Not every four-day week in the wild is unconditional. In the UK 2022 pilot, participating firms used a family of variants: for some staff the fifth day came with fewer other leave days; in some firms staff remained on call and could be called in; in others, keeping the day depended on meeting performance conditions. Leaders considering a pilot often ask whether they should do the same.
Conditional designs are permissible. They are not free. Each condition erodes part of the benefit the day is supposed to deliver, and the erosion falls unevenly. This article sets out the variants, what they buy, what they cost, and the disclosure standard we would insist on.
The three common conditions
On-call-in arrangements. The day off is real unless the phones demand otherwise. The genuine use case is coverage: a small team serving clients five days cannot leave a queue unanswered. Handled well, calls are rare, rotated, logged and compensated with time back the following week. Handled badly, “on call” becomes “usually working, but calling it a day off.” Staff cannot commit to childcare, study or rest because the day is not actually theirs. Recovery research points to predictability as part of the benefit; a day that may be reclaimed at 8 a.m. delivers a fraction of it.
Performance-contingent days. The day continues while targets are met and is “paused” when they are not. This attracts leaders who fear a permanent, irreversible commitment. The difficulty is definition. Team targets that dip because demand falls, a supplier fails, or a colleague is off sick can withdraw the day for reasons nobody controlled — and staff experience that withdrawal as punishment for the system’s failures. The condition also invites gaming: targets soften so the day survives, and your measurement honesty, the thing a pilot most needs, is the first casualty.
Leave-adjusted designs. Some firms traded the weekly day partly against annual leave entitlements. This can be a legitimate accounting choice — the total time benefit is smaller but still real. It is also the condition most likely to surprise staff at contract stage, and surprise is corrosive.
What conditions buy, and what they cost
The honest case for conditions is reversibility and coverage. A board that will approve a pilot with a pause trigger may refuse one without. A service that must answer on Fridays needs a fallback that is more than goodwill.
The cost is dose. The Nature Human Behaviour study published in 2025 (2,896 staff across 141 organisations, with control companies) found a dose-response: staff who actually cut more hours gained more in burnout, satisfaction and health, and it mattered that hours genuinely fell — around a quarter of trial participants reported no real change in hours. Conditions that frequently reclaim the day, or leave people half-working it, shrink the dose and then puzzle over the smaller effect.
There is a second cost: monitoring burden. Every condition needs a rule, an owner, a log and an appeal route. A firm that cannot yet count actual hours worked has no business running performance-contingent days, because it will be judging from presence and impressions.
A decision checklist
Work through these before choosing a conditional design:
- State the condition in the written trial terms, in plain language, before launch. No conditions discovered mid-pilot.
- Make call-ins rare, rotated and logged, with time back within a fortnight. Count them; a rising call-in rate is a staffing finding, not a staff failing.
- If performance gates the day, gate it on team outputs the team controls, measured over a period long enough to absorb a bad week.
- Publish how often the condition has been triggered, to staff, monthly.
- Guaranteed minimum: name the least any individual will receive (for example, at least two clear days per month), so nobody plans a life around nothing.
- Set a sunset: conditions agreed for the pilot are reviewed at the decision date, not quietly made permanent.
Our default advice is simple. Start guaranteed for the pilot period, measure honestly, and use the decision date — not a weekly lever — as your control mechanism. A day off that management can switch off will be treated by staff as a day of permitted absence, and permitted absence does not redesign work, calm a nervous system, or recruit anybody.
If a coverage condition is genuinely unavoidable, say so, price it, and disclose it. The firms that lose trust are rarely the ones with conditions. They are the ones whose conditions were a surprise.
Sources
- UK four-day week pilot design variants and results — Autonomy, University of Cambridge, Boston College and 4 Day Week Global, February 2023.
- Fan, Schor, Kelly & Gu, four-day week trial study — Nature Human Behaviour, 2025.
Day5Group is a consultancy that assists in the transition to a 4 day work week. More about Day5Group · All blog articles