Skip to main content


Welp, we're trying the four-day workweek. Here's how, why, and how we'll be gauging success.

Published onJul 12, 2021

A few weeks ago, thinking about how to spend as much time as possible engaged in deep work, I wrote:

We still have a long way to go before we can afford to compress our days to only take time for deep work.

Just days later, tons of research discussing pandemic-induced experiments with 4-day workweeks started trickling out. The basic premise of these experiments is that for knowledge work, productivity is highly variable and depends largely on mental state, not number of hours worked. A number of companies’ experiments have found that moving to a four-day workweek (or from ~40 hours/week to ~32) paradoxically increases productivity during time spent working by giving workers extra time to take care of needs outside of work that otherwise intrude on the workday and lead to productivity-hindering stress. This also seems to have the effect of increasing overall happiness and satisfaction because workers have more time to spend on their friends, families, and lives outside of work, reducing external stress while at work.

Looking at others’ experiments, particularly Buffer’s, and feeling some fatigue ourselves, we decided to experiment with making Fridays optional as a special benefit, starting this week and continuing at least through October 1. For us, that means that neither productivity expectations nor compensation will change. But we’re no longer meeting at all on Fridays, and no one is expected to respond to Slack or email. We expect that some folks may use a few hours on Fridays to catch up a bit on work odds and ends, and that’s okay. Over time, we hope the change results in a decrease in average weekly hours worked for full-time employees from ~40 to ~32.

I’m lucky to work at a small, nimble non-profit1 that sits at the intersection of academia, technology, and publishing and has both the ability and the will to try experiments like these. But we didn’t rush into it headlong. In a way, we’ve been preparing for it for years by continually working with our employees to co-create and refine our culture, hiring, benefits, and work expectations as we’ve moved from a loose group inside MIT to a fiscally sponsored non-profit to our own entity.

Before last year’s winter break, we spent 2 days meeting as a team to discuss improving our working environment, which included, among other topics, a long and fruitful discussion about the definition of work itself in a knowledge environment (i.e., is a midday bike-ride work if you’re using it to integrate the morning’s meetings?2). Following those discussions, we’ve worked hard to compress “shallow work”3 into as little time as possible to provide more time for deep work, principally by implementing the brilliant Remeet system.

To make sure the experiment is successful and keep track of any unintended consequences, we’ll be tracking our progress both qualitatively and quantitatively. We’ll be holding regular feedback sessions for the team to weigh in on how it’s going, and we’ve already rolled out an employee satisfaction survey4, which we’ll send around again in the middle, and at the end of, the experiment.

My guess is that, like other companies who have tried the experiment, we’ll see concrete gains in worker satisfaction and happiness without any decreases in productivity. However it turns out, I’m very happy that we’re in a position to try it with employee buy-in and ways to meaningfully measure the results.


No comments here