[megaduck]

I enforce a militant "40 hours only" policy for my employees, and it's incredible for productivity. When you work longer hours defect rates increase, code quality drops, and things are on fire all the time. You also lose really good people due to burnout.

Sustainable pace is super important for creating and maintaining high performance teams. Crassly, it's just a more profitable way of doing business.

I wish more managers would stop buying into the myth of "time in seat == productivity", and look at the real output of their teams. When you actually run the numbers there's a lot of results that run counter to conventional wisdom.

[jcadam]

Pfft. A man-hour of labor is a man-hour of labor. Which is why I fully understand why my employer is requiring me (and the others on my team) to make up the 48 hours of labor I lost when I evacuated due to Hurricane Irma (even though the office was of course closed for much of that time).

I'm spending the extra hours perusing/posting on HN and looking at job listings :|

[panarky]

> Pfft. A man-hour of labor is a man-hour of labor.

Not all hours are created equally.

I do my best work with hyperfocus.

I don't just mean more work, but far higher quality work that's at the edge of my intellectual capabilities.

This is the work that's most rewarding and gives results that I'm proud of.

But I can't get significant time "in the zone" on a 40-hour week.

All the normal but necessary distractions take at least four hours a day, and I need at least an hour or two of "non-zone" work before I can hit my stride and achieve flow.

So I find 60-70 hours a week is essential to doing my best work.

But I can't sustain that level of effort continuously for years.

So I balance that out with a week a month of low-intensity "work" (slacking/procrastinating/socializing) plus at least twelve weeks per year of contiguous time for uninterrupted travel, long-distance trekking, rebuilding relationships damaged by hyperfocus, etc.

[hedora]

I think you'll find that pushing back on the four hours of usual distractions per day will be more productive than working long hours.

Personally, I find the usual distractions more tiring than coding, based on years of experience.

I estimate at least 2 hours of work lost per hour of meeting, once you add in meeting prep, calendar wrangling, walking, after and before meeting back chatter, realizing you have 30 minutes till lunch/commute after/before the meeting, being tired from presenting/listening/arguing, etc.

One strategy is to have meeting-only or meeting-free days, so meetings primarily ruin your productivity for other meetings and not your actual work.

Meeting-only days scale with team size, but can backfire by enabling the total number of hours in meetings to increase.

[watwut]

"I think you'll find that pushing back on the four hours of usual distractions per day will be more productive than working long hours."

Pretty much this. When you are explaing long hours by on the job ineffectively, then you need to deal with on the job ineffectivity - especially if you are in any kind of leadership position. Otherwise you end up rewarding innefective workers and punish workers with better organizational skill.

General feeling that it is ok to work ineffectively because we stay late anyway was one of the things I resented the most in previous job. And people who caused interruptions and wasted everybody time were seen as "hard workers" because they stayed late.

[thecardcheat]

I think @jcadam was being snarky, not serious.

[somethingabout]

How do you get 3 months of vacation?

[billforsternz]

I enjoyed this +1 but sadly and inevitably there are replies saying essentially "no no, all man hours are not equal". Why doesn't the intertube get satire/sarcasm/irony ?

[vacri]

Text is a bad way of conveying tone, though there was the smiley at the end. I interpreted it the same as you, but I can see how a lot of people would miss the tone.

Poe's Law is why the internet doesn't get sarcasm - that without additional information, it's impossible to differentiate between an extreme position and one mocking it.

https://en.wikipedia.org/wiki/Poe%27s_law

[billforsternz]

The comment we're talking about was skillfully constructed, there's only one way of interpreting it. On a phone so quoting from memory, but the second paragraph starting "that's why I am spending the time perusing HN..." is just an elegant way of saying that the first paragraph was tongue in cheek.

[s73ver_]

I really hope the rest of your team is doing so, too. Companies like that don't deserve good employees.

[jcadam]

Oh, yea. Although, this sort of thing is actually fairly common for my industry (defense contracting), so I'm not sure any of us will end up anyplace better.

[anitil]

The funny thing is, if you just eat the cost, any halfway conscientious worker is going to pay it back in kind. Maybe a few extra hours here or there, maybe a little bit more emotional investment in the work.

