It was the worst sales call I’ve ever had. I normally felt comfortable talking to prospects about my web design firm, Cantilever, but this time I was inexplicably nervous. I was sweaty and dry-mouthed, stumbling over my words and struggling to hit my normal cues. The lead was a great fit, and they were referred to us by a satisfied Cantilever client, but I was losing them. I cut off my guests mid-sentence, yet left awkward pauses between my own. Being uncomfortable, I let the conversation die out quickly without even making the case for hiring us, practically suggesting that the client not bother.
What was different? I was overworking.
Not long before that call, Cantilever had landed two big projects at once, and we were all in crunch mode. I was spending weekends at the office and sneaking in late coding sessions after the kids were in bed. Our clients counted on us, and I was anxious about letting them down. On top of that, I was visiting one of our large clients regularly off-site, leading to wasted time in transit and travel stress. I was eating too much and spending too little time with my family.
In this compromised state, the hours I spent with this sales lead were not merely less effective than normal, they were downright harmful. Good leads come around only once in a while, and I blew it. I could have postponed or asked one a more rested colleague for help.
I would have better served Cantilever by taking a nap than taking the call. Working longer backfired.
In 2014, researchers investigated the relationship between hours worked and real output using data collected by the British government during World War I. Within the factory they studied, they found that output dropped precipitously after the first 56 hours of work per week. Furthermore, they observed a “Productivity Cliff” — when these workers put in more than 63 hours per week, it actually hurt output, because maxed-out workers are more likely to make costly errors or become ill. They found that for these workers, putting in 70 hours produced the same output as 56 did.
A recent report in Salon summarized these findings for a peacetime context:
Industrial workers have eight good, reliable hours a day in them. On average, you get no more widgets out of a 10-hour day than you do out of an eight-hour day.
Knowledge work can stretch even less far. Modern office workers log absurdly low levels of productive time per hour at work. The philosophy of “Deep Work”, popularized by the writer Cal Newport, encourages knowledge workers to set aside 4–7 uninterrupted hours per day to confront their most challenging tasks. Push further, and productivity will decrease. Push even more, and you could actually harm output.
On software projects like the ones we do at Cantilever, mistakes due to overwork can be epically harmful. Squeezing in extra hours, one might deploy bad code that costs more time to remediate than would have been spent to do it properly. An overworked team member might design a confusing user interface that must be corrected with an expensive rebuild at the end of a project. The consequences to morale can be just as deleterious. Harvard Business Review reports:
…when we’re low on energy, we tend to misread those around us, typically in a more negative fashion. Happy faces appear more neutral. Neutral face start to look like frowns. What’s more, when we’re fatigued we find it harder to resist lashing out at perceived slights. Not only to do we incorrectly perceive the world around us more negativity, we’re also more likely to act upon that information.
Every person, environment, and type of work has its individual “Productivity Cliff.” Anecdotally, in my own experience and after managing and working alongside dozens of programmers and designers, I’d estimate that the most eager individual can do around 30 hours of strong, productive coding or design in a week before productivity goes south. Add in a few hours of meetings and time for lunch and breaks, and you’re at a standard 40-hour modern work week.
One might add another 15 or even 30 hours per week over the course of a few months, but productivity eventually deteriorates back to the 30-hour baseline level, a phenomenon that Salon also observed.
This is the double penalty of overwork. Not only is it potentially counterproductive in the moment, once it becomes a habit, it compounds. Push for too long, and 50 hour weeks start to resemble the productivity of 40 hour weeks. A few months later, you need 60 to reach the same level. Overwork can become an addiction, and like many addictions, it isn’t long before the addict needs to increase their dosage.
Overwork can be counterproductive because of what “time off” actually represents. A break from one’s primary work is time to time to get inspired, learn new skills, and network. It’s time to mentally reset, exercise, cook healthier meals, spend time with family and friends. Time off can be profoundly productive. For many creative people, it’s the fuel their work depends on.
The composition of those extra hours is also problematic. If a coder can really only code productively for 8 hours a day, what are they going to be doing for the rest of an 80-hour week? We’ve witnessed projects where the development leads spend more time explaining their status to upper management than actually working on the problems they’ve been commissioned to solve. Overwork breeds inefficiency as workers strive to meet their culturally mandated hours marks while staying sane.
