Facebook COO Sharyl Sandberg has kicked up a mini-controversy by admitting to Makers.com that she leaves the office at 5:30PM every day, and has done so for years. In the Valley, where work is a religion, leaving early is heresy.
Earlier this week “Jon” published The 501 Developer Manifesto, a call for developers to spend less time working.
These calls for less time at the office are counter balanced by a recent talk by Google executive Marissa Mayer at an 92|Y event. Mayer dismissed the phenomena of “burn out” as resentment and boasted of working 130 hours a week at times.
Research suggests that Sandberg is probably the more productive executive, and those 501ers may be on to something. In a lengthy essay titled “Bring back the 40-hour work week,” Alternet editor Sara Robinson looks at the history of long working hours and reminds us why the 40 hour limit was imposed in the first place: working more than 40 hours a week has been shown to be counterproductive. It’s a relevant conversation for IT workers, who according to ComputerWorld average 71 hours of work per week.
Robinson draws on research compiled by her husband Evan Robinson, a software engineer, in a white paper titled “Why Crunch Modes Doesn’t Work: Six Lessons.”
Much of the work cited by the Robinsons is focused on manufacturing work, Sara Robinson writes that knowledge workers actually max out at about six hours per day, not eight. I haven’t been able to find her source, and as of this writing she has not returned my e-mail asking what the source is. It could be a survey conducted by Microsoft in 2005 intended to promote the value of the company’s productivity software. Out of a 45 hour work week, survey respondents consider about 17 hours to be unproductive. That’s 28 productive hours per week, or just 5.6 hours per day given a five day work week. This is self-reported productivity, and much of that lost productivity is actually due to meetings, but this is not out of line with the industrial research cited by the Robinsons.
Evan Robinson notes that it’s very difficult to measure programmer productivity, so we’re left having to draw on the lessons learned in other fields.
Evan Robinson refers frequently to studies of the effects of sleep loss on performance, including one study conducted by Colonel Gregory Belenky at the Walter Reed Army Institute of Research titled “Sleep, Sleep Deprivation, and Human Performance in Continuous Operations.” This report finds that loss of sleep affects the ability to do mental work far more than it effects our ability to perform manual labor. If nothing else, sleep loss caused by excessive working hours will cut into developer productivity.
Much of the reduction in productivity comes from making serious errors that take time to correct. Evan Robinson writes: “Crunch raises the odds of a significant error, like shipping software that erases customer’s hard drives, or deleting the source tree, or spilling Coke into a server that hasn’t been backed up recently, or setting the building on fire. (Yes, I’ve seen the first three of these actually happen in the last, bleary days of Crunch Mode. The fourth one is probably only a matter of time.”
Passion is a Fashion
According to Sara Robinson, managers have known for generations that worker productivity declines past a certain limit (eight hours a day, five hours a week). What changed to make people want to work so many hours? Robinson says the emergence of Silicon Valley as an economic power house and the singular minded, hyper-passionate people who worked there.
According to an entry on Folklore.org, a collection of anecdotes about the creation of the Macintosh:
As pressure mounted to finish the software in time to meet our January 1984 deadline, we began to work longer and longer hours. By the fall of 1983, it wasn’t unusual to find most of the software team in their cubicles on any given evening, week day or not, still tapping away at their keyboards at 11pm or even later.
The rest of the Macintosh team, which had now swelled to almost a hundred people, nearing the limit that Steve Jobs swore we would never exceed, tended to work more traditional hours, but as our deadline loomed, many of them began to stay late as well to help us test the software during evening testing marathons. Food was brought in as a majority of the team stayed late to help put the software through its paces, competing to see who could find the most bugs, of which there were still plenty, even as the weeks wore on.
This lead Apple’s finance team to create the “90 HRS / WK AND LOVING IT!” sweatshirts pictured above.
Robinson blames management guru Tom Peters for spreading the idea of “passion” as a substitute for rest and relaxation to the world outside Silicon Valley. “Though Peters didn’t advocate this explicitly, it was implicitly understood that to ‘passionate’ people, 40-hour weeks were old-fashioned and boring,” Robinson writes. “In the new workplace, people would find their ultimate meaning and happiness in the sheer unrivaled joy of work. They wouldn’t want to be anywhere else.”
