Talent Buys You One Good Season. Culture Buys You the Next Ten.

What I learned about building durable organizations from a place that is forced, every single year, to do the thing most companies avoid: let its entire senior leadership walk out the door.

Every May, our team loses its best people.

Not for cause. They graduate. In competitive high school robotics, the calendar does what most companies never have the discipline to do on their own. At the end of each season we lose our seniors, the students who by then have become our strongest engineers and our most capable leaders. Over the summer a group of freshmen and sophomores arrives with almost no experience. Everyone who stays moves up into a role they have never done. Then in January the season starts, and we have roughly ten weeks to design a robot, build it, test it, practice with it, and be ready to operate it under real pressure at a competition against teams that did the exact same thing.

So we run a startup that is structurally required to rebuild its leadership layer from scratch every year, on a fixed clock, staffed entirely by volunteers who happen to be teenagers.

That sounds like a disadvantage. Over time I have come to see it as the most honest teacher of organizational design I have worked with. When you cannot lean on the same people season after season, you find out very fast whether your results come from a handful of strong individuals or from something that outlasts them.

One of our team values is to SUSTAIN, to leave the team stronger than we found it. Over the past few seasons that one word has become the way I think about the whole problem, because it spells out the seven tests we hold ourselves to:

• Standards Before Scores
• Use Every Strength
• Scale Leaders, Not Heroes
• Transfer Ownership, Preserve Stewardship
• Apprentice Leaders Before the Pressure Arrives
• Inspect the System Honestly
• Next Leaders Are the Metric

S: Standards Before Scores

We borrow our operating philosophy from Bill Walsh, the coach who turned the 49ers around and wrote "The Score Takes Care of Itself." The idea is simple and hard to live by. You do not chase the result. You obsess over the standard of execution, the attention to detail, the way each role is performed, and you let the score follow. The moment you optimize for the outcome directly, you start cutting the corners that produced it, and you cannot sustain it.

There is a version of a robotics team that lives off a great season from years ago. A strong group of individuals delivers once, the trophies go on the shelf, and the team coasts on that reputation while the habits quietly decay. I have watched it happen. It is the BlackBerry problem in miniature: still winning on paper, already losing on fundamentals.

The teams that repeat are the ones where the standard is the culture, not the scoreboard. And that culture has to be inclusive by necessity, not sentiment. On a robot, one person's mistake shows up on the field for everyone. Every role matters, every person is worth developing, and the standard is the same whether you are the lead or the newest member. High standards, real execution, and genuine respect are not three things. They are one thing.

U: Use Every Strength

I often describe what we do using Formula 1, because it captures something people miss about engineering excellence: it requires very different people, in very different roles, working on very different timeframes. There is a factory where engineers design the car. There are craftsmen who make complex parts fit and meet spec. There is a trackside crew who prepare, fix, and operate the car under pressure with no time to think. Those are not the same skill, and you cannot staff them with the same type of person.

Our team works the same way. Scouts watch every match and analyze the other robots. Strategists turn that data into a plan for each alliance. The drive team executes and adjusts inside the match, live, against opponents doing the same. Design, build, and operation each demand a different temperament and a different sense of time.

This is where diversity earns its place. If you need different skills in different roles to reach the highest level, then failing to find and use everyone's particular strength is leaving performance on the table. A team that only knows how to value its loudest builders is wasting the people who would have made it great.

S: Scale Leaders, Not Heroes

The default move, in robotics and in industry, is to hand the top leadership role to the strongest engineer. It is the same instinct that pulls a founder toward being the hero who carries everything. That works when you need quick results, and in a crisis it is often the right call. I have made that call myself more than once. But when you are building something meant to last, it backfires, for the same reason chasing the score does. The chain is predictable. The strongest engineer becomes the leader, hits a bottleneck, and says the words every one of us has said: it is faster if I just do it myself. Work concentrates in a few hands. Those few burn out and start to resent that no one is helping. Meanwhile everyone else feels pushed aside and stops showing up. And because this is all volunteer time, "stops showing up" is not a metaphor. The workshop empties out, the remaining few work longer hours, mistakes go up, and you end up with a worse robot and a weaker team.

A few years ago we pivoted. We now select leaders far more on their ability to organize, motivate, and steer people than on raw engineering skill, while balancing that against everyone's real desire to shape the robot itself. That balance makes the job harder than it used to be, not easier. We are not looking for the superhero leader it is so tempting to try to be. We are looking for the leader who is at their best when they make everyone around them better, whose own success is indistinguishable from the team's.

We wrote this into our handbook last season, and it now sits at the center of how we appoint leaders, knowing the team loses its top layer every May.

T: Transfer Ownership, Preserve Stewardship

Mentors do run the team. What we do not do is the day-to-day execution. We operate on three levels, and they carry very different amounts of authority.

The first is coaching. We transfer experience to the students, sometimes as knowledge, though they already know a great deal, and just as often by challenging them, questioning a decision or a design so they have to defend or rethink it. That work is always pointed at the student growing, never at getting to the answer faster.