[TimPC]

Not all hours are created equal. If you're doing simple enough tasks perhaps your output is relatively stable. When I work on hard problems I've had single hours that can outperform a day or two. Getting rid of most/all of those key creative moments by keeping employees in a constant state of fatigue from large hours is bad. Studies show for typical employees after about 6 weeks of 70h work weeks a 70h week is producing less than a 40h week for a fresh non-overworked employee. So pretty quickly hours become half-hours in terms of production.

[jtmarmon]

That's ridiculous

[tablet]

I partially agree with you, but personal productivity varies. So in our company we have a rule "work as much as you like, usually we expect no less than 30 hours, but it is OK to do 50. Anything beyond these ranges is odd".

[dclowd9901]

What would you say if I told you in a given work day, I only work maybe 3-5 hours (including meetings) and still meet all targets?

This has been consistent for every job I've ever had.

[jrs235]

Well many managers and execs might respond with "well then, give 'em more work! We obviously aren't giving them enough to do!"

Salary exempt is usually a losing game for developers. Crappy places will say "we pay you a salary to get the work done" and under staff and/or overwork. If it takes you 60 hours then you have to work 60 hours. But if it only took someone 30 hours they can't leave early because a) deadlines will be tightened to ensure 40+ hours or b) more work and assignments will be given.

[pc86]

If we're assuming salaries are based on a 40-hour work week, neither A nor B is "bad" for the developer - it's exactly what they signed up for.

[kirykl]

Same experience. I've used the extra time to start internal side projects, for which a few I've earned $$ for. Surprisingly no one asks how I have the extra time

[jcadam]

I can definitely believe that. But in my line of work I'm expected to bill at least 8 hrs/day to the client regardless of my personal productivity.

[panarky]

> Pfft. A man-hour of labor is a man-hour of labor

> I'm expected to bill at least 8 hrs/day

I bet you're capable of far deeper, more intellectually stimulating and rewarding work than you're doing now.

[jsudhams]

I used to be someone who worked only 4 hours max in any company and use remaining time to be on top of tech stacks(SysAdmin). But last four years i self trained on stock trading and it close to be becoming my main job in another year or so. Though i now look for trader jobs where i can trade for profit % but the salaries are no where near in my country for traders.

[dfischer]

What was your growth path to excel and mature as a trader?

[Grangar]

I'm stuck in the billable hour purgatory as well. It seems to be a trend that becomes more and more pervasive.

For my next job I'll move to an inhouse position, hopefully things will be better in a setting like that.

[jcadam]

I'd like to think so, but the pay is good and there isn't much else in this area.

[thecardcheat]

The larger the org and the closer to government (unless you ask for more work), the more likely I find this to be the case.

[michael_storm]

Everyone is different, so good managers set targets based on past performance. If you set the bar at what you can accomplish by working 3-5 hours per day, they'll assume it takes you 8 hours per day, and give you work based on that. If that works for you, great.

[xyz451]

I would ask where do you work and what sort of industry are you in?

W/O knowing anything about your situation sounds like some sort of financial services job.

[srtjstjsj]

I'd say you could expand your role and earn more, if you want and can find an opportunity.

[Vektorweg]

Some thinking-intensive work can bring you at the edge in just 3 hours a day. After that, you generate more problems than solutions. Especially mathematicians are prone to this. Similar cases in software development need to be respected.

[srtjstjsj]

3 hours of thinking-intensive work is fine. Then take lunch/walk/shower, and sit down to write documentation, update bug reports, chat with a coworker about your ideas, review someone else's code.....

[tw1010]

Grothendieck worked twelve hours per day every day of the week for a decade.

Source: https://math.stackexchange.com/a/874290/161054

[eecc]

Yeah we should all - at he miminum - outperform a mathematical giant.

[s73ver_]

Good for him. I have no desire to sacrifice all of my free time to make someone else rich.

[naravara]

>I wish more managers would stop buying into the myth of "time in seat == productivity", and look at the real output of their teams.

I think a big part of this is the generally ambiguous nature of knowledge work. If you could quantify work into discrete widgets it’s easy to mark people to output, but on stuff that’s fuzzy and hard to quantify it’s not so easy.

It actually takes a fair bit of work and intimate knowledge of how stuff gets done on the ground to make realistic projections about output or to understand what “real output” even means.

[webnrrd2k]

Just as a first approximation, I think other knowlege workers who have enough money and clout to be in control of their schedule whole work as a good model for optimal number of hours spent working over the long term. For example, look at doctor's who have their own practice. They (hopefully) like their work, and can determine their own schedule. How many hours do they work? How do they structure their time?

