The Office Became a Protocol

An office is a communication system that nobody has to maintain. It broadcasts continuously and for free: who is here, who is busy, who is talking to whom, what mood the project is in, whether the thing you were worried about is still a problem. None of that is written down. It is absorbed through proximity, and it is enormously valuable, which is why removing it hurt so much and why the companies that recovered fastest were the ones that stopped trying to recreate it.

The lesson of the distributed years is not that remote work is better or worse. It is that distance converts implicit coordination into explicit coordination, and explicit coordination is a thing you have to actually build.

What the building was doing

Four jobs, mostly, all of them invisible until they stopped.

• Status broadcast. You could see that the design was on the wall, that the tester looked stressed, that two people had been in a room for an hour. Remote teams replace this with written status — or they replace it with anxiety.
• Low-cost interruption. The thirty-second question asked over a desk divider, which is trivially cheap in person and becomes a scheduled meeting at distance. Teams that did not find a substitute watched small uncertainties compound into large ones.
• Boundary enforcement. The building starts and ends the day. Without it, work leaks in both directions, and the leak is asymmetric: it flows into evenings far more readily than it flows out of mornings.
• Belonging. The birthday, the shared lunch, the walk to the station. This is the one people miss most and the one that video calls replace least well, because it depended on being unscheduled.

Distance does not create coordination problems. It removes the free subsidy that was hiding them.

Writing as infrastructure

The single strongest predictor of whether a distributed team functioned well was whether it wrote things down. Not more messages — more documents. There is a large difference between a chat channel, which is a river, and a written decision record, which is a place.

Teams that made this shift produced a specific set of artefacts: a decision log with dates and reasoning, project pages that stated current status without anyone having to ask, onboarding documentation written by the last person to be onboarded, and — crucially — a norm that if it was not written, it did not happen. That last norm is the hard one, because it removes the informal power of people who were previously well-connected rather than well-organised.

The payoff is that written coordination scales in ways proximity cannot. It works across time zones. It works for the person who was on leave. It works for the person hired next year. An office communicates brilliantly to the twelve people in it and not at all to anyone else.

Synchronous time became expensive

When meetings were free — you were all in the building anyway — nobody costed them. At distance they became the scarcest resource, and good teams started spending them deliberately: for decisions that need argument, for relationship-building, for the delicate conversations that text handles badly, and for very little else.

The corresponding discipline is that everything else moves to asynchronous by default. This is genuinely harder than it sounds, because asynchronous communication requires the sender to do more work — to anticipate questions, to provide context, to write clearly enough that no follow-up is needed. Organisations get the coordination quality they are willing to pay for in writing effort.

What it replaced

The commute, obviously — the largest single block of time returned to a working life in decades, though roughly half of it flowed straight back into work rather than into rest.

More consequentially, it replaced geography as a hiring constraint. The most durable change of this period is that a company in an expensive city can hire someone who does not want to live there, and a person in a small town can take a job that would previously have required moving. That is a real redistribution of opportunity — and it also means local employers now compete with distant salaries, which is not universally good news for the small town.

It also replaced presence with output as the default proxy for contribution. In principle this is fairer. In practice it depends entirely on whether output is measurable in a given role, and for a lot of relational and supportive work, it is not — which is why the people whose contribution was hardest to see on a dashboard were often the ones who lost most.

What it cost

The junior experience. Learning by proximity is real. New employees who joined remotely reported weaker networks, slower ramp-up and less sense of belonging, and this shows up in retention figures a year later. The fix — deliberate mentorship, structured onboarding, scheduled unstructured time — is expensive and easy to skip.

Proximity bias. The worst configuration is not fully remote. It is hybrid done carelessly, where the people in the room make the decisions and the people on the screen learn about them afterwards. Any hybrid arrangement that does not aggressively defend remote participants creates a two-tier organisation within about six months.

Surveillance creep. Managers who lost the ability to see people working reached, in some cases, for tools that measure activity — keystrokes, idle time, screenshots. These reliably measure the appearance of work and reliably destroy trust, which is the actual substance holding a distributed team together.

Boundaryless days. Without the building to end the day, the day ends when you decide it does, and a lot of people never decided. The most effective countermeasure has been embarrassingly low-tech: a firm, visible, leader-modelled stop time.