Why Crunch Mode Doesn't Work (2005) | Hacker News Reader

Why Crunch Mode Doesn't Work (2005)

legacy.igda.org

The critiques I've seen of crunch mode always seem to assume that it's a normal state of affairs, and thus leads inexorably to burnout. We'll have maybe one or two weeks of crunch every quarter or so when we have a major release that everyone needs to coordinate on and no one wants to be blamed for holding up. For those one or two weeks out of twelve, we'll work sixty or maybe seventy hours. It's absolutely effective, and while I've felt tired at the end of it I've never felt burned out. It's also offset by having very relaxed working hours the rest of the time.

I get that some firms see the productivity boost that happens during crunch mode and try to make that the new normal, but is this even that common? Do most firms not do as mine does and just crunch when you need to and have normal hours the rest of the time?

I used to work for a major game company on a major title that had a new release each year around the same time of year. Every year, for approximately 3-4 months prior the release of the game, the team would work 10-12 hours per day, 6-7 days per week. As far as I could tell, no one else considered this unusual.

I, too, worked for EA on sports titles.

Superman Returns was worse: 9 months of 60+ hour, 6 days a week crunch.

Did you see a super huge payback on those hours or was it just life force draining from your shell?

Well, the game tanked, we didn't get huge bonuses and next to no comp time so... mostly just drained life force.

It was a technically challenging project, though, and I learned a lot on it.

is this even that common?

Common enough for people to talk about it. In some industries (especially games) it's become the norm. It requires mature management-employee relations and not short term profit maximisation.

It is quite common in games, and this article is from the perspective of games. Crunch for one or two weeks certainly can be effective, even if a little tiring, but in the games industry a crunch of three months is in many places a short, light crunch. It could be hard to understand a critique on crunch if you haven't experienced a bad one, but really try to imagine eighty hours a week for six straight months. That might not be common in general, but it is happening in games companies, and I've been through more than one. After you've done that a few times, honestly, its hard to hear someone talk about a one or two week "crunch" and take it seriously. I don't mean that to be demeaning or belittling, my point is that it can be hard from any side to see the other side's point of view. Crunch times in games, among other industries, are sometimes really, really awful and take away very significant amounts of time from people's lives and families.

No offense, but if your company's crunch time is so predictable, there must be a mistake in your scheduling. I've worked at companies where there is literally no crunch time, ever.

Worth reiterating, I've seen good people run themselves into the ground with an endless 'crunch', because they felt it was a valid strategy. 'Running themselves into the ground' might seem wishy-washy, but to make it more concrete: people have _had to_ take time off work days before a big deadline because they couldn't work. I'd rather the marathon than the brick-wall crunch.

I wonder if the science on the 40-hour work week applies mostly to specialized jobs, or if it can be broadly applied to more generalized roles, i.e. _entrepreneur_, where one switches contexts several times per day/week. How many 'successful' startup founders only worked about 40 hours a week versus 60+? Is Elon deluding himself when he claims his 100+ hour weeks allow him to accomplish more than his less-disciplined competitors?

If the daily work is built around being in the flow and rhythm then mental muscle is not that much stressed.

You wake up, you know that your limo is waiting in 20 minutes. You'll start with a scheduled call on the way to gym. At the gym personal trainer makes sure you're giving maximum. Then the limo takes you to the private jet so you can catch the conference in Europe at 4pm. etc.

My point is, if the world around you revolves in a rhythm where falling out is not an option, you can find the balanced flow and rhythm so 100+ hour week full of balanced work is absolutely normal. You just stop wasting non-work time over the weekend and use it for maximum recovery and family-time.

Add a bit of humour and laughter to these 100+ hours and stop caring about issues you can't fix and it's a paradise to live in.

Huh?

If you have the insane financial position that someone like Musk does, it gives you the ability to edit a lot of annoyance/stress out of your life. Most of the day to day crap that most of us deal with can be automated or delegated away.

I suspect sleep, and routine, is more the determining factor. I imagine most people should be capable of maintaining high levels of productivity for 77+ hour weeks if they could get 9+ hours of solid sleep every night. That leaves enough time for an 11 work day, 1 one hour workout, 30 minutes of commute, and 2.5 hours of misc time. As a single persons game.

However, that's a big SHOULD/COULD/WOULD. Real world statistics probably reflect that most people either aren't capable or simply wouldn't maintain this kind of routine.

Elon probably has built up a huge support system around him that he knows, both implicitly and explicitly, has his back and will help him out.

As opposed to "your project is behind, time for 11½ hour work days." I got that earlier this year. At least I was paid hourly at the time.

Experiments have repeatedly shown a solid increase in day-total productivity when going from 10 hour workdays to 8 hour workdays; similarly, there is an increase in total weekly productivity when comparing 6 days a week to 5 day workweek.

I doubt that the people who work[ed] on saturdays had a much different daily routine than those who didn't; but still, working that extra day only results in less work done.

No, no less work done in total.

You work that extra hour/day, week after week, and you see real productivity gains (not necessarily per-hour productivity gains, but marginally speaking, you make progress faster). As others have mentioned before, this trick can work very well in the short term.

