Why the four-day week keeps passing its trials

The pattern across pilots is remarkably consistent: output holds, wellbeing improves, and the overwhelming majority of participating organisations choose to continue. That consistency is genuine and also easy to over-read, because of who runs these trials and how they measure.
What the trials actually found
Across coordinated pilots in the UK, Iceland, Portugal, Spain and elsewhere, the headline results repeat. Revenue and productivity broadly held steady or improved modestly. Self-reported burnout, stress and fatigue fell substantially. Sick days and resignations declined. And when trials ended, the large majority of participating companies retained the arrangement.
The most common design was "100-80-100": full pay, eighty percent of the hours, with an expectation of maintained output. This matters, because it is a productivity intervention rather than a reduction in work, and conflating the two is where most arguments about it go wrong.
The mechanism, which is less mysterious than it sounds
The improvement does not come from people working harder in less time. It comes from removing work that was not producing anything, and a four-day week forces that removal in a way that ordinary efficiency initiatives never do.
Participating organisations report the same specific changes: fewer and shorter meetings, with many replaced by written updates; blocks of uninterrupted time protected deliberately; clearer prioritisation because there was no room for everything; and reduced tolerance for work whose purpose nobody could articulate. Most knowledge work contains a meaningful amount of activity that survives only because nobody has been forced to justify it. A hard constraint on hours is an unusually effective forcing function.
The catch, stated plainly
There are four legitimate objections, and dismissing them is why advocates lose the argument with people who actually run operations.
- Selection bias. Organisations that volunteer for a four-day-week trial are not typical. They tend to be smaller, knowledge-based, already reasonably well managed and led by someone enthusiastic. That is not the average workplace, and results from that population do not straightforwardly transfer.
- Measurement difficulty. Wellbeing outcomes are self-reported by people who want the arrangement to continue. Productivity in knowledge work is genuinely hard to measure, and revenue over a six-month window is noisy. The direction of the findings is credible; the precision is not.
- Coverage-based work does not compress. A hospital ward, a retail floor, a call centre or a production line needs bodies present for a fixed number of hours. Reducing hours per person means hiring more people, which is a cost increase rather than an efficiency gain. Most four-day-week evidence comes from work where output is not tied to attendance.
- The slack was finite. The first round of cuts removes genuine waste. It is unclear whether the gains persist once that waste is gone, and the longest-running studies are still short by the standards of this question.
The variants, which are not equivalent
"Four-day week" covers arrangements with very different effects. Reduced hours at full pay is the version the trials tested. Compressed hours — the same forty hours in four longer days — is a different thing entirely, and the evidence for it is weaker, with some indication that very long days erode the wellbeing benefit.
Staggered days off, where the business stays open five days but individuals work four, preserve customer coverage but lose the collective quiet that participants cite as the main productivity source. Seasonal versions, applied only in quieter months, are the most common real-world compromise and the least studied.
What implementation actually requires
The organisations that succeed do the operational work first. They audit where time goes and cut meetings before cutting days. They agree explicitly what will not be done any more, rather than hoping it compresses. They define response-time expectations for customers and communicate them. And they decide in advance how the arrangement will be judged, with a stated route to reverting.
The ones that struggle announce the day off and expect the efficiency to appear on its own. What happens instead is that the same work is done in four days by people quietly working on the fifth, which produces the worst outcome available: unchanged output, hidden hours and a policy everyone has learned to distrust.
Why so few organisations have adopted it anyway
Given consistent positive results, low adoption needs explaining. Part of it is genuine unsuitability — much of the economy is coverage-based. Part is coordination: an organisation whose clients, suppliers and regulators all operate five days pays a real cost for being out of step.
And part is straightforwardly that changing how work is organised is difficult and risky for whoever proposes it. The downside of a failed attempt lands on an identifiable person; the upside is diffuse. That asymmetry explains a great deal of organisational inertia, entirely independent of the evidence.
What the results probably do mean
The defensible conclusion is narrower than the headlines but still significant: in knowledge work, a substantial amount of the standard week is not producing output, and a hard constraint is an effective way to find it. Whether the answer is four days, or shorter meetings and protected focus time within five, is a separate question the trials do not settle.
Which is arguably the more useful finding. Most of the reported benefit came from changes any organisation could make on Monday without touching the working week at all.
The questions to ask before proposing one
Anyone considering raising this internally is better served by operational answers than by the trial results, because the trial results are what the objections will be aimed at.
Four questions do most of the work. Is our output tied to hours of coverage, or to work completed? If it is coverage, the arrangement means additional headcount and should be argued on retention grounds rather than productivity. What specifically stops being done, named concretely rather than assumed away? What do customers experience, and have we told them? And how would we know it was failing — which metric, measured over what period, with what threshold for reverting?
An organisation that can answer those four has usually already found most of the benefit in the process of answering them, which is the quietly recurring lesson from every pilot that has run.
Trial designs, sectors and measurement methods vary considerably. Read the underlying studies rather than summaries before drawing conclusions for a specific workplace.