A pilot without a scorecard ends in an argument about impressions. A pilot with one ends in a decision. The difference is agreed before anyone takes a day off: twelve numbers, baselined, with floors and tripwires attached, so that month six produces a verdict rather than a debate about whose anecdotes count.
Why twelve, and why in advance
Twelve is enough to cover the business honestly and few enough that a leadership team will actually read the pack each month. The advance part matters more than the exact list. Metrics chosen after results arrive will always — innocently — favour the person choosing. Metrics agreed at launch protect the pilot twice: staff trust that the goalposts will not move, and sceptics on the board get their hardest questions wired in from the start rather than saved for the decision meeting.
The UK 2022 pilot and its one-year follow-up showed the value of a mixed scorecard: wellbeing and retention measures moved strongly, revenue moved modestly, and neither alone would have told the truth. At one year, 89 percent of the original firms were still running a shorter week and around half had made it permanent — a durability figure that only means something because the cohort kept reporting against the same questions.
The twelve numbers
Business outputs (pick per team). 1) Your team’s defining output units — cases closed, milestones delivered, orders shipped — reported as totals and per hour worked. Reporting per hour is what separates genuine redesign from speed-up. 2) Revenue or bookings for the pilot window, treated as context rather than proof, because markets move for reasons that have nothing to do with Fridays.
Quality. 3) Error or rework rate. 4) Complaints or audit findings. Quality is where intensification surfaces first, and it is the metric sceptics trust most when it holds.
Customer. 5) First-response time. 6) Resolution time and service-level attainment. 7) A satisfaction measure — CSAT, NPS or your equivalent — sampled consistently. Agree a customer floor in advance: the level below which the pilot pauses and redesigns, whatever the wellbeing scores say.
People. 8) Burnout and wellbeing on a short validated scale, surveyed at baseline, midpoint and end. 9) Actual hours worked, including day-off activity in aggregate. 10) Sickness absence and resignations. 11) Work intensity — can staff complete work within hours, at a sustainable pace? This is the honesty metric; leave it out and the board should ask why.
Finance. 12) Labour cost as a percentage of revenue, including overtime and agency spend. In hard-to-fill roles, agency spend is often where a shorter week pays back first. South Cambridgeshire’s much-discussed council trial and the care-sector experience both turned on the cost of covering vacancies, not on salary lines.
Baselines, subgroups and decision rules
Collect four to eight weeks of baseline data before launch, and the same period last year if your business is seasonal. Then cut every people metric by team, role type and caring status where you lawfully can. The average is the least interesting number on the page; pilots fail subgroups first.
Finally, write the decision rules while everyone is still calm. For example: continue if the customer floor held, quality did not deteriorate, actual hours genuinely fell and wellbeing improved; extend if results are mixed and the fixes are obvious; stop if customer or quality floors breach for two consecutive periods after a remediation attempt. Publish those rules to staff. A pilot whose criteria are public can be extended or ended without breaking trust — the evidence was always going to decide.
Practical takeaways
- Agree twelve numbers — outputs, quality, customer, people, finance — before launch, not after.
- Report output per hour worked as well as in total, so redesign is not confused with speed-up.
- Set a customer floor and an intensity tripwire that trigger redesign, not quiet endurance.
- Baseline for four to eight weeks, and cut people metrics by team and caring status.
- Publish the continue / extend / stop rules to staff at the start; let evidence carry the decision.
Sources
- Autonomy, University of Cambridge and Boston College, UK four-day week pilot results, reported February 2023 — https://www.sciencedaily.com/releases/2023/02/230221113132.htm
- Autonomy, UK pilot one-year follow-up, reported February 2024 — https://www.hrdconnect.com/2024/02/26/one-year-on-uks-4-day-work-week-trial-has-been-a-success/
- 4 Day Week Global, South Africa pilot results, reported December 2023 — https://allwork.space/2023/12/south-africas-4-day-work-week-pilot-a-trailblazing-success-for-work-life-balance/
Day5Group is a consultancy that assists in the transition to a 4 day work week. More about Day5Group · All blog articles