The phrase of the month is quiet quitting, which on inspection turns out to describe people doing the job they are paid for and declining the unpaid extra. It has struck a nerve because it names something managers have felt since the spring: discretionary effort is being withdrawn, and pay rises alone are not bringing it back. Against that backdrop, around seventy British companies and more than three thousand employees are halfway through a six-month four-day-week trial that began in June and will not report until later this year. Which makes this a good moment to be precise about what the existing evidence actually shows, because the gap between the headlines and the findings is wide.
What the completed trials found
Three bodies of evidence are worth taking seriously: the Icelandic public-sector programme that ran from 2015 to 2019 across thousands of workers, the New Zealand professional services firm that trialled and then adopted a four-day week in 2018, and the Japanese subsidiary experiment in 2019 that reported a large jump in sales per employee. Read carefully, they say three things. Output held where organisations removed work rather than compressing it. The units that succeeded cut meetings, shortened the remainder, replaced standing updates with written ones and cancelled low-value activity. The units that simply asked people to do five days of work in four did not. Wellbeing improved substantially and consistently. Stress, burnout and perceived work-life balance moved in the same direction in every trial, by margins large enough to survive considerable scepticism. This is the most robust finding in the literature, and notably it is more robust than the productivity claim. The evidence base is weaker than its reputation. Participants self-selected, samples were small, much of the measurement was self-reported, and the public-sector programme mostly reduced hours to thirty-five or thirty-six rather than to four days. Anyone citing a specific productivity percentage is quoting a single case study.
Three different policies share one name
Reduced hours at full pay — roughly eighty per cent of the time for one hundred per cent of the pay, with a productivity commitment. This is what the current trials are testing. Compressed hours — the same weekly total worked in four longer days. Belgium legislated a right to request exactly this in February. It is a scheduling change, not a reduction, and its effects are entirely different. Flexible or annualised arrangements — a shorter week in quieter periods, longer in peak. These three have different costs, different legal implications and different consequences for fatigue. Conflating them is the fastest way to a pilot that answers no question at all.
| Arrangement | What changes |
|---|---|
| Reduced hours | Less contracted working time; test output and coverage. |
| Compressed week | Existing hours are worked across fewer days. |
| Flexible schedule | Working times vary; total hours need not fall. |
Qualitative summary of this article's source text, not a measured outcome or performance estimate.
You cannot remove a day by working faster
This is the mechanism, and it is less inspiring than the marketing. Nobody in the successful trials became twenty per cent more efficient at their actual work. What happened is that organisations spent the preparation period auditing where the week went and discovered that a substantial share of it was coordination: status meetings, update meetings, meetings to prepare for meetings, and the interruption cost around them. The day comes out of the coordination budget, or it does not come out at all. Which means the preparation work — cutting the meeting inventory, converting recurring updates into written artefacts, protecting focus blocks, and explicitly cancelling activities — is not the enabler of the pilot. It is the pilot. Several organisations have found that doing only that part, with no change to the working week, delivered most of the wellbeing benefit.
Where it does not transfer
Coverage work — support desks, clinical care, retail floors, manufacturing lines, warehouses, logistics — does not behave this way. There is no coordination fat to remove, and the hours exist because the service window exists. For these populations a four-day week is a rostering and headcount question with a direct cost, and pretending otherwise is how pilots collapse into resentment between departments. That does not mean nothing is available. Predictable rosters published further in advance, autonomy over shift swaps, and guaranteed consecutive days off frequently rank higher in these workforces than a fourth day would, and cost far less.
Designing a pilot that produces an answer
Choose two or three output measures that already exist — tickets resolved, cases closed, revenue per head, cycle time. Inventing a metric for the pilot guarantees an unfalsifiable result. Baseline for eight to twelve weeks before anything changes. Run for at least three months, preferably six. The first month is a honeymoon and the third is where the real pattern appears. Keep a comparison group — another team, another site, or the same team the previous year. Measure the hidden costs explicitly: customer response latency, handover errors, overtime creep, and how much work people are doing on the supposed day off. Use a validated wellbeing instrument at baseline, midpoint and end, rather than a cheerful internal survey. Write the stop and continue criteria before you start, with numbers attached, and have the executive sponsor sign them. Publish the results internally, including the parts that failed. A pilot whose findings cannot embarrass anyone was never an experiment.
Five ways these pilots fail
The compressed-day trap, where people work ten-hour days and report worse fatigue than before. The invisible fifth day, where staff quietly work anyway and the measurement shows a miracle. Manager exemption, where leadership keeps meeting on the off day so nobody can genuinely disconnect. Customer mismatch, where clients and suppliers operate a five-day week and the gap lands on whoever is most junior. And the unmeasurable pilot, which ends with everyone agreeing it felt good and no basis for a decision.
