Four Days Only Works If You Stop Watching People

**Four Days Only Works If You Stop Watching People** I've never seen a four-day week fail because people got lazy. I've seen two fail because the managers couldn't tell what anyone was doing, panicked, and started filling the gap with meetings. That's the actual risk, and it shows up early. A company cuts Friday, keeps every single ritual it had before, and now Monday to Thursday is a wall of standups and syncs and "quick catch-ups" — all of which exist to answer one unspoken question: is everyone still working? Six months later someone runs the numbers, output looks flat or worse, and the whole thing gets quietly rolled back as a nice experiment that didn't suit the business. It wasn't the four days. It was that nobody had built a way to see the work without interrupting it. **Why the watching instinct kicks in** Worth saying plainly: managers who do this usually aren't tyrants. They're people who lost their instruments. In an office you absorb an enormous amount without trying. You can see who's stuck at 3pm, who's been in the same doc for two days, who's gone quiet in a way that means something. Remote, all of that vanishes at once, and what's left is a Slack sidebar full of green dots that prove nothing except that a mouse moved. Microsoft put numbers on the gap in its 2022 Work Trend Index. Across 20,000 people in 11 countries, 87% of employees said they were productive. At the same time, 85% of leaders said the shift to hybrid made it hard to feel confident their people were productive. Same work, two completely different readings of it. About half of managers of hybrid workers admitted outright that they struggled to trust their team to do good work unsupervised. The obvious response — watch harder, install something — makes it worse, and there's decent evidence for that. Research by Chase Thiel and colleagues, written up in *Harvard Business Review*, found monitored employees were substantially more likely to break rules: unapproved breaks, ignoring instructions, deliberately working slowly, taking company property. The explanation is a bit unsettling. When someone else is watching, people stop feeling responsible for their own conduct. You've outsourced their conscience to a dashboard, and it turns out that's not an upgrade. So the surveillance route is closed. What's left is the harder thing: making the work legible so nobody has to guess. **Let the work speak before anyone asks** The clearest tell of a healthy remote team is who starts the conversation about progress. If the manager has to ask, you've built a system that requires the manager to ask, and it will keep requiring it forever. Every "quick update?" costs the asker a bit of dignity and the receiver a bit of autonomy, and once that loop is running it's very hard to stop. Flip it. Work in public channels rather than DMs. Open pull requests as drafts. Share the doc at 40% with a line about what's still rough. Write a short Friday note — genuinely short, ten minutes, three bullets and a blocker. The bar to aim for is that a manager can answer "how's that going?" without talking to anyone. Once that's true, the status meeting dies without anyone having to kill it, which matters a lot when you've got four days instead of five. One caveat, because I've watched teams get this wrong. Visible work means artifacts and decisions, not activity. If people start posting the meetings they attended and the docs they skimmed, you've swapped surveillance for theatre and it's arguably worse, because now you're burning real hours performing. **Give people the constraint, not the instructions** The fastest way to make a capable adult feel micromanaged is to make them ask before they act. Ten small approvals a week is a curriculum, and what it teaches is that their judgment isn't wanted. Fix that by writing down, once, what people can decide alone. Spend under a certain amount. Ship behind a flag. Reply to any customer. Rewrite copy. Refactor anything they own. Then a shorter list of things that need a conversation first — pricing, headcount, public commitments, anything genuinely hard to undo. Most decisions people currently escalate are easily reversible and get treated as permanent because nobody ever said otherwise. Then hand over context instead of steps. Before something starts, the person doing it should know which constraint actually binds (is it the date, the budget, one specific customer?), what "good enough" looks like here, and what would make you scrap the whole thing. Give someone those three and they'll usually land where you'd have landed, minus the meeting where you land it for them. **Make it cheap to move a date** Reliability is the least glamorous part of trust and the one that compounds fastest. Say Thursday, deliver Thursday, repeat for a few months — that does more than any offsite. But reliability is about the prediction, not the deadline. Someone who says on Tuesday "Thursday's slipping to Monday, here's why" is being more trustworthy than someone who goes silent and delivers at 11:58pm. The rule I'd put in writing: you can move any date once, for free, as long as you move it before it arrives. What damages trust isn't slippage, it's finding out about it yourself. The trap underneath this one is subtle. Most managers punish early warnings without realising it — not with anger, but with a sudden burst of attention, a follow-up meeting, "let's do a daily check-in until this is sorted." The team learns that flagging a risk costs them their autonomy, so next time they wait and hope. If bad news reaches you late, that's rarely a problem with the team. **Count what shipped** If you change one thing operationally, retire the activity metrics. Hours logged, messages sent, response times, meeting attendance, green dots. Track what actually tells you something: did it ship, is the customer better off, how long from start to done, did the estimates roughly match reality, how much rework came back. And one leading indicator I've come to rate highly — how fast problems surface. High-trust teams report them early and messily. Low-trust teams report them late, in tidy language, usually about a week after they became unfixable. This matters more on four days than five, because a compressed week has no padding. You can't absorb a wasted Wednesday. Every outcome-based measure you put in place buys back hours that used to go on proving you were busy. **The honest bit** None of this is fast. Trust rebuilds at the speed of repeated evidence, and if a team has been managed by surveillance before, they'll assume the new approach is temporary until they watch it survive a real problem three or four times. The moment that counts isn't the announcement. It's the first time someone flags a slipped deadline and you thank them and add nothing. The four-day week doesn't create the trust question. It just strips out the slack that let you avoid answering it.