Ask a leadership team to define productivity for a four-day pilot and you will often get one of two failures: a single revenue figure that proves nothing on its own, or a dashboard of twenty metrics that no one reads and everyone games. Neither survives contact with a sceptical finance director.
The alternative is unglamorous and effective: two or three honest proxies per function, agreed before the pilot, baselined for at least four to eight weeks, and reported alongside actual hours worked. This article borrows the scorecard logic from the major pilot programmes and makes it function-specific.
The rules that keep proxies honest
Three rules matter more than the particular measures you choose.
First, pair every output measure with a quality measure. Units shipped without defects, tickets resolved without reopeners, projects delivered without rework. Output alone rewards rushing, and rushing is how a shorter week becomes quieter harm.
Second, always divide by hours actually worked, and publish the hours. Total output can hold steady because the pilot “worked” or because the work migrated to Friday evenings. Timesheets may be imperfect, but a pilot with no hours measure is asking staff to grade their own honesty. After-hours system activity and overtime claims are useful tripwires alongside self-report.
Third, respect small denominators. If a team of nine had two resignations last year and one this year, that is not a 50% retention improvement; it is a small number moving. The UK pilot’s headline leaving and sick-day percentages came with exactly this caution from the researchers themselves. Report counts alongside percentages, state the base period, and agree in advance how much movement is signal rather than weather.
Proxies by function
Pick from these; do not adopt them all.
Sales. Revenue per full-time-equivalent (and per hour worked), and conversion rate through the pipeline. Revenue alone confuses market weather with working pattern; conversion helps separate the two.
Customer support. Resolved cases weighted by complexity, quality-assurance-adjusted, plus first-response and full-resolution times. Raw ticket counts reward cherry-picking easy work. Your customer floor — the service level below which the pilot pauses — belongs here.
Engineering and product. Cycle time from started to shipped, escaped-defect rate, and shipped value against the roadmap you committed to. Avoid story-point totals: they inflate under pressure and everyone in the room knows it.
Professional services. Delivered milestones on time, write-offs, and client satisfaction at project close. If your firm sells days, utilisation must be re-based for 32 hours before the pilot, or the proxy will “prove” failure by construction.
Manufacturing. Overall equipment effectiveness, scrap and rework, and safety incidents read through a fatigue lens. A proxy set without safety is incomplete in any setting where hours and physical work meet.
Care and coverage services. Client or patient outcomes, waiting times, and agency hours bought. If shorter time requires more staff, agency spend is where the honest cost conversation happens.
Intensity belongs on the same page
Add one people-side tripwire beside the proxies: a short, repeated intensity pulse (“this week’s pace felt sustainable,” agreed scale) plus actual hours. Krystal, the UK hosting firm that reverted, reported stress rising and support quality falling — the pattern of work compressed rather than redesigned. Its lesson is not that measurement failed; it is that intensity was the finding, and it was heeded too late to redesign within the trial.
A practical template: for each function, one page. Three proxies, each with baseline, floor, and review date; hours worked next to them; one quality measure; one intensity pulse. Review monthly in the pilot. If a proxy is being gamed, you will see quality sag while output “improves” — that divergence is itself the most honest metric you own.
Measure less, but mean it. A scorecard small enough to be read is the one that will still be running when the pilot decision arrives.
Sources
- Pilot scorecard approach across business, quality, customer, people and financial measures — UK pilot and 4 Day Week Global programme reporting, 2022–2023.
- Germany four-day week pilot measures and outcomes — University of Münster and Intraprenör, October 2024.
- Krystal’s reversal and reported work-intensification pattern — Not Ltd reporting, 2023.
Day5Group is a consultancy that assists in the transition to a 4 day work week. More about Day5Group · All blog articles