The four-day week, backed by the biggest trial yet

Most of the modern wellbeing literature is a catalogue of things that don’t work. Mindfulness apps that move nothing. Resilience workshops that, in the largest study of them, trended slightly negative. Step challenges and meditation seminars and wellness portals that survey after survey finds make no measurable dent in how people actually feel. If you read enough of it, you start to suspect the entire field is a polite way of asking exhausted people to cope harder.

So it is worth pausing on the one intervention that keeps coming back with strong, positive, hard-to-explain-away results — and noticing what makes it different from all the apps. It does not ask the individual to change. It changes the work.

The intervention is the four-day week, and in 2025 it got its most serious test yet.

The biggest trial to date

The headline study, by Wen Fan and Juliet Schor and colleagues, appeared in Nature Human Behaviour in July 2025. It is the largest look at a reorganised, shorter working week so far: about 141 organisations and roughly 2,896 employees, spread across six countries — Australia, Canada, Ireland, New Zealand, the United Kingdom, and the United States. Companies moved to a four-day schedule while holding pay flat, and researchers tracked how workers felt before and after, comparing them against employees at organisations that did not make the switch.

The design matters because it is what most wellbeing research lacks. The studies that find mindfulness apps doing nothing are damning precisely because they compare people who used a program with people who didn’t, or who measure before and after with a control group to lean on. This trial did the same on the other side of the ledger — and where the apps came up empty, the shorter week did not.

Across the trial, self-reported burnout fell sharply: on the order of 0.44 points on a five-point scale, which in some of the participating trials worked out to roughly a 71 percent reduction in burnout symptoms. Job satisfaction rose. Self-rated mental health improved. So did physical health. None of these is the kind of number the wellness-app literature produces; they are the kind it conspicuously fails to.

And the part that makes executives lean in: output did not collapse to pay for it. By the trial’s own measures, productivity broadly held or improved. About nine in ten of the companies that ran the experiment chose to keep the four-day schedule after it ended — which is, in its own way, the most honest data point in the study. Firms do not voluntarily keep an arrangement that quietly wrecks their results.

What the control group tells you

The improvements are only meaningful because of what happened to everyone else. The trial included a comparison set of organisations — a dozen control companies — that carried on with the ordinary five-day week. Over the same window, those workers showed no comparable gains. Burnout did not fall for them. Their wellbeing did not lift on its own.

That contrast is the whole argument. If burnout had fallen everywhere in 2025 — because the economy eased, or the season changed, or some unrelated mood lifted — you could not credit the four-day week for the trial group’s gains. The flat control line is what lets you draw the causal arrow. Something specific happened to the people who got the shorter week, and it did not happen to the people who didn’t.

Researchers also looked at why the trial group felt better, and the pathway is unglamorous and believable. The improvements ran through reduced fatigue, better sleep, and a higher sense of work ability — the feeling of being equal to the demands of the job. This is not a runner’s-high story about an extra day of leisure. It is a recovery story. Given a real chance to rest, bodies and minds did what they do, and the wellbeing numbers followed.

The mechanism nobody puts on the poster: 100–80–100

Here is where most enthusiastic write-ups quietly mislead you. The four-day week that works is not “the same job, minus a day.” If you simply delete Friday and change nothing else, you have not given people a four-day week — you have given them a five-day workload to cram into four, which is a recipe for more strain, not less.

The trials are built on a different bargain, usually summarised as 100–80–100: workers keep 100 percent of their pay, drop to about 80 percent of their hours, and commit to delivering 100 percent of the output. The number that is missing from that formula is the one that does the work. To hold output while cutting a fifth of the hours, the organisation has to find and remove a fifth of the time that was being wasted.

And it turns out there is usually that much to remove. Before these trials begin, participating companies do real preparation: they audit how time is actually spent and start cutting the low-value parts. The standing meeting that could have been a message. The recurring sync where six people listen to an update meant for two. The status report nobody reads. The interruption-driven workday — for knowledge and data teams especially — where attention is shredded into fragments too short to finish anything hard. The four-day week is the forcing function that finally makes a company confront all of it, because once the hours are genuinely scarce, busywork can no longer hide.

So the honest framing is almost the opposite of how it is usually sold. The extra day off is not the cause of the gains. It is the reward for the redesign — and the deadline that makes the redesign happen. Take the day off without the redesign and you get the crammed-four-days disaster. Do the redesign without the day off and you have, in fact, done most of the valuable part anyway.

Why “shorter” does not have to mean “less”

This is the piece that breaks intuition, so it is worth stating plainly. The instinct — fewer hours must mean less work done — assumes that hours and output are the same thing. For a lot of modern work, they badly aren’t.

A factory line is roughly linear: stop it an hour early and you make an hour fewer widgets. But cognitive work does not run on a line. A debugging session, a model that needs rethinking, a tricky piece of analysis — these depend on held concentration, and concentration degrades as people tire. The tenth hour of a long week is not as productive as the third; it is often where the bugs get introduced rather than fixed, generating rework that eats into some later week. A genuinely rested engineer working a focused 32-hour week can plausibly out-produce an exhausted one nominally “working” 45, because so many of those 45 hours were spent fragmented, fatigued, or quietly cleaning up the mistakes of hour ten.

That is why the productivity result is less mysterious than it first appears. The trials did not discover that you get the same work from less effort. They discovered that a lot of the deleted hours were never producing much, and that the rest got more productive once people were rested enough to think.

The honest caveats

None of this makes the four-day week a universal switch you can flip. A few cautions keep the enthusiasm grounded.

The wellbeing and productivity measures are largely self-reported, and the organisations that volunteer for a four-day-week trial are not a random sample of all employers — they are, by definition, places already willing to rethink how they work, which may flatter the results. The schedule also travels unevenly across kinds of work. Output that is genuinely tied to time-in-seat or continuous coverage — an emergency room, a support desk that must be staffed every hour, a production line — cannot simply compress the way a software team’s can; making a four-day week real there means hiring and rota redesign, not just trimming meetings. And the gains depend entirely on the redesign actually being done. A company that announces a four-day week as a perk, skips the hard work of cutting low-value tasks, and then watches everyone work frantic ten-hour days will conclude the idea failed, when what failed was the implementation.

Treat the evidence, then, as genuinely strong but conditional. It is the best positive result in the wellbeing field — and it comes with instructions that most of the excited coverage leaves off.

What it means if you can’t get a four-day week

Most readers of this cannot unilaterally move their team to four days. That decision sits above your pay grade, and your organisation may never make it. But the reason the four-day week works is not locked inside the policy. The policy is just a dramatic way of forcing a principle that you can apply on your own, this week, with no one’s permission: protect focused time, and cut low-value work.

Look at what the redesign actually consists of, and almost all of it is available at the level of an individual or a single team. You can decline the meeting that could have been a message, and propose ending the recurring sync nobody needs. You can defend a block of genuinely uninterrupted time the way the trials force companies to defend it, and treat the reflex to be instantly responsive as the productivity tax it is. You can ask, of any recurring task, the question the four-day week forces on a whole company — if my working hours were suddenly cut by a fifth, would this survive the cut? — and quietly stop doing the things that wouldn’t.

That is the real lesson hiding inside the biggest trial yet. The headline is the extra day off, but the finding underneath is more useful and more portable: a surprising share of the standard working week produces nothing, and rest is not the opposite of productivity but one of its inputs. You may not get to drop a day. You can still drop the busywork — and protect the hours where the work that actually matters gets done.