Regardless of where it came from, or whether being “passionate” actually affects job performance, the concept stuck and has become the norm for corporate America. In his book Jobs That Don’t Suck, a book of advice for getting keeping “good” jobs, Charlie Drozdyk writes that ambitious workers should always show up early, stay late, come in on weekends and always take lunch at their desks – even if you’re just playing solitaire the whole time. Dryzdyk’s advice to young workers typifies the attitude seen in many companies:
If you just said to yourself, “Forget it: I’m not coming in on weekends unless they pay me to come in on weekends,” then your attitude sucks. Try to get your money back for this book, because I don’t even want to be in the same room with you. You’re an idiot, and you should get a job at the post office (not that people who work for the post office are idiots by any means. Many people in my family were postal employees), where you will get paid for coming in on Saturdays.
37 Signals co-founder Jason Fried has railed against this approach for a long time. In his book Rework Fried writes:
Not only is this workaholism unnecessary, it’s stupid. Working more doesn’t mean you care more or get more done. It just means you work more.
Workaholics wind up creating more problems than they solve. First off, working like that just isn’t sustainable over time. When the burnout crash comes–and it will–it’ll hit that much harder.
Workaholics miss the point, too. They try to fix problems by throwing sheer hours at them. They try to make up for intellectual laziness with brute force. This results in inelegant solutions.
So why don’t more companies recognize this and encourage workers to spend less time working?
Is It a Conspiracy?
Mayer suggests that what most of us call burnout isn’t caused by overwork, a lack of sleep, poor diet, lack of exercise or missing out on having a personal life. According to Mayer, this non-existent burnout out thing is actually resentment – resentment caused by missing out on the things that are most important to you. Mayer says it’s possible to work 130 hours per week, without getting enough sleep or exercise, as long as you find time for the absolute most important things in your life:
When she has sensed that an employee was becoming fatigued or annoyed with long hours, Mayer has taken the person aside and asked them what really mattered to them outside of work. For one employee, making nightly 1 a.m. phone calls to her team in Bangalore, India didn’t bother her. What did was missing her children’s soccer games and dance recitals because she was stuck at work. “So, we say you’re never going to miss another soccer game or be late for a recital.”
Apart from the fact that this conflicts with the existing body of research on burnout, there’s something a bit sinister about this. Mayer seems to be asking employees to sacrifice their health and the vast majority of time that could be spent with friends and family at work, with the promise that they don’t have to give up their families entirely. They can still catch soccer games, it’s just everything else that has to go.
It’s easy to see this as a conspiracy on Google’s part: convince the world that burnout doesn’t exist, provide three meals a day on campus so that workers never need to leave, and then maybe someday everyone will just be literally living at work and living for work. But if the research shows that most workers can’t do that much work, then it be in Google’s best interest NOT to encourage this sort of behavior. It could be that the executives running Google and other companies simply unaware of the body of evidence that suggests that overtime kills productivity. But one would think that if there were productivity gains to be had by having employees work less, that would quickly become common knowledge. Why are companies like SAS, which has a 35 hour work week, the exception and not the norm?
Another possibility is that it’s denial, not ignorance, that drives companies to ignore over a century of research. Both executives and employees want to believe that they’re special. Executives believe their companies attract the best people, and that anyone who can’t handle working over 71 hours a week will be weeded out and replaced by someone who can.
My best guess is that Mayer probably really is insanely passionate about her job, and that there are a number of executives in the Valley and beyond that expect that everyone can and should work as hard and as long as they do. Robinson writes:
Asperger’s Syndrome wasn’t named and identified until 1994, but by the 1950s, the defense industries in California’s Santa Clara Valley were already drawing in brilliant young men and women who fit the profile: single-minded, socially awkward, emotionally detached, and blessed (or cursed) with a singular, unique, laser-like focus on some particular area of obsessive interest. For these people, work wasn’t just work; it was their life’s passion, and they devoted every waking hour to it, usually to the exclusion of non-work relationships, exercise, sleep, food, and sometimes even personal care. The popular stereotype of the geek was born in some real truths about the specific kinds of people who were drawn to tech in those early years.
