Cut the work before you cut the hours. In the UK's 2022 four-day-week pilot, revenue rose 1.4% on average from the start of the trial to the end among the 23 organisations that supplied figures, and the companies got there by rebuilding meetings and workflows first. Those firms volunteered. That is one documented cohort, not a rule.
The shorter week is usually sold as a perk. In the published trial data it behaves more like an operations project with a deadline attached: you have four days of capacity and five days of commitments, so something in the process has to go. For a ten-person company, that pressure is the whole mechanism. This is general information about documented company practice, not legal, HR, or employment-law advice — contracts, overtime rules, and payroll changes need your own counsel.
What did the trials actually measure?
They measured self-reported well-being and company-reported business metrics over six months, not randomised productivity. The largest UK pilot ran June to December 2022 with 61 organisations and roughly 2,900 employees at 20% fewer hours and no pay cut, according to the University of Cambridge sociology department's summary of the results.
The headline outcomes from that pilot: 71% of employees reported lower burnout, 39% reported less stress, absenteeism fell 65%, and resignations fell 57% against the prior year. Fifty-six of the 61 companies said they intended to continue.
The academic follow-up is broader. Wen Fan and Juliet Schor of Boston College studied 2,896 workers across 141 organisations in six countries and published the results in Nature Human Behaviour in 2025; workers reported less mental strain, better sleep and less fatigue, with the biggest gains among those whose weekly hours dropped by eight or more, as reported by Newsweek. The authors flagged the limits themselves: companies opted in, skewed small, and clustered in English-speaking countries, and they called for randomised, government-backed trials.
Read that as a ceiling on what you can claim internally. The evidence supports "this has worked for self-selected small firms that redesigned their work." It does not support "this will raise our output."
Which schedule model fits a small team?
Pick the model that matches your coverage obligation, not the one that sounds cleanest. The pilot report published by Autonomy documents five patterns companies used, and the split was not tidy: of 44 responding companies, 14 gave everyone Friday off and 11 had no common day off at all.
| Model | How it works | Fits when |
|---|---|---|
| Fifth-day stoppage | The whole company closes one extra day | No live coverage duty; you want protected collaboration days |
| Staggered | Teams alternate days off, often with a paired backup | Support or sales lines must stay open five days |
| Decentralised | Each department designs its own pattern | Functions have genuinely different rhythms |
| Annualised | Hours average out across the year | Demand is seasonal and lumpy |
| Conditional | The short week is tied to hitting performance conditions | You need a reversible trial with a stated bar |
The staggered and decentralised versions cost more coordination and are where a small team usually breaks. If you cannot name who covers a Thursday escalation before the trial starts, you are not ready to start.
How do you find the hours to cut?
You find them in meetings, interruptions, and reporting, in that order. The redesign steps the pilot report lists are unglamorous, and they are the part founders tend to skip.
- Audit the standing calendar. List every recurring meeting, its length, and its attendee count. Kill, shorten, or shrink the invite list on each one.
- Set an asynchronous default. The report describes companies reforming email and messaging protocols so fewer decisions need a live conversation.
- Protect focus blocks. Participating firms built "heads down" periods and reorganised calendars around single-tasking rather than parallel work.
- Automate the recurring reports. Workflow automation and project-management software show up repeatedly in the same list of changes.
- Only then remove the day. The hours have to already be gone from the process, or they reappear as unpaid evening work.
Boston College's account of the pilot describes the same mechanism from the employee side: firms eliminated low-value activities, particularly meetings, and staff used the free day for appointments and rest rather than compressing errands into working days, per the university's write-up of the research.
What should you track before you start?
Baseline the four numbers the trials moved, and take the reading before you announce anything. Without a before, you will be arguing about vibes in month three, and a small team's month-to-month noise will swamp any real signal.
Track monthly sick days per employee, voluntary resignations per 100 employees, revenue against the same period last year, and actual hours worked. In the UK pilot, absenteeism fell from about 2.0 to 0.7 days a month and resignations fell from 2.0 to 0.8 per 100 employees, per the Autonomy report — those are the shapes to watch, not targets to promise.
One caution on hours: the same report found average working time fell from 38 to 34 hours a week, and 62% of employees said the pace of work increased. A shorter week that raises intensity without removing tasks is a different trade than the one the data describes. Ask about pace directly in your check-ins.
When is a four-day week the wrong move?
When your delivery is billed by the hour, when one person is the single point of failure on a live system, or when you are already behind on commitments to customers. None of the trial mechanisms fix an understaffed team; they redistribute time in a team that has slack hidden in its calendar.
It is also the wrong move as a retention patch applied in a hurry. The measured retention gains came alongside six months of process work and, in the Cambridge summary of the trial, 18 companies making the change permanent only after that run. A trial you can reverse, with a stated end date and a written bar for continuing, is the version that survives contact with a bad quarter.
If you run one, say out loud that it is conditional. Then honour whichever answer the numbers give.
For a related lifestyle perspective, read Glyciome Awarded Landmark Patent for Prebiotic Cervical Cancer Prevention Approach Focused on Under-Served Populations.
For more context, read Complete Guide to Comic Book Collecting for Beginners.
For more context, read Three documented ways to keep AI out of your evenings.
For more context, read What the WARN Act Actually Owes Workers When a Plant Closes.