[naravara]

>They (hopefully) like their work, and can determine their own schedule.

Ehhhh. They’re not really as “in control” of their schedules as you would think. Their time is being boxed by financial pressures, generally imposed by insurance and Medicaid/Medicare reimbursement rates, and all the administrative requirements of both complying with burdensome regulations surrounding both their medical practice AND running what amounts to a small business.

They don’t even really necessarily like their work either. Running your own medical practice has a really high burnout rate and many of the people who do it now tend to be older and/or not fully dependent on the practice for their income (e.g. independently wealthy, willing to retire but just wants supplemental income, married and can rely on spousal support financially, etc.).

[Joeri]

Doctors are not as in-control as you think. "What am I supposed to do? Turn away patients who need care?" (I've heard this verbatim.)

Basically self-employed people either have too little work or too much, I rarely hear from someone who has the balance right.

[curun1r]

I'm mostly the same, but I'm less militant about it and, instead, try to empower my team to handle it themselves. That said, I explicitly promise that any time outside of the normal 40 hours should be taken as comp time. So if someone spends 3 hours one evening dealing with a production issue or we have an upper-management-imposed push to get a feature out, there should be an equivalent number of hours taken off when the employee would have otherwise worked. And it's not time that can be banked to use to increase a vacation, it should be taken immediately and if I notice that someone isn't taking it, I'll tell them to leave the office. I've found that employees who aren't overworked can pull a 70 hour week and actually be that much more productive, but they can't do it on a regular basis and they need to recover afterwards.

Also, to combat the "time in seat" fallacy, I encourage my team to take walks every couple of hours, even if it's just around the office, though they should get outside if it's not raining. They can do it in groups, pairs or alone, but I've found that developers are more creative and better able to think through consequences/permutations when they don't spend too much time being sedentary. And it has the dual benefit of breaking the notion that sitting at your desk is the modern day timecard. As you've correctly noted, getting stuff done and achieving quality are what we should measure, not ability to spend time typing and looking at a screen.

[fma]

Walks are very productive in exercise and clearing head. I often try to take a walk with some people on my team. We just talk about whatever comes into our head - life, hobbies, etc. Sometimes it's work and we end up brainstorming.

We joke if we get caught walking we tell management that we're in a meeting.

[danjayh]

My employer has traditionally been very good about 40 hour weeks. However, recently everyone on the team has been doing 55-80 hours per week. On about week 4, I pointed out to my manager that the last time we were on an extended stretch of OT, the company put in place a policy to pay us at an hourly rate equal to our annual salary / 2080 for all OT hours worked. He responded "Yeah, but [project x] was on OT for over a year! Besides, OT is expected at most other companies, and none pay for it." (never mind that most at least have bonus programs). It's now been over 3 months of OT for my team, and personally I think there is no end in sight. Of some slight consolation, the less senior team members do get paid for a portion of their OT (usually around 1/2 to 2/3).

Also, our company doesn't do bonuses at all, doesn't offer anything in the form of comp time, and has an unlimited vacation policy that's basically meant 'next to no vacation' for the past half year or so. Some members of my team have hinted to me that they're looking for new jobs, but I don't think management has a clue.

[lightbulbjim]

I would be outta there.

[Aeolun]

I think at the point a company reaches a state like that. Management better expect every single employee to be looking for a way out...

[michaelbuckbee]

The most productive organizations I've seen have been those that decouple "butt in seat" time from output - it's a poor proxy for adding value.

[arjie]

What is your methodology for determining the change in defect rates and code quality?

[megaduck]

We use Pivotal Tracker, and it's pretty easy to see how many bugs are popping up and how often.

Code quality shows up in your velocity over time. My favorite saying right now is "quality is future speed". We also do regular reviews to assess things like cyclomatic complexity and how well SOLID principles are being followed.

[sidlls]

The quantity of bugs popping up and code quality haven't got what I'd call a well-understood relationship. How do you reconcile that?

[arjie]

Interesting. Thank you very much for sharing.

[RonanTheGrey]

You are my new favorite person. I love running into these anecdotes - I keep thinking I'm alone in believing that a business can be run according to this policy.

[dogruck]

How did you determine that 40 hours is optimal? Why not 30? Or 50?

