Hardly Working
Why Your Vacation Policy Doesn’t Work
One of my new colleagues admitted recently that his wife was having a difficult time getting her mind around our company’s vacation policy. Like many other tech companies, we have what is referred to as an open, unlimited or untracked vacation policy. Our vacation days are neither accrued nor limited; instead, employees are trusted to take the right amount of vacation for them to ensure a proper work/life balance.
I was introduced to this type of policy in 2008 by a friend who worked at Netflix. Netflix is generally regarded as a pioneer of the open policy — with no policing or tracking. They rely on their “freedom and responsibility” culture to govern the number of days employees take off each year. The moment I heard about the policy, I knew I wanted to adopt it for our new company, textPlus.
Up until that point, I had only been a part of “traditional” vacation plans where employees accrued vacation days as they worked throughout the year. Generally, vacation days accrued more quickly as tenure grew. This accrual would be capped and paid out as cash when the employee left the company.
This traditional plan has been, and continues to be, in widespread use, but I believe it is fundamentally flawed.
Here’s why:
1) The technology industry, in particular, doesn’t follow the industrial model where “shifts” begin and end with a work whistle. The entirety of our work is done between our ears, intermingled with everything else. We’re on email every day, logged in every night, on call at three in the morning and obsessing about details around the clock. There’s no firewall between “work” thoughts and “leisure” thoughts, so the concept of separating and tracking “work” and “leisure” hours is no longer as relevant.
2) Entrepreneurs are taught that execution and results are the coin of the realm — not hours in the saddle.
3) Vacation is deserved, and meant to be taken, not accrued or converted into cash. Hard workers need time away from the office to recharge, see friends and family, master a hobby and attend Burning Man. Or whatever floats their boat.
4) Too little vacation is bad for business. All work and no play leads to less happiness, diminished productivity, and dwindling creativity. Plus, it generally makes people less pleasant to be around. Jim McCarthy coined one of my favorite software mantras: “Team = Software.” If the majority of a team is burned out, the software will follow — and nobody wants that.
5) Accrued vacation is flawed at its very core. If vacation is not being taken then it’s not vacation. I’m more than happy to pay an employee while they take a road trip, explore Rome or hike through the rainforests of Costa Rica, but why should I pay them double time to not take vacation? Isn’t that effectively what I’m doing for two, three or four of their work weeks each year, depending on their accrual rate? Furthermore, accrued vacation requires management; it’s a liability on the balance sheet, and payout timing is hard to project.
Fortunately, Netflix showed us a better way — offer freedom to take as much vacation as is needed. In exchange for that flexibility, however, employees are expected to take responsibility for producing extraordinary work.
Our version of the Netflix system has been in place for almost seven years now, and in my eyes, it’s been a success. We evaluate it every year (more on that below) and each year it seems to pass muster. For the most part, our system requires very little overhead. Of course, the plan is not without its challenges, many of which are echoed in critiques of the Netflix model.
The most common critiques of the open vacation policy:
1) The policy will be abused and people will take too much vacation time.
2) The policy is intended to guarantee that employees will take too little vacation time.
3) The policy is vague and confusing for employees.
4) The policy is hard to coordinate and managers can still withhold approvals .
Let’s start with the notion of abusing the privilege. Are some employees going to “take advantage” of the system and disappear for long stretches at a time? Maybe so, but in the seven years we’ve had our policy in place, we’ve had very few cases of misconduct.
Avoiding exploitation starts with the hiring process. At textPlus, we take great care when hiring into our “freedom and responsibility” culture and have found that those employees who are a fit would never think of abusing the generous vacation policy. Their ethos lines up with the “responsibility” side of the equation. Regardless of philosophy, it is difficult to abuse a vacation policy while meeting expectations and delivering extraordinary results.
Another way to avoid abuse is to draw very clear lines between vacation and other forms of absence, for example: sick days, family leave, maternity/paternity leave, etc. We’ve made a hard separation between those separate concepts. Vacation is strictly defined as days off that replenish an employee; enabling him/her to take some much needed R&R.
