The Five-day Week Is the Most Expensive Habit in Knowledge Work

I have not had a weekend, in the traditional sense, for a few years now. I currently work around 60 hours a week, sometimes more. More than I ever did running my digital agency, and I am less tired than I was at 40. The week has become flatter. Some Tuesdays I am not at the laptop until late afternoon. Some Saturdays I am writing before breakfast and have done the most useful work of the week by lunch.

The change was not deliberate. It happened gradually after I stopped running my Amsterdam-based agency, moved to México and started working on my own things from Mérida. With no team to coordinate with and no clients dictating my calendar, I started noticing when my thinking was actually sharpest. The pattern was not Monday-to-Friday. It never had been. The Monday-to-Friday rhythm had been imposed on a pattern that was actually closer to "two great mornings, one terrible afternoon, one extraordinary stretch on a Sunday when nobody was emailing".

When I stopped enforcing the calendar, the pattern revealed itself.

I am writing about this because I think the Monday-to-Friday work-week is the most expensive habit in knowledge work, and most employers have no idea what it is costing them. It is also the most easily changed. The change does not require new technology, new policy, or new investment. Besides trust it requires the willingness to stop assuming the calendar is doing useful work.

An industrial settlement

The Monday-to-Friday block is not a natural rhythm. It is an industrial settlement.

Henry Ford adopted the five-day week at the Ford Motor Company in 1926, having concluded that assembly-line workers produced more usefully with two days of rest than with one. The U.S. Fair Labor Standards Act codified the 40-hour standard in 1938. The shape spread from factories to offices, and from offices to global default, in the decades that followed. By the time most readers of this article entered the workforce, the five-day week was so naturalized that questioning it felt eccentric.

The work the five-day week was designed for, physical, repetitive, location-bound, requiring synchronized shift coverage, is not the work most knowledge workers do. The output of an engineer, a designer, a writer, a strategist, a researcher is not produced evenly across an 8-hour block. Some hours produce real thinking. Some hours produce performance of thinking. The Monday-to-Friday container treats both kinds of hours as equivalent. They are not equivalent, and pretending they are is costing companies more than they realize.

I know this from both sides. I am a developer by trade. I ran a digital agency with around thirty designers and developers for twelve years.

One of our best developers used to send me Slack messages around 1am. The messages were always sharp - a clean solution to a problem we had been stuck on during the day, a small improvement nobody had asked for, an architectural question that we would have spent a whole meeting on if he had raised it during office hours. I knew, watching this happen, that he was doing his best work between 10pm and 2am (funny enough, the same period I am doing my best work currently too). I also knew that I expected him in the standup at 9:30 the next morning, drained and half-present. The agency was paying for his groggy mornings and getting his sharp nights for free, because the calendar did not have a category for "the hours when he is actually useful".

I enforced the 9-to-6 anyway, because that was what I knew. Most of that structure existed because it was what we knew, not because it produced the best work. The work-week ran the agency, in some quiet way, more than the agency ran the work-week. That is the thing that surprises me most in retrospect. The structure I was defending was not protecting anything. It was just the thing we were used to.

Who decides when the hours happen

The question employers should be asking is not how many days knowledge workers should work. It is who should decide when those hours happen.

Right now, the default answer is the employer. The employer decides that work happens Monday to Friday between roughly 9 and 6. The worker fits themselves into that container. If their best thinking happens at 7am or 11pm or on a Sunday morning when the house is quiet, that thinking has nowhere to go. They will be in the wrong room at the right hour and the right room at the wrong hour. The company will quietly receive a worse version of their work without ever knowing.

This is a real cost. It does not show up in any dashboard. It shows up as work that is technically completed but not actually as good as it could have been, week after week, year after year, in the accumulated output of every knowledge worker whose actual cognitive rhythm does not match the calendar imposed on them.

The alternative is not less work. It is the same total hours, with placement decided by the worker. Coordination windows protected by the team. Norms that say Saturday morning is a legitimate working hour if you want it to be, and Tuesday morning is a legitimate non-working one if you do not. The company keeps its output. The worker gets the thing money cannot buy: the feeling of being trusted to manage their own time. This should be available to workers who want it, not imposed on workers who do not - the entire point is choice.

What the evidence actually supports

The argument I am making here rests on two layers of evidence, and it is worth being explicit about which one carries which weight.

