[sam0x17]

Because burn-out is the main driving factor behind the productivity changes, not the hypotheticals you are talking about. In engineering, people are burned out by friday, and even outside of engineering, people tend to phone it in by that point in the week because they are just done. They are operating at less than 50% mental or physical capacity. If you instead just give them that time to recharge, they'll get that much more done early in the week the following week. They'll also be in a better headspace because although you are only decreasing the days they work by 1/5th, you are increasing their uninterrupted weekend/family time by 1/3rd. In a more egalitarian society, you might imagine that ratio to be more like 1:1 to be fair.

At my previous startup we saw an 8% productivity _increase_ from moving to a 4-day work week, and this was across a variety of metrics like issues closed, lines of code, etc. We don't even know if 4 is optimal, it could be 3. Only way to find out is to try it.

[bumby]

My anecdotal experience doesn't support this. I worked in an org that shifted to a 9/80 schedule with a three day weekend every other week. Prior to this, Fridays always seemed to be the days where people coasted...they would come in kinda late, have a coffee break, then have a breakfast break, then an extended bathroom break, maybe get an hour of stuff done before an extended lunch. And of course, Friday afternoon meetings were frowned upon. After the change, this same behavior shifted to Thursdays, which used to be normally productive day.

Maybe they were burned out from the extra hour of the previous days, but considering the Thursdays of the remaining 5 day weeks didn't display this behavior, I doubt it. Now there was the complication that once management saw a dip in productivity, there was a near mutiny at the mere suggestion of going back to a normal workweek.

The implication is no matter how long/short the work week is, people will adjust to make the end a slack day.

[sam0x17]

Imagine the flipside though. Imagine you work every day of the week. Surely this would have lower productivity than having one day or two days of weekend unless we are to believe weekends don't have a rejuvinating effect. If you have a problem where a segment of your workforce is slacking off on the last day (whatever that may be), that is its own problem that will exist unless you have no days off (which would obviously be pathological and most likely illegal for other reasons). So I find this argument a bit trite, because it either already applies to your situation regardless, or you work at a company where people simply don't slack off the last day. And I do think this whole "people slack off on the last day" thing is very trumped up. I've seen plenty of examples of people finishing what they were expected to get done and tuning out at the end of the week, but I've seen much more scenarios where people have an unending mountain of work and their efficiency goes way down later in the day on thursday and for all of friday. I think this is the norm in most industries, and it can be solved with a 4-day work week.

[bumby]

>Surely this would have lower productivity than having one day or two days of weekend unless we are to believe weekends don't have a rejuvinating effect.

This seems to land in the area of untested hypothesis. Sure, we can speculate on all kinds of mechanisms. I could speculate that by going to 3 day weekends would lead to a bigger "rejuvination effect" and more productivity. But the data gathered from my org showed the opposite. Productivity went down. And now management had a lot of pushback to just get back to their baseline (more productive) schedule.

Now I'm willing to concede that it's going to be different industry to industry. It may very well work well in SWE and I think we see a lot of that bias in tech-centric circles. I also think it’s an error to assume it holds for most industries. My example was from an R&D area and not strictly software engineering. Do you, for example, think your doctors office would have more patient throughput if they went down to 4 eight-hour days? Or manufacturing? I'd be skeptical until I see the data. And I concede that a lot of work cultural differences also matter. And leadership matters. My original post was not claiming some definitive answer, just giving some pause to the sentiment that reduced hours is a generalizable rule to increase productivity*. I think a lot of people let their cognitive biases get the best of them and run with the idea.

* I also think "productivity" is the wrong way to frame the problem. Some things, like work-life balance, are a net good and worth a hit on productivity. The economy should serve society and not the other way around.

[sam0x17]

> I also think "productivity" is the wrong way to frame the problem. Some things, like work-life balance, are a net good and worth a hit on productivity. The economy should serve society and not the other way around.

I mostly agree with what you're saying, especially about going across industries, but I do believe it is quite easy to lower productivity, regardless of the industry, by dramatically imbalancing work-life balance. Put another way, having good work-life balance is a net good, partially _because_ people with good work-life balance are more productive than people who are burned out.

Your medical doctor isn't going to be very good at diagnosing patients when he's in his 140th working hour of the week, if he were to attempt such a thing (with ~3 hours of sleep per night and no time off other than that)

[bumby]

>having good work-life balance is a net good, partially _because_ people with good work-life balance are more productive

This is actually what I'm pushing against. I don't like productivity being a primary principle. I'm saying it's worthwhile irrespective of the impact on productivity. (the use of 'net good' was a bad choice on my part because it implies a balancing act).

It's quite common for healthcare workers (particularly surgeons) to work extremely long hours, but that's being re-thought in some areas (esp. in regards to residents). To your point, though, I think a lot of the rethinking is driven by trying to reduce medical errors. It's more related to quality than productivity, but I don't think they can be easily parsed.

[banannaise]

Given that 32 hours is still far too long for someone to be doing active mental work in a week, what you're likely seeing is not the same number of slack hours, but the remaining slack hours being shifted around

[bumby]

Does this change if it’s not SWE work? The example in my anecdote is from R&D work that involves a lot of wrenching on hardware.