I've run 11 cohorts across 3 years. For the first two, I was mainly managing what was in front of me — applications, meetings, sessions, feedback. Making sure things didn't fall apart. That's running a program.
Building a program is different. Building is when you start asking: why is this task recurring? Why does this always take three hours when it should take 30 minutes? Why are we still using a spreadsheet for this?
The moment you cross the line
I can tell you exactly when I crossed from running to building. It was during our third cohort, looking at a list of 300 startup applications in a spreadsheet that I had to sort manually. I thought: this is not a tool problem. This is a problem of how we've set up the workflow.
So I built an internal scoring and filtering tool. A simple web form connected to a database, with automatic scoring weights and a review dashboard. Cohort prep went from a full week to two days. That moment — when you look at your own work and say "I could build this out of existence" — is the moment you start building instead of running.
The trap program managers fall into
The trap is operational comfort. When you're good at running a program, it feels valuable. And it is. But if you're doing the same tasks repeatedly without questioning whether those tasks need to exist, you're leaving the most important work undone. The best operators I know treat their own workflows the way engineers treat code: constantly looking for what can be abstracted, automated, or eliminated.
What it looks like in practice
Building a program means: documenting recurring decisions so you don't make them from scratch every cycle. Creating systems for things that happen every cohort — scoring, onboarding, comms, reporting. Building tools when a tool would obviously help and nothing exists. Writing playbooks so the program can survive without you.
The goal is that the program works better over time — not because you're working harder, but because the system is better.
Why it matters for innovation programs specifically
In innovation and accelerator programs, the pressure to execute is constant. There's always a deadline — a pitch day, a cohort close, a board update. It's easy to spend all your time in execution mode. But execution mode is a trap. The programs that scale are the ones where someone paused the execution and asked: what are we actually building here? And then went and built it.