Written by I’ve almost burned out a few times now. In each case, I was lucky enough to realize (with the help of supportive managers) that what I was doing was unsustainable, before it was too late.
That this has happened a few times may be surprising. Why didn’t I learn enough from the first time to completely prevent burnout from then on out?
The reality is, new circumstances and challenges kept testing my limits in very different ways. No two of these high-intensity situations were exactly the same—therefore, the tactics that I’d implemented the first time did not necessarily save me from the second time, and so on.
It took me a while to learn the higher-level lesson: adaptability and experimentation are key! Now, when I start to observe the signs of burnout in myself (not being motivated, always feeling stressed or tired, doing less-than-my-best work), I reconsider my current tactics and start testing new ones.
There’s a basic mental model I follow to improve my own sustainability. These steps should benefit anyone in a similar position:
- Decide on deliberate principles to apply to time management decisions.
- Devise a process for allocating time.
- Experiment with tactics to optimize particular activities or responsibilities.
In this post, I’ll share the particulars that I ended up with. Mine are certainly not meant to be prescriptive recommendations—instead, I would encourage anyone feeling symptoms of burnout to work through the above model, and come up with the approach that is uniquely right for them.
Principles
When I stepped back to evaluate my unsustainable situation, I came up with a few key principles to guide the rest of my decision-making. Thinking about it from first principles ensured that I did not misdirect my energy into habits that would ultimately be a dead-end.
Principle 1: Time management is about global optimization
My first real brush with burnout came while I was an engineering manager. Although time management is critically important regardless of one’s role, I had overly constrained myself because of the responsibility I felt to my team members.
Because I was an independent contributor (an engineer) prior to being a manager, I had very strong opinions about what I would want out of a manager. I considered things like “weekly one-on-one meetings with the whole team” to be axiomatic requirements, and that if I let any of those slip, I wouldn’t be a good manager to the people I was supporting.
Unfortunately, there’s intrinsic tension between selflessness and good time management. Because our time is finite, we are forced to make hard decisions about allocating it — and in considering the various trade-offs, we inevitably have to confront the question: which way to spend my time is best for me?
I realized that time management is essentially a global optimization problem, with many variables to simultaneously maximize.
For example, I wanted to maximize all of the following (and more):
- My helpfulness to my team
- My own productivity
- My energy levels and sustainability
Even though I should have been optimizing my time globally, I had fallen into a local maximum: I was so focused on one of my goals (#1) that I neglected to incorporate the others (#2 and especially #3).
With the challenge presented in this way, I decided to approach it like a multi-armed bandit problem (sometimes referred to as the “explore/exploit tradeoff”):
… a problem in which a fixed limited set of resources must be allocated between competing (alternative) choices in a way that maximizes their expected gain, when each choice’s properties are only partially known at the time of allocation, and may become better understood as time passes or by allocating resources to the choice.
To fully explore the space of possibilities here, I decided that I needed to experiment with uncomfortable changes. And even though changes will not always pan out, the process is necessary to approach the global maximum. I needed to test alternatives to what I was doing, and rationally evaluate the costs versus the benefits.
I’ve listed some of my results in the “Tactics” section below.
Principle 2: Focus on comparative advantage
Comparative advantage is an economic principle which roughly states that, even if two parties (or people) are capable of doing the same thing, it’s to the benefit of both to specialize in the things where they have the lowest opportunity cost:
… if two countries capable of producing two commodities engage in the free market, then each country will increase its overall consumption by exporting the good for which it has a comparative advantage while importing the other good….
The principle of comparative advantage suggests that we should each focus on the tasks that we are in the best position to do, or that won’t happen without our intervention, and delegate the rest to others.
These tasks — the ones that only you can do — would be your opportunity cost if you prioritized other work. In a time management context, there’s always some opportunity cost, but it can be minimized by delegating to others.
The famous (and somewhat unintuitive) insight of comparative advantage is that it benefits everyone if each party minimizes their opportunity cost. Even if you could do every task better than everyone else, it’s still globally beneficial for you to delegate.
Principle 3: Latency matters
In a time management context, latency is how long it takes one to do or respond to something. An urgent issue demands a low-latency response, while latency is less important in responding to, e.g., a low-priority feature request.
