Every word, every meeting

Leadership is mostly small moments

Most writing about leadership is about strategy, org charts and vision decks. In my experience, a team decides whether to follow you based on much smaller things: what you say, how you say it, and whether you show up the same way every day.

The questions I hold myself to:

• What are the first thirty seconds you say to your team after someone is let go? Those words tell everyone in the room whether they’re safe, whether you’re honest, and whether you saw it coming.
• What tone do you set in the first meeting after hard news? Whether it’s a rough quarter or a reorg, the team takes its cue from you before they take it from the announcement.
• When someone on your team doesn’t make it, how much of that do you own? Was it your job to make them successful? Usually, at least partly, yes.
• Does your team know you’re behind them? Not in a speech. In how you handle it when one of them gets blamed for something in a meeting they weren’t in.
• How do you talk about your own work? If you’re vague about what you’re doing, your team will be vague too.
• Are you decisive? A team can recover from a wrong call. It struggles to recover from no call.
• How helpful are you, one person at a time? Trust is built in the five-minute unblock, not the all-hands.

None of these show up in a job description. All of them decide whether people trust you.

What running my own company taught me

Before data, I spent three years running my own company. Three lessons from that stay with me.

Pick a direction and move together. In every meeting, I knew what I was talking about and I had a direction for us. We weren’t always moving in the right direction, but whatever direction we moved in, we moved in together, quickly, and without second-guessing. A team that’s aligned and wrong finds out fast and corrects. A team that’s unsure never finds out anything.

You can’t do it alone. The first few hires on any team are its pillars. Not everyone has to be perfect, but you need a couple of people beside you who can lead on their own and hold up their part of the work. Getting those hires right matters more than almost anything else you’ll do.

Keep learning, because the ground keeps moving. A founder who stops learning gets passed. That’s even more true now, when the tools change every few months.

Hiring is where it starts

Because hiring matters that much, I treat it as the highest-leverage hour of my week. In one year I interviewed nearly thirty candidates across five roles, from analysts to the director who now leads the department, with a written scorecard for every one.

The most useful thing I did there wasn’t a single interview. I noticed that panelists who heard each other’s opinions first tended to agree with whoever spoke first. So I proposed that everyone submit feedback privately before any discussion, and the team adopted it. I also pushed for an explicit hiring bar as we scaled, because a bar that only lives in people’s heads drifts.

Leading a team into the AI era

The hardest leadership job right now is taking a team through a change in how work gets done.

• Go first. I moved my own work onto AI before asking anyone else to. That way, when I describe the workflow, I’m describing my week, not a vendor’s demo.
• Set the outlook. A working prototype in two weeks isn’t the finished product. But when the starting line moves that much, the ceiling for what a small team can build in a year moves with it. I want my team to believe that, because they’ll build up to the size of what they believe.
• Build the infrastructure yourself. The context layer I built for our team, a shared, reviewed memory that every analyst’s AI reads, is a leadership act as much as an engineering one. It’s how a team’s knowledge stops belonging to whoever has been around longest.
• Meet people where they are. Some people adopt new tools on day one. Some take real pride in doing the work by hand, and that pride is earned. You don’t move the second group with a mandate. You move them with a fast, real win.

Reach, not headcount

Our data team is small, so my scale isn’t the size of my org. It’s reach. I lead analytics for two teams. I directly manage two analysts and coach others less formally. I’m embedded with clinical operations, provider operations, physician experience, customer experience, marketing and product. I run the weekly debriefs, office hours and quarterly planning with those leaders, and I write the specs, thresholds and data contracts our data engineers build from.

The best measure of a lead is what happens to a system after they hand it off. I coached one of my analysts from onboarding to leading the next major version of one of our production AI systems. The handoff was a written hub: how the system works, why every rule exists, what had been tried and failed, and who to ask about what. A handoff you can’t write down isn’t really a handoff.

What it comes down to

My team trusted me because I worked hard, I looked out for every one of them, and I showed up to every meeting knowing what I was talking about and where we were going. That’s not a personality trait. It’s a set of habits anyone can practice: every word, every meeting.

The goal was never a team that needs me. It’s a team that trusts each other, moves in one direction, and can build the next thing without me in the room.