I agree with critiques of the research cited above, arguing that 40 hours is not necessarily the “Productivity Cliff” for knowledge work. For some people and situations, I think 50 or even 60 productive hours per week is sustainable and efficient. But equally, I believe there are many high-creativity or high-focus roles where 30 or fewer hours may represent the cliff. A fine artist will often spend just a few hours a week literally making art. Doubling the amount of pieces he makes would lower the artistic quality of each one and would decrease his overall creative success. If a surgeon doubles the number of procedures she does in a week, her increased risk of errors in any one procedure may drive her below the point of being actually useful to patients.
Dedication is admirable. One should always strive to maximize one’s output. It’s just that for most people, 40 distraction-free hours per week is just as effective as 80 scattered ones. True productivity demands balance.
While the science seems settled, modern work culture has often ignored the lessons of the past. In software, the “Cult of Overwork” (a term coined by New Yorker contributor James Surowiecki) is rampant. The coder who sleeps in the office is seen as a noble martyr. The manager with the most Tetris-like calendar is seen as the most effective and integral to the firm. In 2014, Erin Hoffman, the wife of an Electronic Arts engineer, famously revealed the intense working hours expectations at the games manufacturer, especially during a “crunch.” Her polemic struck a chord within the games industry and eventually led to a $14.9 million overtime-pay settlement.
The current mandatory hours are 9am to 10pm — seven days a week — with the occasional Saturday evening off for good behavior (at 6:30pm). This averages out to an eighty-five hour work week.
There is no chance that EA was receiving 85 hours worth of output from these beleaguered developers. I would be surprised if they were even getting 40. The engineers attempted to communicate this point:
Complaints that these once more extended hours combined with the team’s existing fatigue would result in a greater number of mistakes made and an even greater amount of wasted energy were ignored.
If increased hours don’t lead to increased output, why do we do them?
For me, the Cult of Overwork is a part of my identity. In the early days of Cantilever, I felt that clients who received emails from me at 2 AM would take it as a signal of our dedication. Like many founders, I believed that if I didn’t outwork my team, they wouldn’t respect me. Whether or not I was actually performing well was hard for the team to see, but they would see my early-morning commit messages.
Signaling is a fundamental part of human psychology. We can’t tell who is dangerous and who is not, so we use heuristics to inform our decisions. We can’t read each other’s DNA, but we can use body shape and facial features to decide in a split second whether a stranger is a potential mate. Overwork is another such signal, an easy proxy for actual working output. It is hard for managers (or, often, colleagues) to determine who is a skillful coder and who is not, so people use the external proxy of overwork to make lazy assumptions about who is effective.
Not that this is conscious. I know better than anyone the allure of being the workhorse, and as a manager, I’m just as susceptible to overrating the performance of people who jam long hours. In the technology sector, defined by iconoclastic obsessives, the Cult of Overwork is the dominant orthodoxy. Like many religions, it demands tribute — but in the form of unnecessary and counterproductive excess hours.
Firms which adhere to this doctrine are therefore not comparing employees and job candidates on their ability to do the work, but on their willingness to pay this tithe. By establishing that the important factor is sacrificial overwork, not just actual output, employers exclude individuals who meet the output criteria but not the sacrificial one.
Who might those people be? Parents, especially those who manage the endless parade of day-altering responsibilities wrapped up in taking care of children. People who take care of their aging parents, a sick spouse, or a family member with special needs. People with health conditions that prohibit overwork. People who commute for too long to stay late. People who need or want to have a second job, or who are in school.
Firms which obey the Cult of Overwork issue an automatic blanket rejection of these people, regardless of their talent, experience, or level of dedication. They inevitably create a monoculture where only the financially and socially advantaged are valid. It’s starkly antithetical to everything most firms state that they want — a claim I believe is genuine, in most cases — but it’s the inevitable product of the culture that they have allowed to take hold.
