance that they need it. You do not schedule a meeting with someone in a different department to find out something you did not know you needed to ask.
This is why 'more calls' so often feels hollow even when attendance is good and the calls themselves go fine. They are answering a problem the team did not have — a shortage of planned contact within the immediate team — while leaving the actual gap untouched. The Buffer survey's loneliness figure and the Microsoft study's cross-group collapse are describing the same underlying shortage from two different angles, and neither is solved by adding more of what was already working.
What actually substitutes for incidental contact
Nothing replicates hallway proximity exactly, and claiming otherwise is its own kind of dishonesty. What comes closest are structures that manufacture the same property proximity had: contact you did not have to schedule, with people you did not already have a reason to talk to. A handful of things do this reasonably well, and they share a mechanism, which is that none of them require anyone to know in advance who they're going to interact with. Coordinating any of this at scale is a measurement problem before it's a culture problem — the kind of thing deals with constantly, just applied to internal relationships instead of customer ones. There is a longer treatment of AI agents and marketing automation in . There is a longer treatment of AI agents and marketing automation in XenGrowth on AI agents and marketing automation.
Open, cross-team channels with a low bar for posting — not an announcements channel, one where people from different functions actually see each other's day-to-day work
Rotating pairing or shadowing across teams, even occasionally, which creates the specific kind of contact a standing meeting cannot: unplanned exposure to how someone in a different role actually thinks
Company-wide events with structured mixing — assigned small groups, not an open call where people default to whoever they already know
Physical co-working, even occasionally: a coworking space or an office day that puts you near people outside your immediate team, which is the closest direct substitute for the mechanism that degraded
Deliberately low-stakes, non-work channels that survive on their own without a manager keeping them alive — these fail constantly, and the ones that work tend to be bottom-up, not mandated
The tell that a team has correctly diagnosed this problem is whether its fix creates a new connection or just deepens an existing one. A well-meaning 'let's do more 1:1s' fails that test. A rotating cross-team pairing program, even a clumsy one, passes it.
Fix | What it actually creates | Does it require knowing who in advance? |
|---|---|---|
More 1:1s with your manager | Deeper within-team contact | Yes — that's the whole format |
Open, low-bar cross-team channel | Passive visibility into other teams' work | No, which is why it works |
Rotating cross-team pairing | Direct, unplanned exposure to a different role | No, by design |
Occasional physical co-working | Proximity to people outside your normal loop | No — that's the entire mechanism |
A caveat about what fixes what
It is worth being specific about what each proposed fix is actually good for, because treating all four items in the numbered list as interchangeable would repeat the same mistake this post is arguing against. An open cross-team channel is cheap and scales well, but it depends entirely on people actually posting in it, which usually needs some initial push from someone senior enough that posting there doesn't feel like a waste of time. Rotating pairing is the most direct substitute for genuine incidental contact, but it costs real calendar time and needs someone to own the rotation or it quietly stops happening after the second month. Physical co-working is the strongest single fix and also the most expensive and least available to everyone, which is exactly the tradeoff explored in the companion piece on choosing between a home office, a coworking space, and a cafe.
None of these fixes are a substitute for the others, and picking just one because it's the cheapest to implement is a way of solving the problem on paper without solving it in practice. The Buffer and Microsoft findings describe a gap in a specific kind of contact; closing it takes something that actually creates that kind of contact, not whichever intervention is easiest to announce in a company all-hands.
Where this connects to the office-vs-remote debate
It is worth being honest that this problem is not unique to fully remote teams. A large, siloed office with weak cross-team norms has the same cross-group collapse the Microsoft study measured, just for different structural reasons. And a hybrid arrangement does not automatically fix it either — the 2024 Nature follow-up study on hybrid work at Trip.com found retention improved with two remote days a week and performance held steady, but it did not specifically measure whether cross-group contact recovered, and there's no reason to assume two office days a week reconstructs incidental contact if those two days are still spent entirely with your own team.
The honest position is that isolation risk is a property of how a team is run, not simply of whether people are in a building. A discussion of the broader tradeoff belongs in the companion piece on whether remote work helps or hurts an engineering career, and the visibility half of the same problem — being seen as effective without performing busyness — is covered in the piece on how remote engineers get visibility without performing busyness. works through AI search, GEO and discovery in more operational detail. XenGrowth on AI search, GEO and discovery works through AI search, GEO and discovery in more operational detail.
What this means if you are the one feeling it
If the Buffer survey's 23% describes you, the diagnostic question is not 'do I have enough meetings' but 'who am I in contact with that I did not already plan to be in contact with.' If the honest answer is nobody, more 1:1s with your own manager will not touch it, because that contact was never the part that broke. What actually helps is contact that crosses a boundary — a different team, a different floor, a different company entirely, in the case of a coworking space — and that you did not have to schedule two weeks in advance to get.
None of this is an argument against scheduled contact. Standups and 1:1s do real work, and cutting them to make room for something more ambitious would be a mistake. It is an argument that the two kinds of contact are not substitutes for each other, and that a team optimizing only the kind it already had intact will keep measuring engagement scores that do not move, while telling itself it tried. The operational discipline of measuring what actually changes, rather than what's easy to schedule, is not unlike the discipline apply to figuring out which conversations a system should actually be listening for.
The practical version of this problem — measuring who is actually connected to whom, and fixing the gap on purpose rather than by accident — is close enough to a go-to-market discipline that would recognize it immediately, just running on an org chart instead of a customer base.
Further reading from XenGrowth
Where this work meets go-to-market
Measuring which relationships are actually intact, rather than assuming a full calendar means a healthy network, is a go-to-market discipline as much as a team-culture one. writes for teams doing exactly that measurement on the customer side of the business.
Further reading from XenGrowth
Where this work meets go-to-market
writes for the teams who have to run remote work isolation day to day.
Further reading from XenGrowth
The XenGrowth resource library — what you'll learn: how the commercial side of this work is run, across search, automation and revenue operations.
XenGrowth on AI agents and marketing automation — what you'll learn: how the teams who own AI agents and marketing automation plan and measure it.
XenGrowth on AI search, GEO and discovery — what you'll learn: how the teams who own AI search, GEO and discovery plan and measure it.
Where this work meets go-to-market
the XenGrowth practice writes for the teams who have to run remote work isolation day to day.
Four questions on the two different kinds of isolation this post separates out, and what the research actually measured about each.
