Remote work and non-obvious downsides

With the news that Facebook is going semipermanent work from home I've been spending time thinking about what happens to the softer sides of company performance.

Ten years or so ago, there was a big push to encourage random people to bump into each other; I remember that the architects who did Meraki's interior at 500TF saying that different snacks were put in kitchenettes to encourage people going to different sections of the company, and they built "choke points" to force people to mix.

(That made me wonder if we should just have created a completely random seating plan and watched what happened.)

I'm always a little suspicious of things like this because it's fundamentally not measurable and is inevitably oversold but a) it can't do any harm and b) we always need to be open to systems that have non-obvious outputs.

And it worked! There were a number of people I talked to who had wandered upstairs to make an espresso that I wouldn't have otherwise. Did it have a appreciable effect on our bottom line? That's impossible to judge, but it certainly made things better (for me; for my interlocutors maybe not)

So when companies start mass WFH policies, I wonder about the unseen stuff. There's the obvious (no commutes! Less rent!), the semi-obvious (do you need random interactions, how do you brainstorm, ensuring all voices get heard) and then the really non-obvious (what happens to culture in the medium-term, how do you measure morale, are there weird limits to scaling and how would that become known).

There are plenty of companies that have been successful doing the fully-remote thing; Gitlab seems to be the most cited. But they're relatively small. And there are many large companies that are at least semidistributed; Netflix is sometimes cited as one, although anyone who has driven down Winchester knows the reality lags the hype.

The closest I could conceive was Cisco's massive experiment with telecommuting, although most of the data I could find highlighted the cost savings (obvious) and the employee satisfaction (also obvious).

This isn't some paean to the joys of office life or contrarian take about the move to virtual environments. Instead the question we need to ask is Why wasn't there a mass move to virtual offices before? It was pointed out that up until March, the trend in Silicon Valley was for more consolidation and bigger campuses, like the Apple Park, Facebook's Menlo Park campus, the Googleplex, etc. In fact, a lot of these massive developments were sold on increased productivity, innovation, and retention because of the random interactions, amenities, etc.

So to believe that going to extreme virtualization will somehow be better than status quo ante one has to believe that a lot of very smart people who spent a lot of money over the last ten years promoting centralization were at best wrong and inefficient and at worst, actively harming the very people and work they were trying to help.

Fundamentally if there are downsides to this, they are going to be hard to articulate, hard to measure, and potentially impossible to correct. We need to be thinking now about potential "soft issues" before cultures around distributed teams become entrenched.

Obviously there will have to be much more distribution; things aren't going back to the way they were any time soon and it is prudent to be prepared in the future. But let's keep an eye out for non-obvious downsides to this move to telecommuting. That's the thing about non-obvious problems; you have to look for them. By the time they become obvious, it's too late.