I won’t touch the Asperger’s issue. That’s another field of research entirely and I don’t know that it’s safe to draw any conclusions about the nature of that condition, or any other, and its connection to tech work. But there do seem to be a number of outliers who can easily work long hours. They aren’t necessarily more “passionate” about their work, but they have more stamina.
The problem is all workers are expected to pretend like they are these outliers or risk being branded as dispassionate slackers. In his book, Drozdyk specifically advises workers to be “fake and dishonest” and never to complain about anything – especially long working hours. No one wants to look like a slacker when the next round of layoffs come, but constant overwork will still drag a project or company down.
It’s difficult to determine how many of these outliers exist, but we can look to sleep research to get a sense of how rare they must be.
You Need More Sleep Than You Think You Do
Last year The Wall Street Journal reported research conducted by Daniel J. Buysse, a psychiatrist at the University of Pittsburgh Medical Center on the so-called “short sleepers” – people who can get by with very little sleep. Buysse doesn’t know how many short sleepers there are – possibly 1% to 3% of the populations – but that far more people think they are short sleepers. In other words, many of us are more sleep deprived than we realize.
Also last year The New York Times reported on research conducted by David Dinges at the Sleep and Chronobiology Laboratory at the Hospital at University of Pennsylvania. Dinges concluded that almost everyone needs at least eight hours of sleep per night, and you can’t train yourself to get by on less.
If fatigue from overwork follows a similar pattern to chronic sleep deprivation, there are probably a lot fewer productivity outliers than we think there are. Although Buysse says that many short sleepers gravitate towards tech careers, we can still expect them to be a minority even at tech firms. According information released by the Bureau of Labor Statistics, computer and mathematical occupations account for 3,406,720 jobs as of May 2011. According to other data released by BLS, there were 237,830,000 people of working age in the United States in 2010. That puts computer and mathematics professionals at around 1.4% of the working age population (these numbers aren’t perfect since they are drawn from different years, and the categorization of tech jobs isn’t perfect). Even if we assume all short sleepers are of working age, tech companies would need to attract 50-100% of them to fill all positions – and still may come up short. Plus, all these short sleepers would need to be intelligent and technically minded enough for the work.
Mayer cites Albert Einstein and Winston Churchill as examples of people who worked extensive hours long into their lives, but the truth is most of us are probably more like Archimedes: we need to take breaks in order to make breakthroughs.
How Much Are You Really Getting Done?
Sara Robinson mentioned meetings and e-mail as part of what eats up an eight hour day for knowledge workers. The Robinsons don’t address whether activities and meetings count as “work” during a knowledge worker’s day, but there is some evidence that even very boring meetings require a degree of mental effort. In 2009 Canada.com ran a story on research conducted by Jackie Andrade of the University of Plymouth on the subject of “doodling.” Andrade’s research found that doodling actually sharpens the mind. One possible reason:
When people are bored they have high levels of brain activity, Andrade says. “When you’re bored, you think nothing much is going on, but actually your brain is looking for something to do.”
So we daydream. But daydreaming takes considerable mental effort, particularly when we get stuck in a daydream. “So that sucks mental resources and energy away from the other task we’re meant to be doing,” Andrade says.
Doodling occupies the mind, but isn’t as exhausting as day dreaming. Andrade’s research suggests that it would be better to cut back on unproductive meetings than it would be to expect workers to count meetings as “breaks” (though apparently doodling during boring meetings will make them less exhausting).
This opens up a lot of questions about what counts as “work,” though. We don’t know how much our day to day activities, such as Cooking dinner, reading Russian novels, building model airplanes, reading to our kids, helping them with their homework or just laying around watching TV, affect our cognitive performance at work the next day.
Side Projects and Volunteering
Perhaps the most controversial aspect of the 501 Developers Manifesto is this section:
- Write a technical blog
- Contribute to open source projects
- Attend user groups in your spare time
- Mostly only read books about coding and productivity
- Push to GitHub while sitting on the toilet
- Are committed to maximum awesomeness at all times, or would have us believe it
…we respect you for it. There’s probably some pity in there too, but honestly, it’s mostly respect.
