Hardware is a Marathon
Exploring the physics definition of Work to reduce 'thrash' and build hardware like an Olympian.
As I dive into the real work of a new hardware startup, I have been turning to successful friends and mentors to get their advice on things to look out for. I have spoken with people from early-stage hardware startups, to former SpaceX lead engineers and managers, to current SVPs and COOs at hardware unicorns.
What came up again and again was the need to work hard. This is, of course, both essential and obvious as table stakes. Certainly, it was requisite in successfully deploying hundreds of millions of dollars of hardware over the past decade. What’s funny though, is that it’s also curiously unclear as to what “work hard” means.
In my experience working closely with dozens of teams over the last decade, I have found that too often the idea of “work hard” is interpreted solely to mean “work long hours over a long period of time.” Certainly, the time we’re at work is going to be part of it, but is it really the whole story? I doubt it.
There were periods as a bush pilot growing up, and again as a launch engineer during a campaign, that I worked seven days a week, averaging 17 hours a day, for literal months on end. No breaks, no days off. More than a handful of times, I’ve been at work manning critical systems for over 24 hours without respite. As such, I’ve had a lot of time to think about the outcomes from those efforts and my take is more nuanced.
Many factors like working on the right things, minimizing thrash, and having clear deliverables must all weigh in. Otherwise, we’re no better than Boxer the horse walking in circles, doing a lot of work, but not getting much really done. Still, the idea persists that the simple secret is just to be there. I wanted to find a new way to poke at it.
A Physical Definition of Work
To get some clarity, I decided to think about the physics definition of Work.
In its simplest form:
Where F is force and S is the distance that force is applied over.
This makes intuitive sense. Applying a force to go some distance passes the gut check for defining Work1. But there are some interesting things to note here. First is that depending on the frame of reference, pushing a box across the room and then pushing it back would mean no distance was covered, and thus no Work was done2. Others would argue you did twice the work (going there and back) but I like the former frame of reference because we all know intuitively that just because energy was expended, that doesn’t mean Work was done!3 They have same units, but Work is different because we must make progress for it to count. In this case, that’s our distance covered.
This all makes good sense, but it’s also a bit too simplistic of a take. We rarely move in straight lines, and neither do our projects. To get a better view, let’s add a critical term.
Work That Moves Us Along a Winding Path
In order to model those projects that tend to follow more of an up-and-down, twist-and-turn path to their completion, a better definition of work looks like this:
Note that now we’re working with an integral, which let’s us have a more organic shape to the distance covered, and that we’ve added a cosine term.
What the heck is that all about?
Don’t worry, it’s not a big deal. We can think of that cosine as an honesty term. Without it, we’re saying that all effort (e.g. force applied) moves us father down our intended path. But that isn’t the case, is it? We all know that we might spend a lot of effort on things that take us absolutely nowhere. For an honest definition of Work, we must only count the force that moves us along the path to our goal!4
Sometimes wasted effort is goofy stuff like the previous footnote about the guy who sent emails all night instead of building his prototype, but that’s rare in high performing teams. In these latter cases, the issue can often be attributed to thrash. Here, I am defining thrash as that lost time moving between efforts in your day-to-day or being unclear where to go next.5
Common causes of thrash for hardware teams include:
Effort of Starting New Tasks
Effort of Task Switching
Working Towards Unclear Deliverables
Minimizing thrash is about maximizing lean process. Woah, easy there! I sense your revulsion and desperate jump for the little red ‘x’ in the upper corner of your browser tab. Stick with me! Too much unnecessary process has rightfully gotten a bad rap, but too little also kills companies through indecision.6
I think the answer (as with most things) is the Middle Way. An ebb and flow of effort at the right times, directed properly down the right path (cosine(theta) = 1). To illustrate that, let’s look at how an Olympian trains.
How Olympic Marathoners Train
We always say building hardware is a marathon, right? I build teams with the Olympians of their discipline. The Eliud Kipchoge of thermofluids, the Haile Gebrselassie of structural design, the Catherine Ndereba of deployment. We all plan to have the absolute best in our fields to help bring the vision of the company to life. Now, imagine assembling that team and then telling them to run a full marathon every single day.7
We agree that would be insane, right?
We know that’s not how training works and yet, that’s how we must interpret the common advice of “work hard, all the time” for hardware engineers the world over, even though we’re also squishy humans trying our best to perform optimally while warding off diminishing returns (or worse, injury and burnout). In my view, that’s poor management.
I believe the secret to engaging strong teams lies with good planning, sane milestones, and an aggressive (yet achievable) schedule co-developed with the stakeholders8. That represents, in my experience, the aforementioned Middle Way. Like any plan, it can change, but it’s important to have one because otherwise you enter a mode of pure reactivity, waiting for things to happen to you and responding rather than proactively moving forward. Within that schedule exist times to push, and also exist times to recover and assess the result. Update the plan, move forward again.
Now, depending on which side of the table you’re on, you’re either agreeing with me or praying for (and investing in) the arrival of tireless AI robots9.
Just like in our physical definition of work, the key is strong, hard, unyielding effort at the right times. Riding the wave of build up to the crest of an event and then going into active recovery on the other side. Never yielding wholly in output or thrashing uselessly on things that tire without providing progress, and also not keeping the pedal to the metal 24/7.
The magic, for us as founders, is to craft programs that understand the pushes and then maintain the team’s energy for when the hard times really come.
Note that I will use both ‘work’ and ‘Work’ where the former is being at a job and the latter proper noun is the physics definition.
If you go “S” feet and then go “-S” feet back, your sum total is zero. Sorry, buster.
I think of the startup CEO I was mentoring who screamed at me, “I WAS UP UNTIL 3:00 AM SENDING EMAILS! I AM SO BUSY!” to which I replied, “And yet you have all the parts for your first prototype sitting on a shelf untouched.” Effort doesn’t mean progress.
For those interested, theta is defined here as the angle between the applied force and the path, meaning it is a dynamic angle moment to moment, but always produces the component of the force along the line, while leaving out the orthogonal component (sine of theta).
It’s not a perfect analogy, but it works.
I’ll write more about this in the future, but iterating weak design through physical hardware dev (especially when unsure about the goal) gets expensive fast.
Of course, your job as their boss is to follow them in a little golf cart screaming, “FASTER! FASTER!” That’ll work for sure, right?
More on this another time, too, but note that those stakeholders are managers, employees, investors, customers, vendors, and regulators to name just a few.