The second is continuity. Mentors typically stay more than four years while students cycle through in four, so we are the team's memory and its long horizon. We make sure decisions hold up across seasons, whether that is finances, training programs, or projects that have to be maintained long after the students who started them have graduated.

The third, for lead mentors, is a veto. It works a lot like a board seat backed by a majority stake. We use it rarely, and we reserve it for the things that outlast any single season: the health of the culture, the relationship with the school and our other stakeholders, and multi-year financial sustainability. A concrete example: the students own their season budget, but we expect it to be historically grounded and backed by a real plan to raise the money, because our 501(c)(3) funds the season up front. They understand that a budget has to balance, and that they can only ask for more when they have a plan for it. If they cannot, we can decline. The veto is not a club. It is a standard, made visible.

The hard part is holding two things at once. We cannot tell students they own the team and then quietly reverse every decision we do not like. But ownership also cannot mean handing teenagers risks whose consequences will still be here long after they have graduated. The goal is neither control nor abdication. It is stewardship.

A: Apprentice Leaders Before the Pressure Arrives

Almost none of our leadership development happens during the season. The season runs from January into May: the first ten weeks are the build, and then the competitions carry on until the last one, which we hope is the World Championship. That whole run is when everything has to come together under pressure. The leaders themselves are formed earlier, in the years before and in the off-season, where we use small project-driven work to put people into the model for real, so they can prove it to themselves before it counts.

The clearest evidence that this works is our junior varsity team, 9584. We built it partly as a place where newer students could practice the model early, and partly to set the standard for what we expect. Three years on, the students who came up through 9584 are now more than half of our leadership, and they set the bar for who gets selected onto 9584 next. The pipeline is producing the culture, and the culture is staffing the pipeline.

The reason we start this early is that some lessons only arrive on a delay. Delegation sounds inefficient until you have become the bottleneck. Clear roles sound bureaucratic until two people each assume the other one had it. Succession planning sounds premature until the one person who understands a system graduates and takes it with them. You cannot teach those in ten weeks. You can only give people enough seasons to feel them.

I: Inspect the System Honestly

The hardest problem is seeing students accurately, and it is structural, not a matter of effort. As volunteers with day jobs, we cannot be present all the time. A lot of our read on a student comes from objective results and from peer feedback, not from watching them directly. I ran a team in India from the United States during COVID, and I learned then what it is to assess someone through a narrow tube, seeing a small slice of what they actually do. It can never be fully fair. The most useful thing I did was stop pretending otherwise and say so plainly. When you cannot watch everything, your teammates become your instruments, and your standards have to be legible enough that people can hold themselves to them when you are not in the room.

That partial view makes two kinds of students especially hard to place. There are the over-committers, who take on far too much and then struggle to finish as their interest shifts or the work turns out bigger than they thought. And there are the quiet ones, genuinely skilled, who believe they are important contributors but wait to be called on and grow frustrated when the call does not come. They look like opposites. They share a root: a gap between how someone sees themselves and what the team actually needs from them. Our application process has to tell the student who presents well but does not always deliver apart from the introvert who is excellent and just needs a real setting to show it. We are not there yet. We used to run closer to sink or swim, and we have moved away from it, because for a large group it did not work and could turn toxic and needlessly competitive.

The second open problem is teaching leadership to people who have not yet lived the experiences that make the lessons obvious. It is hard to explain why something matters to someone who has not felt it go wrong. We are addressing it by defining more explicit individual roles so the expectations are concrete, but it is still very much a work in progress. Some of it can only be learned by failing at it, and that puts the responsibility on us as mentors to hold open a space where controlled failure is safe. A robot that does not work out can be one of the best things to happen to a group, as long as it sets them up to succeed the season after. A bad robot is not always a bad outcome.

The third is that we may have moved burnout rather than removed it. The old model burned out a few heroes. The new one loads coordination and standard-setting onto our people-leaders and mentors. We watch for it, we back it with leadership training, and we now keep an understudy for every key role so the load can be shared and handed off. It is a real improvement, and not yet where I like it to be.

N: Next Leaders Are the Metric

The turnover I started with, the thing that looks like our biggest weakness, is really the point. Because we lose our leadership every year, we cannot fool ourselves for long about whether we are building a team or renting one. Talent can win you a season. Only culture wins you the next ten, and culture is just the set of standards that survive the people who set them.

So the question I have started asking of every team I am part of, robotics and professional alike, is not whether we are winning right now. It is whether we are building leaders faster than we lose them. For a robotics team the answer comes back on a fixed date. Ours arrives every May, when the seniors we spent years building walk out the door.

None of this is mine alone. It reflects years of work with a dedicated group of volunteer mentors who show up season after season, my co-lead mentor Leonard Speiser among them, and the team is far better for them.

I went deeper on this in two workshop talks with the team: one on how we reworked our leadership structure, and one on why and how we started the JV team. Both are a few seasons old and show where a lot of this thinking started.