Our team has started holding meetings for what we call “premortems.” It’s an exercise we did at Sequoia’s AMP program, where we looked a couple of years into the future and asked, “What will kill the company?” In a premortem we take any task, big or small, and write down everything we can think of that could go wrong. We’ve always had postmortems at Athelas, but it quickly became clear that 90% of the issues seen in hindsight were preventable with some forethought.

You can apply premortems to something as simple as a product demo or as complex as a company-defining event. When we were planning to deploy Athelas at long-term care facilities, for example, we’d been assuming our devices would just connect to the surrounding internet. Everyone has WiFi, right? But in the premortem, someone pointed out that might not be true in this case—and it turned out many of these customers had no internet at all. So we began shipping with our own hotspots, and those critical first days went much more smoothly.

Shorten your iteration cycles as much as you can. We’re a health care company, so FDA approvals and clinical trials mean that our product cycles are naturally long. Our goal is to test each independent assumption with small iteration cycles much more efficiently. For example, before dumping resources into major, long-term clinical trials, we’ll run micro-usability tests and a smaller set of experiments with proxy endpoints.

That same compression of iteration cycle matters on the user experience, as well. We want people’s jaws to drop the first time they use our product. But it’s one thing to delight them that first time and another to keep them coming back—that’s how you know they’re getting a truly magical experience. If not, you probably need a couple more iterations.

About a year ago, I completely turned off email and text notifications on my phone and started setting aside specific chunks of time to respond during the day. There’s really nothing that needs an answer within a minute, and it’s saved me a ton of context switching costs. Before, if I was working on a presentation or talking to a colleague when a notification popped up, just glancing at it for a second could completely sidetrack a thought process. I was no longer fully engaged in what I was doing—and wasn’t able to give all my attention to that notification, either. Setting aside that time has made everything more efficient.

I’ve also started to put my phone in airplane mode at night. I might miss an early-morning call occasionally, but in the long term I think the extra sleep is much more important.

Speaking Mandarin really feels like a missing skill set for me because so many of our vendors and potential customers are in China. Being able to converse in someone’s native tongue builds trust and removes the need for intermediaries like translators. Every language has its own nuances, and it’s not just what you say, but how you say it. When I’m speaking English or Hindi, I can make complex points clearly without worrying about something getting taken the wrong way. I wish I could say the same for Mandarin.

I just reread The Martian, by Andy Weir. It’s a great, fun read and touches on a lot of what it’s like to be a startup founder. The main character works with limited resources in a hostile environment, and gets slammed in the face with a slew of unexpected problems. But he just adds them to his to-do list like any other task and keeps trudging on.

I also recently finished flipping through the annotated screenplay of The Godfather, which is one of my favorite movies. It seems perfect when you watch it, but when you read the director’s comments and all the little anecdotes, the chaos of the whole production becomes apparent. An amazing amount of care went into every single scene.

And I really liked Creativity, Inc., by Amy Wallace and Edwin Catmull. It was interesting to see Pixar’s origin story peppered with management lessons about putting people first and constructing a fertile creative environment. One of my key takeaways was how much a CEO’s job comes down to effective roadblock anticipation and removal.

As an engineer, I love the thrill of building something for the first time. But I’ve come to realize that building a product at scale, with consistency, is the harder problem to solve. Even if you’ve nailed the prototype, bringing it to the masses means making it a million more times the exact same way. Manufacturing and supply chain management are arts of their own.

In our first production build at Athelas, we made 50 devices based on our successful, handmade and tested prototype—and each device failed in its own unique way. We identified every problem on the assembly line one by one—whether it was a vendor that didn’t meet our quality specs or an operator who wasn’t fully trained—and then re-ran the build with much better results. Every person involved with the product needed to deeply understand the importance of getting even the smallest detail perfect.

I take two or three minutes right before every meeting or call to organize my thoughts. I’ll write down a few things I want to get across and a few questions that need answers. We plan what we want to say before the big, really important meetings—but for me, it’s been helpful to do the same thing for the smaller, daily ones, too. Those few moments are a highly leverageable unit of time—they can turn a half-hour call into just 5 or 6 minutes.