[megaduck]

I'm not convinced it's optimal! However, it's a "known good" number that's easy to convince people to follow along with. Running experiments is hard and expensive, so we just haven't tackled that yet.

My personal suspicion is that optimal is probably in the 35-45 range, but I'd love to have more data.

[otakucode]

It is "known good" specifically for repetitive manufacturing labor. That's what all of our business structures are optimized for. As work changes to be primarily mental, things will change. It's good that you're thinking about this sort of thing, it will give you a significant advantage in the coming years. Lots of research into how capable humans are of extended periods of mental exertion shows that we can be productive far less than with physical exertion. I suppose it's a good thing that thanks to computers, mental exertion ends up producing productivity multipliers rather than just incremental improvements then!

[Aeolun]

My suspicion is it's somewhere in the 30 - 40 range. For me personally my productivity starts declining after 30, but I'm fairly certain some of my coworkers can do it longer.

[Stranger43]

It's what most of the English speaking societies have centred on over time.

Other places say Scandinavia might put it at 37 and some Asian companies might be closer to 45.

But around 40 hours pr week is where most modern societies tend to think work/life balance is reasonable both for the economy and the individual.

The old union slogan used to be 8 hours of work 8 hours of sleep and 8 hours of leisure. And in many ways that the origin of our 40 hours woork week when we only works 5 days a week.

[dogruck]

That's linked to the question of "how many hours can we extract from these human widgets before they revolt?"

It doesn't hold true in times of crisis, such as war. When the US ramped up for WWII, people were happily working around the clock.

And I'm not sure that it applies to the knowledge workers. My own view is that knowledge workers are productive in bursts.

[Aeolun]

People were happily working around the clock, but I'm fairly certain the people breaking Enigma didn't try doing that more than a fair number of hours a day (fuck if I know what it is, but it's not equatable to making bombs or tanks).

[dogruck]

That's a great question. I wonder what the working habits of the code breakers and other knowledge workers were like. Anybody know more details?

[cardiffspaceman]

The US was only in that period from Dec 1941 to Aug 1945, just over 3 1/2 years. Assuming that the build-up years (when "everyone knew" we were going to have a war) were not as intense.

[dogruck]

The work didn't stop when the war ended.

Regardless, my point is that when there is a motivating force, people work hard.

[s73ver_]

Europe isn't going to fall to Hitler if I don't put in extra hours at my job, though.

[RhysU]

Have a source on the "old union slogan"? I would like to gauge vs some other language regarding tripartite division of the 24-hour day.

[sfRattan]

https://en.wikipedia.org/wiki/Eight-hour_day

For something more 'reputable' than wikipedia, check the footnotes of the article, specifically:

http://www.encyclopedia.chicagohistory.org/pages/417.html

[dominotw]

>I enforce a militant "40 hours only" policy for my employees, and it's incredible for productivity.

why not 35, 39 or 42 hours?

[megaduck]

Predictability. We pair program, so we need everybody to be there at the same time.

People sometimes work less, due to doctor appointments, child care, home repairs, and such. However, they never work more.

[randomdata]

How does a total of 40 hours per week affect the predicability of being there at the same time in a way that 35, 39, or 42 hours maximum per week could not?

[danvasquez29]

my teams do the same thing, it works fine as long as there's some degree of predictability in the schedule. My team does a standard 40 hour work week but only 6 hours per day of required "core hours" used for pairing. The other two hours ar e flexible so the folks can adjust their schedules if they want to get in earlier in the morning and leave earlier.

9-4 you'll be in the office and ready to pair, and lunch for everyone is 12-1 regardless of when you got there; but you can come in at 7 and leave at 4 if you want, or get in at 9 and stay til 6 or work at home later if that works better.

you need that strictness of "pairing time starts now" otherwise you have people coming and going and needing breaks at all different times and it gets disruptive.

[randomdata]

But the daily schedule is only dependent on the maximum numbers of hours to the extent that the hours have to fit within that weekly window. If we assume a five day work week, and the maximum weekly total is 35 instead of 40, then you may only have 5 core hours (or 6 core hours and one fewer flexible hour per day), but the predicability remains exactly the same. Changing the total hours per week does not mean having to throw out a schedule entirely.

I am afraid I still don't quite see what impact the total maximum hours for a week has to do with predictability of scheduling. What am I missing?