Hard Work Isn’t the Point of the Office

The pandemic disrupted soft work—the gossip, eavesdropping, and casual relationship-building that aren’t a formal part of your job.

Illustration of a watercooler on a magic carpet
Matt Chase

About the author: Derek Thompson is a staff writer at The Atlantic and the author of the Work in Progress newsletter. He is also the author of Hit Makers and the host of the podcast Plain English.

As the Age of Delta scrambles back-to-office timelines, I find myself wondering what offices are good for in the first place.

I am pro-office. I miss a good eavesdropping, the promise of midday gossip, the “quick random question” that blooms into a half-hour conversation, and, theoretically, the magical combustion of creativity forged by these connections.

These things aren’t what I’m directly paid to do when I’m in the office, and they’re not what I’m annually evaluated for doing. Instead, they’re what I think of as “soft work.” “Hard work,” for me, is reading, researching, calling people, transcribing conversations, and writing articles. For others, it might include managing employees, working in Excel or PowerPoint, or reading and writing a zillion emails. (This kind of hard work, I should note, doesn’t have to be physically difficult.) If the past year and a half has taught us anything, it’s that white-collar workers can do hard work from home just about as well as they can do it in the office—and maybe even better, precisely because their colleagues aren’t interrupting them.

Soft work is different. It’s the vague middle space of weekday activity that isn’t hard work but also isn’t not-work. (You know what not-work is: calling your dentist at noon, getting lost on YouTube, falling down a Wikipedia rabbit hole about the mating habits of seahorses. It’s all fine; just don’t call it work.) Soft work is getting coffee with a co-worker. It’s catching up about the NFL on Monday morning. If networking, schmoozing, gossiping, and mildly annoying people on your floor with “Hey, does this idea suck?” are species of behavior, soft work is the genus that contains them.

Offices’ biggest advantage isn’t hard work. It’s soft work. That’s not just my theory. It was the unmistakable conclusion of a new paper from Dave Holtz of UC Berkeley and Siddharth Suri, Longqi Yang, and Sonia Jaffe of Microsoft. They studied the effect of the first six months of the pandemic on 61,000 Microsoft employees. Remote work caused “collaboration to be more static and siloed,” they concluded, noting a sharp decrease in communication across groups that made it harder for “employees to acquire and share new information.”

This is a complicated paper, but the best way to understand their findings is to think about two types of connections at the office. There are in-group connections—the ones we have within a formal team under a single manager. And there are out-group connections—the informal friendships we have throughout the company. These out-group relationships come in many forms. Maybe, before the pandemic, you occasionally got lunch with someone to swap stories about raising young daughters. Maybe, on the way to the bathroom, you often stopped by somebody’s desk to talk about basketball. Or maybe you bumped into the person who sat next to you at your new-employee orientation and liked to talk crap about the company from time to time.

Remote work took a battering ram to all out-group ties. The researchers classified and studied several of them—ties across teams, ties between loosely affiliated workers, ties between people who barely interacted at all—and found that the numbers all plummeted.

Only one kind of communication went up in the pandemic. It was the individual clustering coefficient, which is a lot of syllables that mean something like “team closeness.”

“Think of your 10 closest friends, and if those friends know each other or not,” Holtz told me. “If they’re all strangers to each other, that’s low clustering coefficient. But if your best 10 friends are all friends with the other nine, that’s a high clustering coefficient. What we found in our study is that communication outside of groups declined, but communications within teams became more densely connected.” Within Microsoft during the pandemic last year, the silos got deeper, and the walls between silos got higher.

So the trillion-dollar question is: Who cares?

For a decade or more, productivity experts have been telling us that cross-group collaboration and weak ties are the skeleton key to unlocking radical creativity. In the abstract, this seems plausible. The history of invention is, in large part, a history of smart people making great leaps forward by combining disparate ideas in domains adjacent to their expertise. As the British philosopher Liam Bright has put it, “anything can be related to anything,” so we ought to build systems that allow people to see previously undiscovered connections between siloed domains.

This sounds like a loud endorsement of the office. But the internet does all of this at a much larger scale than any office floor. (After all, I discovered Bright’s observation on Twitter.) What’s more, if today’s corporate conventional wisdom were true, the pandemic would have created a hellscape for productivity. Instead, corporate earnings are rising, wages are rising, and official measures of productivity are rising too, in practically every state. Microsoft’s cross-group communication might still be in the dumps, but its performance certainly isn’t; in June, the firm announced that its quarterly net income had risen 47 percent annually.

Moreover, the closer you look at the collaboration research, the more caveats you find. Yes, collaboration in the office has seemed fundamental to solving big problems. This is in part because although casual conversations aren’t immediately productive, they build the trust necessary for groups to freely exchange ideas and feedback without making everybody hate one another. But a great deal of office-based collaboration turns out to have been pure wasted time. Other research suggests that we’ve been consumed by “collaboration overload” for years and that we might be better off clawing back up to 20 percent of collaboration time for ourselves to avoid burning out from an eternal purgatory of circling, circling, circling back.

The debate about offices isn’t always really a debate about offices. In many cases, it’s a tug-of-war over tradition, wherein the pro-office crowd essentially stumps for the status quo and disguises its change aversion by wrapping it in the normative language of Good Offices versus Troublesome Work-From-Home Policies. To cut through that, try a little thought experiment.

Imagine an alternate reality in which no such thing as an office existed in the 1900s and early 2000s. Until 2019, everybody worked remotely—in their homes, coffee shops, and libraries. Then, in 2020, a strange virus broke out. It was transmitted by aloneness. Suddenly workers who had been happily (and sometimes not-so-happily) remote for decades were forced to cluster in large, open indoor spaces with adjacent desks called “cubicles.” Physical rooms were built and reserved for what used to be Zoom calls but were then rebranded as “meetings.” Gossip on Slack and IM was abruptly funneled into hallways. We called our newly synchronous chitchat “watercooler gossip,” although this was an odd coinage as watercoolers were rarely visible on the premises.

In short, for the first time in a century, white-collar Americans were forced to work in one another’s physical proximity, and they discovered, to their horror and delight, that the proportion of soft work in their day had dramatically expanded. Some people, especially extroverts and some managers, welcomed this change. Spending their weekdays just, like, talking to corporeal beings all the time was such a pleasure! Others hated the new arrangement so much they promptly quit …

The point of this hypothetical is twofold. First, inverting reality sometimes helps show that even the most familiar institutions are also inventions. Second, in a very real way, it’s not a hypothetical. Remote is the status quo for many American workers now. People are quitting their jobs rather than be forced back into offices. It is the corporate headquarters, not the bedroom office, that is, or is about to be, the novel intrusion for much of today’s white-collar workforce.

The lesson of the Microsoft study is that offices seem to be excellent at a specific job: They increase soft work across silos. But rather than think of toiling in offices as the good, normal thing to which we’ll inevitably return, we should think of it as one of many working arrangements that deserve our scrutiny and skepticism.