Eighteen months ago, the leadership team of a 45-person SaaS company where I advise asked a question that made everyone in the room lean forward: could we give people Fridays back without shipping less? The engineering leads worried about sprint velocity. The support team worried about ticket queues. Finance worried about paying for five days and receiving four.
I had read the headlines about four-day week trials, but headlines do not survive contact with a real org chart. What convinced me to run the experiment was the pattern underneath the numbers: teams that succeeded did not simply delete a day from the calendar. They redesigned how work moved first, then let the calendar follow.
This guide walks through the exact 90-day pilot framework we used, the async handoff templates that kept coverage intact, and the output metrics that won over our board. If your team is remote-first, you are better positioned than any office-based company to make this work.
Why Remote Teams Have an Edge in the Four-Day Week
When people picture a four-day work week, they usually picture an office experiment. Distributed teams actually hold most of the cards. There is no commute to compress, no meeting rooms to fight over, and no manager walking the floor to see who is at a desk. The work already lives in tools, and async communication is already a survival skill.
The flexible-work data backs this up. Buffer’s 2024 State of Remote Work report found that 98% of remote workers want to keep working remotely at least some of the time for the rest of their careers. Flexibility is the strongest currency an employer has right now, and a four-day week is flexibility in its most concentrated form.
That said, remote teams fail at this too. The pattern I see most often is a team that deletes Friday but keeps every process from the five-day week. By week three, people are answering Slack on their day off just to keep up. The rest day turns into an unpaid guilt day, and the pilot quietly dies.
A four-day week is not a day you delete. It is an operating model you redesign.
What the UK Trial Data Actually Says
The strongest evidence comes from the UK’s largest coordinated four-day week trial, which tracked 61 companies across multiple industries for six months. The results matter for remote teams because most participants were knowledge-work businesses that look a lot like yours.
| Metric | Trial result | What it means for your pilot |
|---|---|---|
| Companies still running the policy after the trial | 90% | The model survives contact with reality |
| Employees reporting lower burnout | 71% | Burnout is a cost you are already paying today |
| Revenue during the trial | Roughly flat, about +1.4% | Output held steady without extra hours |
| Sick days | Down 65% | Health gains show up in attendance first |
| Resignations | Down 57% | Retention is where the ROI compounds |
Microsoft ran a similar experiment even earlier. Its 2019 Work-Life Choice Challenge closed the Japan offices every Friday in August, and the company reported a 39.9% increase in productivity per hour compared with the previous August, alongside lower electricity costs and fewer printed pages.
The lesson from both is consistent. Output held steady not because people worked harder for four days, but because the trials forced companies to cut low-value meetings and busywork first. That is exactly the muscle remote teams need to build anyway.
Choose Your Pilot Model Before You Touch the Calendar
The trial companies used different structures, and the right one depends on how your team faces customers. Pick the model first, because it determines your coverage plan, your metrics, and the risks leadership will ask about.
| Model | How it works | Best for | Main trade-off |
|---|---|---|---|
| Staggered rotation | Teams take different weekdays off so coverage never drops | Support, operations, customer-facing roles | Some cross-team sync still needs the shared day |
| All-hands | Everyone takes the same day off | Product and engineering orgs with a clear weekly rhythm | Higher visible risk on the first attempt |
| Compressed hours | Four longer days at full pay | Teams with deep, schedulable work | Fatigue risk; the day still has a focus limit |
In my experience, product and engineering orgs usually start with the all-hands model because shared context is easier to maintain. Support and operations almost always need the staggered rotation, since someone has to answer the queue every weekday. The compressed model is a reasonable fallback when leadership cannot yet accept a full day off, but treat it as a stepping stone, not the destination.
Whichever model you pick, write the day rule down in one sentence and make it public. Ours read: no internal meetings and no internal messages on Friday, customer escalations excepted. A policy that needs a paragraph to explain will eventually need a meeting to defend, which defeats the purpose.
The 90-Day Pilot Roadmap
Ninety days is the shortest honest window. It is long enough to cross two planning cycles and short enough that leadership will actually approve it. Here is the schedule we used, and the one I now recommend to every team I advise.
| Phase | Weeks | Focus | Exit criteria |
|---|---|---|---|
| Baseline | 1-2 | Measure current output, response times, sick days | Metrics documented and signed off |
| Model and cohort | 3-4 | Pick the pilot model and the teams involved | Coverage plan approved |
| Meeting redesign | 5-6 | Audit every recurring meeting; cut what can go async | Weekly meeting hours down by a third |
| Async handoffs | 7-10 | Roll out handoff notes, decision logs, rotations | The off-day runs with no pings |
| Evaluate and decide | 11-13 | Compare against baseline; present to leadership | Continue, adjust, or stop |
Two details decide whether that schedule holds. First, the meeting redesign must happen before the first day off, not after: if the pilot starts from a full calendar, you are measuring compression, not redesign. Second, announce the evaluation date at kickoff. A pilot without an end date drifts into permanence, and permanent experiments stop being measured.
Coverage Without Burnout: Async Handoffs That Work
Coverage is where most pilots wobble. The fix is boring and effective: a written handoff that travels across time zones and days off. On my team, the handoff note has five fields: current state, blockers, next owner, where the context lives, and when to escalate. Writing it takes ten minutes and saves an hour of Slack archaeology the next morning.
For anything too tangled for text, we record a short screen video instead. A three-minute walkthrough of a staging environment communicates more than a page of prose, and the next person can watch it at double speed on their own morning.
Customer support teams need one more layer: a rotation. Stagger the off-days so the queue always has two people on it, publish the rotation a full quarter ahead, and pair it with an escalation path to a manager who is formally on-call and paid for it. If coverage quietly depends on unpaid heroics, the pilot will collapse in exactly that spot.
The Metrics That Convince Leadership
Leadership does not care how people feel about Fridays until the numbers hold. Capture these five before the pilot starts, then compare at the end. Per-day framing is what makes the argument work: divide weekly output by four instead of five, and the conversation changes immediately.
| Metric | How to measure | Healthy signal at the end of the pilot |
|---|---|---|
| Output per working day | Weekly throughput divided by working days | Flat or rising versus baseline |
| Cycle time | Median time from ticket start to done | Stable, not drifting upward |
| Customer response | First-response time and CSAT | Within one business day of baseline |
| Sick days and burnout | HR dashboard plus a monthly pulse survey | Trending down |
| Retention | Regrettable attrition versus the same period last year | Flat or better |
One caution: do not over-instrument. Five metrics measured consistently beat twenty metrics measured once. In our pilot, output per working day and first-response time carried the entire board conversation; the other three were supporting evidence.
Three Failure Modes That Sink Four-Day Pilots
- Compressing five days into four. If the meeting audit happens after the pilot starts instead of before, people simply work longer Tuesday through Thursday. Burnout rises, and the trial gets blamed for it.
- Skipping the baseline. When nobody documented output before the change, every later debate becomes a matter of opinion. Two weeks of measurement is the cheapest insurance you will ever buy.
- Declaring victory on vibes. A happy pulse survey with flat revenue is a real result. A happy pulse survey with slipping cycle times is a problem wearing a smile.
Start smaller than feels impressive: one cohort, one model, ninety days, five metrics. The UK trial data says output holds, Microsoft’s Japan experiment says focus improves, and our own pilot said the same. The teams that succeed are not the ones with the best intentions. They are the ones that redesign the work before they redesign the calendar. If your real bottleneck is meetings rather than days, run a meeting audit first; and if async writing is the weak link, tighten your update templates before you cut a day.