Most of the time, I like to think of my productivity in terms of throughput, and optimize for that. Higher throughput means that I am doing more things, and therefore scaling myself and my impact.
However, when I was most underwater, I found that sacrificing latency to increase my throughput was making things worse. Latency does matter — especially on a distributed team (where there is already additional latency across timezones) and as an engineering manager (where delayed responses could significantly affect team execution).
But neither is optimizing solely for latency the right answer: if one did this, one would end up doing all the “urgent, but not important” work, without leaving any time for the “important, but not urgent” work. My solution has been to hold myself to roughly a 24-hour SLA (a kind of self-imposed commitment), ensuring that nobody has to wait on a response from me for longer than one working day.
Process
The next thing I needed to develop was a process by which to organize and prioritize my work. Without some sort of system, I was just ending up with a big grab-bag of obligations — introducing overhead every time I needed to choose what to work on, and contributing stress because it felt insurmountable.
The process that I landed upon was significantly influenced by the seminal time management book, Getting Things Done.
The to-do list
I created a to-do list for myself with four sections:
- “Today” contains tasks that I absolutely want to do before I leave for the day. By the end of the day, these items should be completed, or pushed back into the “Backlog” section (#3) if they’re actually not important enough to finish right away.
- “Waiting on others” contains tasks which someone else is responsible for (or when “the ball is in their court” now), but which I should check up on at some point. This includes tasks which I have delegated, but I am holding myself accountable for, so I need to follow through and ensure they make forward progress.
- “Backlog” contains all the other tasks that I need to do, in no particular order. I’ll put due dates on tasks here when they have real deadlines, so I get reminders to prioritize them when the date is approaching.
- “Someday/maybe” contains tasks that I may want to do at some point — my wish list or “nice-to-haves.” Basically, this list is for whenever I happen to have free time, or want to procrastinate productively. Nothing that needs to get done should be in this list.
I keep my to-do list in a digital document which syncs to the cloud and all of my devices, so I can edit things on-the-go, but I know several people who swear by pen and paper as well. The important thing is having a system.
Daily triage
To populate the to-do list that I maintain, I need to keep up-to-date on what’s happening around me.
Every morning, I start my day by draining (reading through and clearing out) all of my queues of notifications: posts, documents, chats, issues, code changes, etc. For each notification, I decide on-the-spot between three actions:
- Ignore, by hiding, archiving, or deleting the offending item. If I find myself ignoring notifications from the same source frequently, I will start filtering them out more aggressively (e.g., by turning off notifications for a particular thread or project).
- Take care of immediately — for example, by responding to emails or posts. I only respond immediately when it can be done in a very short period of time (roughly 5 minutes or less), unless the notification is particularly time-sensitive.
- Stash aside for now, and put a note in my to-do list about it. The exact way in which I do this will vary by medium; for emails, I use a folder named “To-do related,” so I can come back later when I’m ready to complete my stashed tasks.
The goal here is simple: I should understand everything that’s on my plate before I start progressing through it. If I don’t triage before starting work, it’s very likely that I won’t be directing my energy into the tasks that are the most important. If I don’t put everything into one place (my to-do list), it’s hard to see the bigger picture.
Weekly triage
Every Monday morning, after I’ve done my usual daily triage, I also review my whole to-do list, and take the opportunity to drop tasks, bucket them differently (e.g., move tasks from “Backlog” to “Someday/maybe”), or delegate them as suggested above.
This is even more important than the daily curation because this is how I prevent unbounded growth of my to-do list. No matter how productive I am, there will always be more to do, so it’s essential that I aggressively drop those things which are low in relative importance, and delegate those things where I don’t have a comparative advantage.
In my Monday triage sessions, I also try to observe if my to-do list is growing week-over-week, despite my best efforts to keep it under control. If it is, this is a signal that I should look to shed some of my responsibilities, because they are generating too much work for me to keep up with.
Monthly, quarterly, or yearly reviews?
The daily and weekly reviews of my to-do list have been critical mainstays of my process ever since I introduced them, but recently, I have started thinking about how to take a broader perspective as well — to introduce time for thinking about goals in addition to just tasks.
To force that shift in perspective, I may start checking in on my goals with a monthly (or longer) cadence, but I haven’t implemented this yet.
Tactics
Finally, once I knew my principles and my basic process was in place, I experimented with different optimizations in how I spend my time. Some failed, some succeeded… some succeeded then broke down later… but I would not have found the successful tactics without being willing to take risks.
