What a four-day-week pilot actually showed

The setup, and the baseline that mattered

Fernline’s leadership had been circling the idea for a year. What finally tipped the decision was baseline data: eight weeks of pre-pilot measurement showing their teams averaged 26.4 focused hours per person per week inside a nominal 40. The rest went to meetings, coordination churn and context-switching residue.

That number reframed the risk in a way no amount of advocacy had managed. They were not being asked to fit 40 hours of work into 32. They were being asked to protect roughly 26 focused hours inside 32, which suddenly sounded plausible rather than heroic, and gave the sceptics on the leadership team something concrete to hold the pilot to.

The pilot ran twelve weeks, Fridays off for everyone, salaries untouched. Three metrics were agreed in advance: focus hours per person, delivery output per team, and out-of-hours activity. Deciding the success criteria before week one turned out to be the single smartest move they made.

Week one was genuinely bad

Nobody puts this bit in the press release. Week one, focus hours dropped 14% below baseline even after adjusting for the shorter week. Everyone was renegotiating everything at once: standups, client calls, handoffs with a US partner who still very much worked Fridays.

The heatmap looked like a construction site. Meetings that had lived on Fridays scattered across the remaining days, colliding with existing focus blocks. Two engineers logged their worst deep-work scores in six months.

By week three the schedule had reorganised itself and the panic emails stopped. If Fernline had run a two-week trial, as one board member had suggested, they would have cancelled the whole thing at the exact moment it was about to start working.

Meetings were the first casualty, deservedly

Total meeting time fell 31% by week six and stayed there. Almost none of it was cancelled by decree. When the calendar shrank by a day, recurring meetings had to justify their slot against actual work, and a remarkable number quietly failed the audition.

The weekly all-hands went fortnightly without anyone appearing to miss it. Three separate status meetings collapsed into one written update that took its author twenty minutes and its readers five. One team discovered its Monday planning session and its Wednesday check-in were the same meeting wearing different names, and quietly buried the Wednesday one.

Average meeting size dropped too, from 6.1 people to 4.8. Scarcity made organisers think about who actually needed to be in the room, a discipline five-day weeks had never demanded of anyone.

Did output hold? Yes, with one asterisk

Across the twelve weeks, focused hours per person settled at 25.1 per week, down just 5% from the five-day baseline of 26.4, achieved in 20% less time. Delivery metrics told the same story: sprint completion held within noise, support resolution times improved slightly, and two product launches shipped on their original dates.

The asterisk is the sales team. Their pipeline activity genuinely suffered, because prospects book calls on Fridays whether or not you work them. By week five, sales had negotiated a rotating schedule, half the team off Friday, half off Monday. Output recovered within a fortnight.

The lesson generalises. A uniform four-day week fits maker teams beautifully and collides with any role whose demand is set by the outside world. Fernline’s pilot survived because they treated that as a scheduling problem rather than a verdict.

The Thursday-night question

The sceptics’ prediction was specific and, to be fair, entirely reasonable: Friday’s work would not disappear, it would migrate to Thursday evenings and a low hum of Sunday dread, leaving people more compressed than before. This is exactly the kind of claim aggregate activity data can test, which is why Fernline wrote it into the metrics before the pilot began.

Mostly, it did not happen. Activity after 7pm on Thursdays rose 9% over baseline in the first month, then faded to 3% by pilot’s end. Weekend logins stayed flat for eleven of thirteen teams. The five-day rhythm, it turns out, was padding as much as producing.

Two teams were the exception, both running month-end financial processes with immovable deadlines. Their Thursday evenings grew and stayed grown. Fernline moved their close calendar rather than pretending the pattern away, which is precisely what the data was for.

What the averages hid

The aggregate story was calm, but the team-level view was not uniform, and the individual experience less so. Parents rated the pilot near-perfect. A handful of early-career staff quietly reported missing the ambient learning of overheard conversations, since the compressed week left less unstructured overlap with seniors.

One engineering lead found the four-day rhythm harder, not easier. Her role was interrupt-driven, and the same volume of interruptions now landed in fewer hours. Her focus score dropped 18% while her team’s rose. The averages would have buried her entirely.

This is our standing argument for reading distributions, not means. A pilot that is 80% wonderful and 20% quietly corrosive needs to know which fifth is hurting before the policy hardens.

What Fernline kept, and what we took away

At the end of week twelve, Fernline made the four-day week permanent, with the sales rotation and the shifted close calendar written into the policy. Twelve months on, regretted attrition is down 38% year on year and their offer-acceptance rate has become a recruiting talking point.

Our takeaway is narrower than the headlines want it to be. The four-day week did not manufacture productivity. It forced an audit of the five-day week, and the audit is what paid. A company could, in principle, run the audit without touching the calendar. Few will, because nothing concentrates minds like a deadline that arrives every Thursday.

If you run your own pilot, measure a baseline first, commit to twelve weeks, and agree what success means before week one. The companies that skip those steps do not end up with a failed experiment. They end up with an argument.

Writes for The Signal about culture and the future of measurable, humane work — drawing on anonymised patterns from the teams and focus hours analysed on Momentum.