A 2018 New York Times report included an anecdote which is both tragic and predictable:
Last fall, I was in a meeting with a leader in women’s health, discussing re-entry-to-work programs for new mothers when, out of the blue, she began complaining about a former employee. This employee on their small team had gotten pregnant, the woman said — and it was a problem: “She was way too focused on her pregnancy. It was distracting her. I didn’t think she was going to be committed enough to the job, so I had to let her go.”
The “Motherhood Penalty” is the well-studied phenomenon that mothers are less likely to be hired, promoted, and to be allocated internal resources than their male colleagues. Employers seem to be hesitant to invest in mothers because they doubt their dedication and long-term value to the firm. The New York Times also reported on a Stanford study in which fake resumés were sent to hundreds of employers:
They were identical, except on some there was a line about being a member of the parent-teacher association, suggesting that the applicant was a parent. Mothers were half as likely to be called back, while fathers were called back slightly more often than the men whose résumés did not mention parenthood.
The idea that mothers are trying to “have it all” — a successful career, thriving family, personal pursuits — is a myth. We are all trying to “have it all”, or we should be. It’s not greed, it’s common sense. If working 80 hours and 50 produces the same output, it is masochism to choose the 80, and sadism to ask it. When firms ask unattached people to submit to this lifestyle, many are driven enough to say yes. Mothers simply don’t have that option.
Historically, fatherhood has paradoxically meant improved career prospects — the so-called “Fatherhood Bonus”. There is bias behind this trend, but in a purely mechanistic sense, it is predictable. Families consolidate and redistribute responsibilities after childbirth. One parent will often focus more on career to ameliorate the loss of income from the parent dedicated more to childcare. Naturally, this focused, motivated individual would be attractive to employers. Traditionally, this has been Dad — but encouragingly, the tide is turning. More and more fathers are taking an equal or even primary role in childcare. In a changing society, mechanistic bias towards a family’s sole earner will no longer be assigned by default to men. I predict that over the course of the 21st century, both the Motherhood Penalty and Fatherhood Bonus will lessen, and the Cult of Overwork will be increasingly salient for men, both financially and politically.
Parents bring unique and invaluable perspective to a team. When you take care of children, you have no choice but to grow more patient and more resilient. In giving so much to a child, you sublimate your ego. You learn how to teach, to motivate, and to energize another human being. You learn how to prioritize, and get the most out of short bursts of productive time. You learn how to push yourself harder than ever before. Who doesn’t want a teammate with those skills?
The Cult of Overwork is not just a challenge to social justice, it is an impediment to corporate success. If a firm wants maximum output, the Cult of Overwork is doubly harmful. Not only will their people hit the Productivity Cliff far more frequently, reducing their real output, but the insidious norm of long hours will filter their workforce down to a homogenous subset of the actual talent they have access to. Diversity is demonstrably and obviously beneficial to group decision making — consider how much more useful it is to have a fork and knife than two knives or two forks. As a recent McKinsey study put it:
More diverse companies, we believe, are better able to win top talent and improve their customer orientation, employee satisfaction, and decision making, and all that leads to a virtuous cycle of increasing returns.
I am learning from my past mistakes. Nowadays, part of our mission at Cantilever is to provide an alternative to the Cult of Overwork in the software industry. We are just getting started, but are constantly attempting to create an environment where people of any background and life circumstance can do their best work.
We have designed our company around remote work, flexible schedules, and sensible standards for communication. Our people work where they want, when they want, wherever they want. As a small firm, we have never formally provided paid parental leave, but we feel it is time to fill that gap. In 2019 we will offer a modest program which we hope to expand year by year.
We don’t do these things just because they are nice for employees. We do them because it creates a key strategic edge. Smarter design, informed by a broader diversity of viewpoints. Fewer preventable errors. Less turnover. Combatting overwork is good business.
The Cult of Overwork is a tempting shortcut, especially when one is growing a business. But there is a time for exertion and a time for recovery. Recognizing and harnessing the rhythm of a healthy, balanced working life is key in getting the most from one’s talent. I’ve learned this lesson time and time again: Peak performance and constant performance are very different. Choose to be at your best by choosing to rest, and if you manage others, let them do the same.
Edited by Rebecca Testrake
Studio Manager & Magical Instigator at Cantilever