My personal experience and the observable practice of distributed-first companies are the bridge to the structural claim. I have lived this for the last years. Companies like GitLab and Automattic have lived it for over a decade - GitLab employs more than 1,500 people across roughly 60 countries with no offices and no fixed work hours, and Automattic operates similarly with around 2,000 employees. Both publish handbooks documenting how the coordination actually works. Neither has collapsed. Both are still hiring. The case for worker-chosen scheduling at scale is built mostly on what these companies have done, not on controlled studies.

Underneath that, the research on time-of-day cognitive performance is robust. Decades of chronotype and circadian studies show that people perform better on analytical work, writing, and decision-making during their personal peak windows, which vary considerably across individuals. Morning types do their best thinking before noon. Evening types do theirs late at night. The fixed 9-to-6 forces both kinds of workers to produce during hours that match neither group particularly well. This part is grounded in research that has been replicated many times.

The 4-day-week experiments - Iceland, the UK pilot, Microsoft Japan - sit somewhere in between. They show that reducing total hours does not reduce output. Productivity held or rose. Burnout dropped. 92% of UK pilot companies kept the model. But those experiments compressed work into fewer days. They did not test what happens when individual workers freely choose their own hours and days. That experiment has mostly not been run at scale outside the distributed-first companies above.

The chronotype evidence supports the case for control over hours. My experience and the observable practice of async-first companies extend that case to control over days. The extension is an inference, not a proof. I think the inference holds. I want to be honest that it is an inference.

The coordination problem

The honest counterargument is about coordination, and it is harder than I initially want to admit.

Knowledge work is not entirely solo. Teams need windows where most people are reachable for the genuinely synchronous parts of the job - the decision conversations, the design reviews, the problems that cannot be resolved asynchronously. If everyone chooses their own hours, those windows shrink, and the work that needs real-time collaboration gets harder.

Coordination under flexible scheduling is harder, not just different. It requires more discipline around documentation, more rigor about decision-recording, more investment in async tooling than most companies currently have. The companies that do this well - GitLab, Automattic, Basecamp, Doist - have spent years building that infrastructure. You cannot announce flexible scheduling on a Monday morning and have it working by Wednesday. The structure has to be built.

The trade-off is between two kinds of cost. Rigid scheduling costs you your best thinking. Flexible scheduling costs you the synchronous tooling and discipline most companies haven't built. The right answer for any specific company depends on which cost is bigger, and most companies have never actually compared the two. They assume the first cost is invisible and the second is prohibitive. In my experience, the opposite is closer to true.

The middle that actually works tends to look like this: a protected core window of three or four hours per day when most people are reachable, with the rest of the week worker-chosen. That gives the team real-time collaboration when it is needed and real focus when it is not. It is not the simplest structure to maintain. It is the one that produces the best work.

The caregiving objection

Another standard objection is that flexible scheduling will harm workers with caregiving responsibilities, because caregiving demand is highest in the evenings and on weekends.

I think this is mostly wrong, and the wrongness matters.

Caregivers currently have the least flexible lives in the workforce. They are the ones who have to negotiate doctor appointments around the Monday-to-Friday calendar, who lose half a day to a school event that should have been a quick stop, who pretend not to need anything because they cannot admit that the standard work shape does not fit. The five-day, fixed-hours model was not built for them. It was built for an industrial worker who had a wife at home managing everything else.

Flexible scheduling, done well, gives caregivers their lives back. The parent who can take Monday afternoon off to watch their kid perform, work Saturday morning instead, and not feel anxious about either choice, has more functional time and more functional rest than the parent locked into a calendar that pretends caregiving does not happen during business hours. The worry is real that "flexibility" can become a hidden expectation of always-on availability. The fix is structural - protected off-hours, genuine choice - not a return to the old default.

The default was not protecting caregivers. It was making them invisible.

The always-on trap

A third objection worth engaging with is the one about always-on culture.

Microsoft's recent Work Trend Index describes an "infinite workday" pattern emerging in companies that moved to remote work without restructuring around it. The 9-to-5 has dissolved, but instead of being replaced by deliberate worker-chosen rhythms, it has been replaced by ambient availability. People answer emails at 9pm. They check Slack on Sunday afternoons. The boundary between work and rest did not move. It was removed.

This is a real risk. If "flexibility" means "we expect you to be reachable whenever you are awake", the new model is worse than the old one. The old model at least had walls.