However three months into it (roughly +150 man-hrs), you make a small blunder that nobody notices at first. One year later, it will blow up and set your whole team back for a whole month (assuming team_size = 5, -800 man-hrs). Everybody will try to catch up by staying an hour later, or working on the weekends... rinse and repeat.

The problem is that cause and effect are so far apart in time that nobody notices why this is happening. Usual suspects such as "rusty codebase" or "technical debt" get assigned with the blame and no lesson is learned.

As far as I understand, the experiments show explicitly less work done, as in, if you do the 60 hour workweek for prolonged time, then you have less widgets produced per week, period.

It's NOT a tradeoff of more volume for less quality and increased risk. All the drawbacks you list are valid, and come on top of less total productivity.

You don't get a product built faster but with more defects and technical debt. You get the product built slower; with more defects and techical debt; and at a cost to the workers personal life in a true lose-lose tradeoff.

Well, it 48 hours a week vs 60 jours is the secret of the success of british & USA vs Germany during WWI :

* WWI = death by explosive * over work == more defaults => more of your soldiers killed less of theirs, more money spent * so british army edicted a ban on overwork ~ 1917 followed by USA.

The question is does it matter more to produce the most defective products, or to produce stuff that works?

I can't comment on context switching, but one thing to keep in mind is that 40 hours is the peak of a bell curve. That implies that some individuals can in fact work 50, 60-hour weeks for months without a significant drop in productivity, and some probably shouldn't work more than 4 days a week. Maybe Elon Musk is the one-in-a-billion guy that can actually stay productive for 100+ hours, but thinking that's reason for you working 50+ hours is like saying you should be able to swim 100m in 70 seconds because they do it in less than 50 in the Olympics.

Or, to add in the more specifically measured element:

40 hours is the peak of the Laffer curve [productivity being inversely proportional to hours worked a week]. I have absolutely no data on variance for the peak of the Laffer curve (if you imagine attempting to gather this data, you'll notice it is extremely difficult). That being said, despite this lack of data, I have heard a lot of people claiming they can work a whole lot more and get more and more done. I do not believe their ad hoc methods have secretly outsmarted science. It seems more likely that they are mistaken (if only for the reason that it is very hard to measure and they universally lack specific reason to choose "my peak productivity happens at >60 hours a week" over "the volume of my work is more noticeable than the quality of my work", which is essentially universal -- but lines of code, as a metric, don't pay the bills).

Correct me if I'm wrong, but I believe the research actually only shows the peak is no higher than 40 hours per week. Are there studies indicating people are more productive working 40-hour weeks than 30 or 35-hour weeks?

Of course your point still stands, they'd just have to be even more of an outlier.

It really depends what you're doing. For me, meetings don't result in nearly the same mental fatigue as coding does.

and just the opposite for me! One meeting can depress me for half a day. But I can get in a coding thrall and go 4 or 5 hours without noticing time pass.

I hear you. I hate having scheduled meeting almost every day, early, that ramble about things that usually don't affect me. My time would have been better spent still in bed until it was time to start building (coding, designing, writing, etc)

Entrepreneurs are, by definition, outliers. I don't think it makes much sense to compare them to -- or include data for them in -- an analysis of the average worker.

That's why other industries don't do "crunch mode." Can you imagine:

"We're really behind on building this bridge" "CRUNCH MODE!"

The usual solution in bridge building is: "MORE PEOPLE!" ... but that doesn't work in software (because of training time, less productivity in the meantime and so on, you know the drill), so we don't have that option. The best idea would be to accept that it will take longer, but that fails flat for political reasons. Bad situation.

I got the impression, that most people are aware, that there is only a finished software or a date, even if they are non-technical.

They got something, even if it's bug-ridden, which they can show and sell someone. And the fixes can be made between the "finish date" and the "first real use". The software that is ready after the finish date just has to work for presentations.

They do. Note the 2nd and 3rd bullets on this announcement: http://www.spanishfork.org/newsevents/events/view.php?id=520

other industries do crunch mode, just in different ways.

tax accountants crunch 3 months a year for 12+ hours a day. it's as predictable as clockwork.

construction crunches by adding 2nd and 3rd shifts to a project to get it done quicker.

investment bankers crunch their entire careers, basically.

political teams crunch during elections.

media crunches during newsworthy events or to wrap a project.

we're not that unique. "the grass is always greener..."

The finance industry, particularly around the front desk, is notorious for freakishly long working hours. I don't work in that industry, but friends that do speak of 100+ hour work weeks, with 80 hours considered normal in some workplaces.

As outlined in the post, that way of working, doesn't (shouldn't) work. Especially when the code produced, or in the case of traders decisions made, might lose vast amounts of money due to a single mistake.

I'm curious, if anyone here works in that kind of an environment - How does this work? Are these numbers exaggerated? Are you all on stimulants? Do you see the kind of creeping errors and codebase decay one might expect?

AFAIK it's more about signalling - demonstrating commitment to your peers and superiors.

Not in that industry, but working at a startup founded by two ex banking (albeit software) guys. A little bit of the culture has come along for the ride.