Practical Guidance for Work Model Pilot Design
- Decide which of the three policies you are testing and say so in writing.
- Audit the meeting and coordination load first; the day comes from there or nowhere.
- Baseline existing metrics for at least two months before changing anything.
- Name the stop criteria and the decision date in advance.
- Design coverage populations separately, with rostering answers rather than delayed inclusion.
- Track overtime, off-day working and response latency as first-class measures.
- Hold managers to the same rules, or the benefit evaporates for everyone below them.
- Test the customer-facing interface — who answers on the fifth day, and how that is communicated.
The Regional Angle
Three regional facts change this conversation, and the first is that this market has already run the experiment. Every year, for a month, working hours across the Gulf are reduced by law during Ramadan — typically by two hours a day, which is a larger cut than any four-day-week trial has attempted. Organisations here already know, informally, what happens: decisions still get made, meetings compress dramatically, knowledge work largely holds, and what suffers is coordination-heavy activity and anything dependent on counterparties keeping a different schedule. In other words, the regional experience supports the central finding of the international trials. What is missing is measurement. Almost nobody baselines output before Ramadan, tracks it through, and compares. Any organisation here considering a shorter-week pilot has an annual natural experiment available at zero cost, repeated for years, and has never instrumented it. Doing so once would produce better local evidence than any consultant's benchmark. The second is that the region changed its working week eighteen months into a pandemic and learned something useful. The Emirates moved the public sector to a four-and-a-half-day week with a Saturday and Sunday weekend at the start of this year, while neighbouring markets kept the Friday and Saturday weekend, and the private sector split — some following, some not. The friction that followed was almost entirely coordination rather than productivity: cross-border meetings, banking cut-offs, shared services covering multiple jurisdictions, and family logistics where spouses work different weeks. That is precisely the lesson for a four-day pilot in a group with entities in several countries. The internal cost is manageable; the external calendar mismatch is where the pain lands, and it must be designed for rather than discovered. The third is contractual. Employment contracts here specify working hours and are registered with the labour authorities, and hours interact with overtime entitlements and with end-of-service accrual. A pilot that quietly changes the working week without addressing the paperwork creates a gap between the registered contract and the actual arrangement, which is exactly the kind of discrepancy that surfaces during an inspection or a termination dispute. The clean approach is to run the trial as a documented flexible working arrangement with an explicit end date, take advice on whether registered terms need amendment, and confirm that pay, overtime treatment and service accrual are unchanged. None of this is difficult. It is simply the step that enthusiastic pilots skip.
The objection worth taking seriously
The strongest objection is about evidence quality, and it is a good objection. The trials were self-selected, the participating firms were mostly small, autonomous knowledge businesses with obvious slack in their weeks, and the most-quoted productivity figures come from single cases with commercial interests in the result. Much of the published research comes from organisations that advocate for the policy. The wellbeing gains may partly reflect novelty and the simple fact of receiving an extra day, which would follow from almost any comparable benefit. And proposing to cut capacity by a fifth during a downturn, with hiring frozen and demand flat, is a difficult conversation to open. The evidence criticism is correct and should be stated openly rather than argued away. There is no large randomised trial and there may never be one. But the wellbeing and retention findings are consistent across quite different designs, countries and sectors, which is worth something even after heavy discounting. The downturn argument cuts both ways: when pay increases are unavailable, non-wage compensation becomes the only lever left, and it is cheaper than the attrition it prevents. And the proposal here is not adoption. It is a measured experiment with pre-committed stop criteria, which costs one quarter and produces a defensible answer in either direction — which is more than the current debate has managed.
Common Questions
Does it work for client-facing teams?
Only with explicit coverage design. Staggered days off across the team, published availability, and an escalation path preserve service; a blanket closure does not survive a demanding client.
What happens to pay and benefits?
In the reduced-hours model, nothing changes — that is the point, and it is why the productivity commitment has to be explicit. Compressed-hours models raise different questions about overtime and fatigue.
How long before we know?
Three months minimum. Anything shorter measures enthusiasm.
What should we expect over the next twelve months?
Expect the current British cohort to report in the autumn and expect the coverage to overstate whatever it finds. Expect more governments to legislate the compressed variant, as Belgium has, because it costs employers nothing. Expect this debate to merge with the discretionary-effort argument now running under the quiet-quitting label, since both are really about how much of a person's week an employer can claim. Expect several visible reversals from organisations that shortened the week without removing any work. And expect regional employers to experiment at the edges — summer schedules, flexible contract types, earlier Friday finishes — rather than adopt the full model, which is a reasonable place to start if the measurement is real.
Work Model Pilot Design — we define which policy you are actually testing, baseline the metrics that already exist, and write the stop criteria before the pilot starts so the result can survive a board challenge.