> My experience working 80-90 hr weeks was that as long as I had flexibility to vary the tasks I didn't sense any drop in productivity. In many cases the fewer context switches increased the amount of "flow", and projects went more quickly than they would have.
Did your experience rigorously measure how many mistakes (bugs, if you're a developer) you made during hours over 56 relative to baseline, and how much it cost someone else down the road in cleanup and opportunity costs to fix them?
With programmers, at least, they can keep writing code well into 100 hours a week and beyond, but the trustworthiness of that code falls off, and that's a problem. It's a well-known result that fixing bugs costs many times more than not making them to begin with.
I saw no difference in quality of work, nor did I get reports on any. I believe the key reason is the ability to change tasks. Been writing code for 4 hours and tired of it? Switch to documentation, regression testing, managerial tasks, ticket triage, spec writing, or something else from your list instead.
Want to go see a movie in the middle of the day, take a nap, or take a two hour lunch to regroup? No one cares.
For startups especially it's do-or-die, so this type of pressure will never go away.
Imagine you had two 5-person startups each building a social media management platform to capitalize on a growing market. The space is quickly becoming crowded, and while not winner-take-all, most entrants will fail.
Company A: They work 90 hour weeks, weekends included. The willingly give up their social lives, take no vacations, and work as hard as they can. They'll sleep at the office if they have to.
Company B: They work 40 hour weeks, and take a normal amount of time off. At the end of the day they go home and relax with family/friends.
Either company could succeed in the end, but which has the better chance? Which attracts more VC interest? Who gives a better impression to prospective clients? Who builds a fully functional platform first?
In my experience Company A often trounces Company B.
That's a pretty silly, cynical, and useless way to look at it. The evaluation should be from a technical point of view of code quality: did the code do what it was supposed to do, and do it well? After all, that's what's being debated in this particular thread.
This is independent of the business value of the code, which of course matters in an entirely different way. However, that's not really the responsibility of the engineer so much as it is the business leaders.
Did your experience rigorously measure how many mistakes (bugs, if you're a developer) you made during hours over 56 relative to baseline, and how much it cost someone else down the road in cleanup and opportunity costs to fix them?
With programmers, at least, they can keep writing code well into 100 hours a week and beyond, but the trustworthiness of that code falls off, and that's a problem. It's a well-known result that fixing bugs costs many times more than not making them to begin with.