The four-day week sounds like a perk that would tank output. Two of the largest trials found the opposite, and the takeaway that matters for engineering leaders has almost nothing to do with the number of days.
Key Takeaways
- In the UK's largest trial (61 companies, ~2,900 workers), revenue held broadly steady and resignations fell 57% (University of Cambridge).
- A separate US/Ireland trial (33 companies, 900+ workers) saw average revenue rise 38% and 97% of workers wanting to keep the schedule (4 Day Week Global).
- Across both, roughly 90% of companies kept the shorter week afterward.
- The real lesson: hours are a poor proxy for output. Manage delivery, not time-in-seat.
What the Trials Actually Showed
The UK pilot ran 61 companies and around 2,900 workers through six months at full pay, per the University of Cambridge team that analyzed it. Revenue stayed broadly steady over the trial (up about 35% year on year, per 4 Day Week Global), resignations dropped 57%, and absenteeism 65%.
The US and Ireland ran their own coordinated trial, and the numbers rhyme. Across 33 companies and more than 900 workers, average revenue rose 38% over the trial period, burnout and stress fell, and average sleep went from 7.02 to 7.72 hours a night. Workers rated the experience 9.1 out of 10, and 97% said they wanted to keep it (4 Day Week Global). When it ended, not one of the 27 companies that responded planned to go back to five days.
Two independent trials, two continents, the same result. That's hard to wave off as a fluke.
Why This Matters Even If You Never Cut a Day
You don't have to adopt a four-day week to learn from it. The result quietly kills a common assumption: that output scales with hours in the seat. It doesn't. A team that produced the same value in four days wasn't using the fifth for value. It was spending it on busywork, meetings, and the drag of a long week.
| Assumption | What the trials showed |
|---|---|
| Fewer hours means less output | Output held or rose |
| Long hours signal commitment | Long hours signaled burnout risk |
| Presence equals productivity | Presence and output are loosely linked |
A Concrete Version
Say your team "works" 45 hours a week and you're proud of the grind. Strip out the four hours of status meetings that could have been a doc, the couple of hours lost re-focusing after interruptions, and the Friday-afternoon fade when everyone's cooked. The real productive core might be 30 hours. The trials are basically that arithmetic run at scale: cut the low-value time, protect the focused time, and output holds. The fifth day was mostly the fade.
The Honest Counterpoint
A four-day week isn't automatically right for every company, and the trials had selection bias. The companies that opted in were already willing to rethink how they work, and some found it genuinely hard, customer-facing and support-heavy businesses especially. "Adopt a four-day week tomorrow" is not the lesson. The durable, transferable one is narrower and safer: hours are a weak proxy for output, so stop managing your engineers by them.
The Real Lesson: Measure Output, Not Hours
If hours are a weak proxy for value, then managing engineers by time-in-seat, or by which timezone they sit in, is managing the wrong variable. Measure delivery with something like DORA metrics and the question of when and where people work mostly dissolves. It's the same reason timezone overlap beats a rigid schedule: you want collaboration when it counts and deep focus the rest of the time, not maximum hours logged. See available engineers.
Frequently Asked Questions
Does a four-day week hurt output?
Two large trials (UK, and US/Ireland) found output held or rose. Revenue was broadly steady in the UK and up 38% on average in the US/Ireland trial, while resignations and burnout dropped.
Should my startup adopt a four-day week?
You don't have to. The trials had selection bias and it doesn't fit every business. The transferable lesson is that hours are a weak proxy for output, so manage by delivery instead.
How do I manage engineers by output?
Track delivery metrics like deployment frequency and lead time (DORA) rather than hours or presence. That works identically for in-office, remote, and nearshore teams.
Why didn't revenue fall with fewer hours?
Because a lot of a standard week is low-value time: status meetings, re-focusing after interruptions, and end-of-week fade. Cutting that while protecting focused work leaves output roughly intact.
The Bottom Line
The four-day week's real finding is that output and hours are only loosely linked, and two independent trials back it up. Whether or not you cut a day, manage by delivery, and the questions of hours, location, and timezone stop being threats.
Roberto Espinoza is CEO of Ruzora, which helps US startups hire pre-vetted senior LATAM engineers in 72 hours. See available engineers.