In contrast, other types of leave are governed by various state and federal laws and thus should have separate policies that adhere to those guidelines. For example, our team knows the difference between a sick day and a vacation day, and understands how many consecutive sick days trigger sick leave. This distinction is particularly relevant now because California mandates paid sick leave (to be capped at three days per year, if desired).
The flipside of abuse is the critique that unlimited vacation policies actually discourage vacation time to be used. The cynical point of view is that these unlimited policies are put in place to encourage fewer vacation days. I can’t speak for all companies, but I know that this is not the point of our policy at textPlus..
The truth is, taking insufficient vacation is a cultural problem, not a systemic one — it exists regardless of which vacation model a company uses. In my two previous companies, Activision and JAMDAT, we carried over large amounts of accrued vacation on the books because employees at those companies didn’t take a lot of time off. In fact, at the management level, accrued vacation was worn as a badge of honor; while the junior team viewed it as a future severance package or safety net. We’ve taken a more active role at textPlus to avoid this cultural paradox, and we’re constantly reviewing our policy to ensure it’s the best fit for the team. A few months ago I asked for volunteers to self report their 2014 vacation days. Ten team members took me up on it, and here are the results:
Analyzing the chart, a few things jump out.
First, the number of actual vacation days, shown in column 2, are light — ranging from three to 12 days (which is also a pretty wide spread). I can only guess as to why vacation days were so low. I’d venture a guess that it had to do with the workload of rewriting our entire platform in 2014. As previously mentioned, the 24/7 work culture of the technology space, as evidenced at JAMDAT and Activision, can be hard to transcend.
It’s important to note that there may have been a data bias with the survey, above; considering respondents volunteered to participate, only if they felt comfortable revealing their days off. Based on these findings, it’s entirely possible that untracked vacation policies do, in fact, encourage people to take less vacation, although I still point at my Activision and JAMDAT as counterpoint examples.
Fortunately, our policy doesn’t end with the number of days that employees take off. We’ve incorporated a number of forced closures each year, including nine bank holidays and ten days between Christmas and New Years. Because of these closures, each employee averages roughly 28 days of vacation a year — just shy of five workweeks. These forced closures have the added benefit of shrinking the inequality gap we saw in the vacation day analysis above — most days taken are no longer a significant multiple of the least.
The final two critiques, guidance and coordination, are issues that come up a lot. First, let’s tackle the question of “guidance.”
Some employees find the open-ended vacation policy a little too open. They don’t love “vagueries” and want to know how much time is appropriate to take, rather than play a guessing game. Similarly, many managers express concern about the potential for serious inequalities across their teams, due to a lack of guidelines. These are very real management issues that we have faced the most over the years.
Because I’m personally more comfortable than most with ambiguities, and because I love the simplicity of the Netflix policy, I’ve been resistant to overly define the answers to these questions. In response, I often share a favorite (unattributed) quote, “I’m happy to deliver 12 months worth of work, just not in 12 months.”
Based on that notion, I ask each individual to gauge what feels right for him/her. I encourage employees to ask themselves: Am I someone who can produce 12 months of work in eight months? Seven? Nine? How much time do I need to deliver extraordinary results? Interestingly enough, our small sample of employees from 2014 revealed a very narrow range of 7.5–7.8 work months.
The final challenge is coordination. Our company has decided that manager approvals are not required for vacation days. However, in the early days, we found some people would disappear for vacation without letting their teams know. To solve for this, we’ve added the explicit notion to our charter: “responsibility for having your work covered when you’re not (working).”
Coordinating absence is an important responsibility for all employees. While we don’t approve or track vacation days, one tool we’ve embraced is a shared vacation calendar that employees update when they’re gone. It’s not used to track days, in fact, we often clear the past days and months. It’s solely used to coordinate and communicate.
Overall, I’m comfortable with where we’ve ended up on our vacation policy. I feel like we’ve solved many of the critiques of the system by combining solutions effectively, including providing forced closures, implementing a tracking calendar with “amnesia,” etc. But it’s a work in progress and I’d love to answer questions and learn of others’ experiences and opinions on the subject.
For more reading, here’s a thoughtful and balanced piece by Jacob Kaplan-Moss who works for Heroku:
