The system you think you have

I wrote a comment in my own code this week. Four words: "the only entry point."

I was wrong. There were four entry points. I had written the other three myself, at different moments, for different reasons, and then stopped thinking about them while the first one became the canonical version in my head.

This is not a software problem. It is a founder problem.

You think you know where your customers come from. You have a number in your head, some source that feels primary. Meanwhile three other channels are quietly delivering people you never attributed, because you built those paths a long time ago and forgot about them. The mental model calcified. Reality kept moving.

You think you know your cost structure. You have a rough sense of what runs the business each month. But the actual bill, when you sit with it line by line, is different. Not dramatically. Enough to matter.

You think you know how your best customers use the product. You watched a few early ones. But the cohort that actually stuck around -- that renewed, that sent referrals -- might be doing something slightly different from what you optimized for. You have not checked lately.

The intimacy problem

The entrepreneur's curse is that you built the thing, so you feel like you know it. You watched it come together. You made every decision. That intimacy becomes a liability. Builder's confidence is earned, and it is also dangerous, because confidence does not update automatically.

Reid Hoffman says if you're not embarrassed by your first version, you launched too late. I think that's right. But there's a second half nobody quotes: after you launch, the teaching does not stop. The version of the business in your head keeps lagging the one on the ground.

The founders I've watched stumble in year two or three didn't usually build the wrong thing. They stopped checking. They ran the old map in a new territory.

The antidote is boring. Look at the actual numbers, not your memory of them. Read the actual output, not the summary. Walk the actual path a customer takes, not the path you designed. Once a month is probably enough. Once a quarter is not.

The gap between the system you think you have and the one you actually have is not a crisis. It's maintenance. The crisis is discovering it by accident, late, under pressure.