Some countries have already started to legislate around this. France introduced the droit à la déconnexion - the right to disconnect - into its Labour Code in 2016. The law does not make it illegal for employers to send messages outside work hours, but it gives employees the legal right to ignore them without being sanctioned. Companies above a certain size are required to negotiate explicit policies on when staff are expected to be reachable. Portugal, Belgium, Ireland, and Australia have followed with their own versions. The Netherlands is debating one.

What these laws recognize is that the answer to the infinite workday is not to retreat to the old structure. It is to put new structure in place. Hours that are genuinely off, even when the worker chooses non-standard ones. Coordination windows the team agrees to and protects. Off-hours that mean something again.

The infinite workday is what happens when employers abandon the old structure without building a new one. The flexible week is what happens when they build the new one deliberately. These are not the same thing, and the companies that conflate them are going to end up with the worst version of both.

Who this applies to

I want to be honest about who this argument applies to.

It applies most cleanly to experienced knowledge workers who have built the self-discipline to organize their own time. Senior engineers, designers, writers, strategists, founders. People who have spent years figuring out how their own attention works and can be trusted to use that knowledge productively. For them, the calendar is mostly cost and not much benefit.

It applies less cleanly to workers earlier in their careers. The 23-year-old junior developer in their first job is often still learning how to focus, how to manage their own time, how to ship reliably. They benefit from external structure in a way more experienced workers do not. Throwing them into full flexibility without scaffolding can produce real problems - missed deadlines, ambient anxiety about whether they're doing enough, loss of the mentorship that comes from being in shared hours with senior colleagues.

The answer is not to keep the rigid calendar for everyone because some people need it. The answer is to make flexibility available to those who can use it, while maintaining more structure for those who are still developing the muscles for self-direction. Different roles, different career stages, different needs. The default should not be "everyone works 9 to 6 because some people would struggle without it". It should be "everyone gets the structure that fits their work and their stage".

This argument is about cognitive knowledge work. It does not apply to shift work, customer-facing service, or any role where presence at specific hours is the job. The whole point is choice. Choice means the worker who prefers Monday-to-Friday gets that too, without penalty.

Working more, feeling better

The deepest thing I have noticed since I stopped enforcing the calendar is that I work more, and feel better about it.

The friction in knowledge work is rarely the work itself. The friction is the requirement to do the work now, on someone else's schedule, whether you are ready or not. Remove that friction, and the work becomes something you do rather than something you are protecting your life from. The total hours often go up, not down. The hours just feel different.

The same logic works for any employee. Someone who knows their employer is fine with them taking a Monday afternoon to watch their kid perform, or skipping a Thursday morning because they are not feeling sharp, does not work less as a result. They usually work more. The anxiety about whether they are allowed to step away is what was costing them, not the absence of the time off itself. Take the anxiety away, and the work flows back in willingly, often at hours nobody required.

The hours you spend working are hours you chose to spend that way. The hours you spend not working are hours you chose not to. Both feel different. Both are better. The worker who feels both will, in aggregate, produce more useful work than the worker who is producing on schedule but quietly resentful about it.

There is also a layer to this argument that nobody is measuring yet. When work hours stop being concentrated into the same five-day rectangle for everyone, traffic distributes. Gyms are less crowded at 6pm. Electricity demand flattens across the week. Restaurants and cafes get patronage at different hours. The peak-load pattern of modern urban life is largely a side effect of synchronized work hours. If knowledge workers stopped synchronizing, much of the rest of society would breathe more easily without anyone planning for it.

This is a smaller argument than the productivity one. But it adds up. The five-day week is not just costing employers their best work. It is shaping cities, infrastructure, and daily life in ways that mostly serve a calendar that does not need to exist in this form anymore.

What I will do differently

When I hire next, I will do this differently. Worker-chosen placement, protected coordination windows, off-hours that genuinely mean off, and enough scaffolding for the people earlier in their careers to learn the discipline self-direction requires. Not as an experiment. As the way the company is built.

Within ten years, mandatory Monday-to-Friday will be a recruiting disadvantage in knowledge work, the way mandatory office attendance has become in the last three. The companies that get there first will be the ones who understood what the research has been saying for years and acted on it before they had to. The companies that wait will spend the same period writing exit interviews and wondering what happened to their best people.

The next phase of the work-from-home shift is not where people work. It is when.

The companies still treating that as a question with a fixed answer are losing their best people right now. They might not have noticed yet. Their best people have.