Why pilots attract attention
A shorter working week is an easy idea to grasp and a hard one to settle by argument. So when organisations run trials, and the trials report positive findings, the results are widely reported. A pilot feels like real-world evidence rather than opinion, and that is partly true: a trial can reveal practical problems and benefits that a thought experiment cannot.
But a pilot is a specific kind of evidence with specific limits. Understanding those limits does not mean dismissing the results. It means knowing which questions they can answer.
What a typical pilot looks like
Designs vary, but many trials share a few features. A group of employers signs up. Over a set period, often several months, they move to a shorter week, commonly while keeping pay unchanged. Researchers or coordinators collect information, which might include employer surveys, employee questionnaires, administrative figures such as revenue or staff turnover, and interviews. A report then summarises what happened.
The definition of “four-day week” itself can differ between schemes. It may mean fewer hours with the same pay, the same hours compressed into fewer days, or something negotiated for each workplace. Results from one version do not automatically carry over to another, so comparing trials requires checking what exactly was tried.
The self-selection problem
The most important limit is who takes part. Employers in pilots have generally chosen to join. Organisations that volunteer may differ from those that do not: they may already be curious about the idea, confident in their management, in sectors where output is easy to reorganise, or sympathetic to the aims of the scheme. This is a straightforward case of selection effects.
If volunteers tend to be well-suited to the change, positive results may say more about the volunteers than about the policy. A fair summary is that pilots often show what can happen when willing, prepared employers try the idea. They do not necessarily show what would happen if it were applied across workplaces that are sceptical, stretched or dependent on constant coverage, such as some services that must remain open.
The same applies within organisations. Staff who are unhappy with a change might leave or give less favourable answers, while those who like it respond enthusiastically. Who completes a survey at the end can differ from who started.
No comparison group
Many pilots lack a control group: similar workplaces that did not change. Without one, it is difficult to separate the effect of the shorter week from everything else happening in the same period. Business conditions, seasonal patterns, new staff, new tools or simply the attention of being part of a high-profile trial could all contribute. Researchers refer to these as confounding variables.
Some evaluations do compare against earlier periods in the same organisation or use matched comparisons, which helps. But a before-and-after comparison within one workplace still cannot rule out other changes in the meantime. Designs closer to natural experiments, where some groups happen to be treated and others not for reasons unrelated to their characteristics, are rarer and tend to be more informative.
Short time horizons and novelty
Most trials are brief compared with the lifetime of a working arrangement. Early enthusiasm may fade, or early difficulties may be resolved. A trial lasting some months says little about how things look after several years. There is also a plain novelty effect: people often behave differently when they know they are being observed and when something new has been introduced.
What gets measured
Reports typically draw on several kinds of measure, each with its own limits.
- Self-reported wellbeing and stress. These rely on how people describe their feelings, and questionnaire answers are shaped by wording and mood. They can also be influenced by wanting the scheme to succeed.
- Productivity and output. Defining output differs hugely by occupation. Where output is hard to count, as in many knowledge roles, firms may rely on managers’ judgement. The broader difficulties of productivity measures apply here.
- Revenue and business performance. These move for many reasons unrelated to the schedule.
- Staff retention and recruitment. Meaningful but slow to observe, and affected by the wider job market.
- Working time actually worked. Staff may report a four-day week while still completing tasks in their own time, so formal schedules and real hours can diverge.
What pilots can reasonably show
Within their limits, pilots can be useful in several ways:
- They can show that a shorter week is feasible for at least some organisations.
- They can surface practical issues, such as cover arrangements, meeting culture and customer expectations.
- They can generate hypotheses about which kinds of workplaces might benefit.
- They can capture how participants experienced the change in their own words.
What they cannot show on their own
- That the results would apply to every sector or size of employer.
- That the policy caused the reported improvements, rather than other factors or the type of volunteer.
- That benefits persist over the long run.
- How much it would cost or save in the economy at large, including effects on other workers and prices.
Such questions call for larger, longer studies with comparison groups, and even then results may differ by context.
A sensible way to read a headline
When a report states that a trial was a success, it is worth asking how participants were recruited, what “success” meant, whether there was a comparison, how long it lasted, and who wrote the report. A scheme run by advocates may emphasise different findings than one run by an independent evaluator, though both may be honest. Treat the results as one useful piece of the picture, which is a fair way to treat most claims about the future of work.