Since I came up with these ideas for myself, with little thought to how they would generalize to others, I won’t go into too much detail. (I can discuss the particulars more later, if there’s interest.)
Calculating where all the time is going
When I first wanted to understand where all my time was going, I actually went through all of my recurring meetings, and put them into a spreadsheet. I created separate “monthly,” “biweekly,” and “weekly” sections, and wrote simple spreadsheet formulas to calculate how much time all of those meetings average out to per week.
From that, I learned that I spend more time in a meeting room than not. More distressingly, I learned that I wouldn’t be able to take off work more than one or two days in a week, without being forced to cancel some meetings. This was a big wake-up call for me, and I immediately started culling those which were least valuable and/or highest cost.
I wouldn’t necessarily suggest tracking meeting load all the time, but doing it as a periodic exercise can really open your eyes to how much time isn’t “yours.”
Biweekly 1:1s by default
As mentioned above, I used to feel that weekly 1:1s (one-on-one meetings) were sacred. But as part of exploring the solution space, I tried 30-minute biweekly 1:1s with my team. Although I have not stuck to them permanently, I found the biweekly cadence quite helpful while I was in crisis mode.
Importantly, I knew that I might miss some signal, or occasionally miss opportunities to help my team (and I was open about this with them when we made this change). However, it yielded more time for other important things like thinking about team strategy and mission. The goal here was to obtain most of the value with half of the time.
I employed a few safeguards as well:
- Even with biweekly 1:1s, I tried to ensure that my 1:1s each week were an effective cross-section of the teams and projects that I supported, so I got some sampled signal from each area. This meant lots of context-switching, which was intimidating (don’t we want to avoid context-switching?) but worked really well in practice.
- My team was distributed, and therefore made it a point to communicate asynchronously as much as possible. In turn, this meant that I could pick up important signal by following along in those places, and not just through 1:1 meetings.
- Where the “default” configuration of 30 minutes, every other week, didn’t work for a particular team member, we adjusted! This doesn’t need to be one-size-fits-all.
Less note-taking
I used to take notes in almost every 1:1 and share them with the person I had met with. I realized that this was creating a lot of stress and extra work, and the added value was minimal.
Instead, I decided to be more probabilistic — if I don’t remember something, it’s probably not that important! When I want to be absolutely sure not to forget something, I will pull out my phone and take a note on-the-spot, but I now avoid bringing anything with me as much as possible.
One note, though: this is different from taking notes to share with people who weren’t in attendance. Group meeting notes can reduce the amount of time that each person spends in meetings, without leaving anyone feeling out-of-the-loop.
Blocking out time, ahead of time
When I have tasks to accomplish that I know will require dedicated, uninterrupted time, I block out my calendar in advance — usually for 1 hour or more.
I find this especially useful when I need to spend time writing a lot of text, because I personally can’t make progress on writing when I’m constantly context-switching to and from other tasks or meetings.
Leaving time for non-essentials
When I first realized I was burning out, I went into “survival mode.” I extricated myself from every responsibility I didn’t absolutely have to have, and became unavailable for anything optional.
Counter-intuitively, dropping everything actually ended up stressing me out more! It turns out that some of this “optional” work was work that I enjoyed doing, like mentoring others, and that these were contributions that I wanted to be making.
Once I became more established in my process, I reintroduced some of these energy-giving activities into my routine, and essentially planned around them as a fixed time expenditure — which has been a good forcing function to organize the rest of my work more efficiently as well.
Conclusion
It’s far too easy to settle into a particular routine and never question whether it’s actually the most effective use of time. I’ve learned that, for any particular task or responsibility, I can often get most of the same value for a fraction of the time, and then “reinvest” that saved time into other things that I want to be doing as well.
I like to push myself, so even after the above changes, I’ve still occasionally struggled with burnout. However, I’ve learned to keep experimenting — to always be working toward the global maximum. By always exploring different principles, processes, and tactics for managing my time (and energy), I can avoid falling into a rut or, worse, a downward spiral.
Further reading
The book Algorithms to Live By, although not about time management per se, is an interesting (and a bit cheesy) exploration of ideas from computer science that could have application in our daily lives. The explore/exploit tradeoff is one of these.