Many programmers contribute code to open source projects on the side not because they want to advance their careers or because they think they’ll be able to build a profitable startup from their work later, but because they legitimately enjoy programming. Unfortunately, the sort of complex thought that goes into open source contributions or other side projects or even volunteer work or activism probably does count towards our cognitive capacity.
I don’t think any company should try to limit developers contributions to open source outside of work, or try to stop employees from joining the boards of non-profits, volunteering at their churches or their children’s schools or otherwise tell employees how to spend their spare time. I also don’t think developers or other workers should stop doing the things that are important to them just to be more productive at work. These are the things that help us lead meaningful, fulfilled lives. But it’s important to keep our limits in mind – spending too much time on one hobby or side project can take away our ability to focus on another.
For employers, it’s worth still worth keeping the work hours short. For workers, it’s worth being mindful of how you spend your energy.
When Overtime Works
It might be OK to do some overtime under the right conditions. Evan Robinson writes about a study by the Business Roundtable called Scheduled Overtime
Effect on Construction Projects:
Productivity drops when working 60-hour weeks compared with 40-hour weeks. Initially, the extra 20 hours a week makes up for the lost productivity and total output increases. But the Business Roundtable study states that construction productivity starts to drop very quickly upon the transition to 60-hour weeks. The fall-off can be seen within days, is obvious within a week…and just keeps sliding from there. In about two months, the cumulative productivity loss has declined to the point where the project would actually be farther ahead if you’d just stuck to 40-hour weeks all along.
In other words, you may be able to use overtime to catch up on lost work, or to make a final push towards the end of the project, but push too much and you’re better off having just stuck to a 40 hour a week schedule.
This paper was criticized in a meta-analysis paper by H. Randolph Thomas titled “Effects of Scheduled Overtime on Labor Productivity.” Thomas noted methodology deficiencies in the paper, including the fact that the data was taken from only one construction project. However, Thomas and Karl A. Raynar later conducted further research on four different construction projects and concluded in the paper “Scheduled Overtime and Labor Productivity” that: “The results compare favorably to other published data including the Business Roundtable (BRT) curves. Therefore, it was concluded that the BRT curve is a reasonable estimate of losses that may occur on average industrial projects.”
I would also note that there are circumstances under which overwork is impossible to avoid. I would not have been able to make the transition from being an IT worker to being a tech journalist if I hadn’t worked part time as a writer on the side. Many people have to work full time while going to school. Some people have to work more than one job to get by. Sometimes freelancers can’t afford to turn down projects with overlapping deadlines.
It may result in overall lower productivity, but sometimes it’s the only way to get something done. The key is to be aware of the effects of overwork and know when it’s worth putting in the extra effort.
What Happens Next?
Chris Nerney writes for IT World:
Even as you read this, thousands of tech workers at Facebook, Google, Zynga and elsewhere are playing the Sandberg card! And when I say “thousands,” I mean none. Because no one who’s putting in 50 or 60+ hours because they’re afraid not to is going to stick out their neck and demand their lives back from the tech jobs that consume them or the venture capitalists who get wealthy on the backs of overworked and stressed-out technology employees. Not in this shaky economy.
I hope Nerney is wrong. The 501 Manifesto is being widely criticized, but it’s also getting a lot of play. Sandberg stepping forward is a good sign, as is Netscape co-founder Jamie Zawinski taking Michael Arrington to task for demanding more from entrepreneurs:
He’s trying to make the point that the only path to success in the software industry is to work insane hours, sleep under your desk, and give up your one and only youth, and if you don’t do that, you’re a pussy. He’s using my words to try and back up that thesis.
I hate this, because it’s not true, and it’s disingenuous.
What is true is that for a VC’s business model to work, it’s necessary for you to give up your life in order for him to become richer.
Meanwhile, companies like 37 Signals and SAS are leading the way on offering sane working hours for employees. But the most important thing might be just taking a long hard look at the data and what it tells us about how much time we should spend working.