@bbbhltz @hacks4pancakes When people join my team, I tell them to go and look at productivity studies. Across different industries (I originally thought this was solely for knowledge workers, but I recently chatted to a researcher who has reproduced the same result in the construction industry), they all show roughly the same shape:
Productivity increases up to 20 hours a week.
It then plateaus up to 40.
It then starts to decrease and is typically negative by about 60.
This is the total net productivity, not the delta. If you are working 60 hour weeks, you would probably be more productive if you just stayed in bed all day.
For programmers, just think about how long it takes to fix a bug that you introduced when you were tired. Fixing mistakes (in any field) is often slow and expensive. Reducing the likelihood of making mistakes is usually much cheaper,
This is for sustained periods. People can often be productive for a 60-hour week if they are well rested, so if you have a one-off urgent deadline, it *may*be okay to work longer hours to meet it, as long as you take enough time off to recover. Averaged out (factoring in the recovery time), this tends to be less productive overall (ignoring the secondary impacts on people who have other commitments, like to see their families, and so on), so it’s generally a bad idea.
I want the most productive 20 hours of each employee each week. I don’t care when they happen (I’ve worked with some people who find they are most productive 2-4am, and that’s fine). Employees are responsible for getting enough rest to make sure that they can be productive for 20 hours each week.
I wrote our vacation policy to be explicit about the point of leave. It is not a gift from the company. It is not a reward for good behaviour. It is an obligation from the employees to the company to ensure that their brains are taken care of so that they can be productive. My contract (which is the model for new employees) has a minimum amount of leave I must take each year and a maximum time I can go without taking at least two days of leave.
The book I most recommend to new managers is PeopleWare and the most important point in that book is that, as a manager, it is not your job to make people work. Most people take pride in their work and want to do it well. Your job is to remove obstacles that stop them from being able to do good work. I don’t think it goes quite far enough because sometimes the biggest obstacle is the employee. If you’re hiring smart and motivated people, the most likely failure mode is that they work too hard and don’t notice their productivity dropping off. Sometimes you have to force them to take a week off (and you need a leave policy that supports you in doing so).
Sorry for the long rant, I haven’t had coffee yet and bad management annoys me, even when it’s depressingly accurate satire.
I am astonished by how much people miss this. If I get a reasonable volume of assigned work done with high quality, why does it matter whether it took 15 hours or 55?
But so many managers care about seeing employees try to be productive consistently throughout X hours, and care nothing about specific productivity.
@david_chisnall @bbbhltz @hacks4pancakes
The major problem, of course is that administration and management will never believe it. They cannot believe it, because it would imply that only field experts can judge productivity, not time sheets or metrics. It would mean that people are not commodities and that management cannot treat them on their terms, but actually have to judge the product.
I have my doubts it can happen at any institution large enough to have a management caste.
@joncruz @bbbhltz @hacks4pancakes One of the most depressing experiences in my time at Microsoft was in the compulsory training for new managers (people who hadn’t been managers at MS before) in 2020. The person running the course asked a question and then proudly told us that he’d have given a different answer when he taught management 20 years ago, to show us how up-to-date he was with new management ideas. The new correct answer was the one that management theory books published in the ‘80s recommended. The worst thing was that a lot of these were written by people who were contrasting Microsoft’s new management style to IBM’s old one.
The rest of the course was about as good as you’d imagine. The first exercise was for everyone to pick a Teams background that best represented their ‘authentic self’ and explain why. This was done with in a group with no prior connection and where the convenor did nothing to make it a safe space for people to express themselves. It was quite telling that the only people who put up family photos were people in heteronormative relationships. This could have been an interesting teachable moment, but the person running the class was completely oblivious to the fact that it had happened.
It was a good set of case studies in what not to do.
This was the course that was made compulsory because Satya decided he cared about good management. Now he says he cares about security. I expect that to go exactly as well.

@david_chisnall @bbbhltz @hacks4pancakes
60 hour weeks being net negative seems a bit suspect, any sources for that?
https://cs.stanford.edu/people/eroberts/cs181/projects/crunchmode/econ-hours-productivity.html and https://www.innovativehumancapital.com/post/what-research-says-about-the-dangers-of-long-working-hours say 60 hour weeks can be less productive than 40 hour weeks (absolute output).
https://www.andrewjensen.net/should-you-work-more-than-40-hours-a-week/ talks about 80 hour weeks snapping back to 40 hours quickly.
Those aren't actual studies, but still it being net negative sounds... incredible :)
@david_chisnall @bbbhltz @hacks4pancakes Just to be clear: you're saying overall productivity per, say, week is about the same for any number of nominal work hours per week between 20 and 40?
Not disputing, just making sure I'm reading you correctly.
@david_chisnall @hacks4pancakes 1. This is total per-week productivity, not per-hour productivity, right?
2. What do you think about people who do multiple jobs? If you ask them to work 20hrs a week as you think is optimal, do you think you should be able to legally prevent them from working at any other job at the same time? Doing two 20hr/week jobs at the same time is quite realistic, and I suspect many people would try this if 20hr was commonplace. I suspect it's more economical to hire for 40 hours and stomach the loss than to hire for 20 and have them work another 20 for a different company, at the same low productivity for both.
When one of your 20 hr/week jobs is being a cashier at a convenience store, you're just being a warm body/deterrent/shelf-restocker. Basically a sneakerbot. No productivity needed