Absolutely - finance uses hours the same way software uses open-source projects created in your spare time. "Passion", "culture fit", etc

People I knew that worked that much, used to spend a lot of time blankly staring at screen or chatting about weather after few weeks. The rest and socialization they were missing out of work somehow creeped into work. I do not think they were aware of the effect, it is just that everything took them somewhat longer time.

It may seen silly but the article could use a brief description of what's "crunch mode" since it was a new term to me.

Google: define:crunch mode > "Crunch mode", also referred to as "crunch time," is the term used by those in the software development industry to describe working extra hours for extended periods of time in order to finish a project or meet a deadline.

New title - "Why CONTINUOUS crunch mode doesn't work."

This feels a bit link-baity, because it says nothing of how short, uncommonly used crunch modes help or hurt productivity - just how super-long work weeks are eventually more detrimental than helpful.

Edit: I would posit that short bursts of overtime - perhaps a single 60-80 hour week at the ramp up to a major release can actually be helpful if not exciting - if used quite sparingly. Research on that theory would be more interesting to me.

Hopefully, such a study would include a study on how often the promise of only one week of overtime turns into multi-week slogs.

Short bursts of overtime is not crunch. Crunch is by definition long term.

Anyway, I recall that you can actually raise short term productivity for up to four weeks or so, but expect lower output following weeks. The productivity falls if crunch runs longer then six weeks. Not sure about in between. That was just one study, so take it or leave it.

Last note: if you have 60-80 hour work weeks before every major release, then there is something wrong with your planning or process. In any case, it does not sounds like the release will be much tested before shipment.

I was almost with you until you got to the last paragraph which is both judgmental and horribly, horribly wrong.

Not only are our releases incredibly well tested (we have millions of players - there is no "untested") but we keep things pretty relaxed during normal development, and pre-release "crunches" result in improved team cohesion and morale. Releases feel like huge wins, and we always celebrate.

Please don't presume to know anything about anyone else without any evidence or first-hand knowledge of a situation. It's rather unbecoming.

Crunch is longer than usual work with a looming deadline. Crunch doesn't have to be long term. Crunch can be working for 7 hrs seeing you are going to make your goal and busting ass for 3 more hours. The slope should match up with the expected outcome, mx+b or else it is futile and not crunch. Crunch should be "doable". What the author was describing is a death-march. The goal isn't to arrive at a location but to kill as many people as possible in the journey. Ref, trail of tears.

The bigger the burn in, the bigger the burn out. The more often the burn in, the bigger the burn out.

If you think burnn-outs are cool, I can package you all of mine with their consequences in a small box and I can send them to you by mail. I'd be delighted, truly. Such a shame it is not possible.

And crunch are just burnin, the maniacal addictive phase before the depressive one. I am not sure living like unbalanced junkies should be considered a good idea. We are not politicians yet.

I'd suggest you read all the way down through the section entitled "What About Short-Term Output?" where this question is addressed.

The problem with short term burst of overtime is the setting of the wrong expectation. People who only see result but don't know the process would expect the same result the next time. You are under the gun to do it again. Doing it enough time and it becomes crunch mode.

The most important bit missed here is - what exactly are you doing?

I started as a physical worker, and when I switched to computers, coming back to physical work felt like taking a holiday.

Now when I'm actively managing people and finances, having a 3-day code crunch is my definition of a walk in the sunshine. Sometimes I literally can't stop smiling during that period, it's such a relief.

(From 2005)

So? Do you have something (newer) arguing reasonably well that it's wrong?

I read that as a reminder that this isn't a new article but rather the older one some of it might have read before. It's fair enough to submit it since some might not have read it though.

HN has gotten in the habit of adding year of publication to submission titles for pre-2014 content recently and this one should probably get it as well.

Yeah, I just thought the title should be updated with that. ea_spouse was still considered current events, but the fact that this stuff from 2005 reads the same today just shows how timeless this research is.

I did seem to have struck a nerve with that comment. In the future I should be explicit that I'm just asking for a title update.

Titles of old articles tend to have (yyyy) slapped after them. So maybe he wants to notify admins that they should add one.

The real problem is that crunch actually works.

[citation needed]

You read a whole article full of evidence to the contrary, and you think you can wave it away by just ignoring it?

I've read it long time ago. I've been in this industry for a long time, and unfortunately crunch does work. I can't say whether for everyone, but it works. It's not question whether I like it, whether its good for the people, families, etc.

But it works!

I'n my experience, crunch mode only "solves" problems that could have been avoided by better organization and/or engineering practice.

I can't think of a single deadline that was "saved" by crunch mode that didn't also have some additional cost - if we had done a better job in the run up and avoided the crunch, we would have been significantly ahead or where we actually were.

Ymmv, naturally.

Where did you read it? All the research and evidence points to the fact that crunches don't work. Overtime (60hr week for two weeks) is different, but extended overtime periods do not work. If you have any citations other than anecdotal evidence, please share with the rest of us

Or maybe you just feel that it works. Thanks for sharing your personal experience, but I think I'll trust the studies.

Reply on news.ycombinator.com