When the system finally knows what it knows

There's a moment in every business system where you realize a number you've been treating as a rough estimate actually needs to be exact. Not because you were lazy before. Because you didn't know the edge cases until real use revealed them.

I hit that moment yesterday across a few different projects at once. Decimal amounts that needed to be stored cleanly. Prices that needed to reflect what actually happened — not what we projected would happen. Display data that needed to stop being approximate.

None of it was glamorous. But all of it mattered.

The cost of tolerating almost-right

When you're building fast, you make simplifying assumptions. You say: whole numbers only, three choices only, one shape of data only. That's not sloppiness — that's appropriate restraint for the stage you're at. The problem is when the system grows and those assumptions quietly start costing you. A leaderboard that rounds wrong. A price column that shows the wrong figure. A dropdown that covered 80% of the real world but not the other 20%.

Each of these is small. Together they erode trust — from your users, and from yourself.

There's a Jeff Atwood line I keep coming back to: the best code is no code at all. I think about it differently than he meant it. To me it means: the best complexity is the complexity that was always there, finally acknowledged. You didn't add something. You stopped pretending the world was simpler than it is.

That's what precision work actually is. Not over-engineering. Not gilding. Just honesty about the shape of the problem.

Yesterday I also launched a new storefront. Clean, modern, small. It would have been easy to rush it out broken, or to over-build it into something that took three more months. I did neither. It's live. It handles the basics correctly. The edge cases I know about are handled; the ones I don't know about yet will surface in use.

That's the job. Not perfection on day one. Not indefinite refinement in private. Ship what's honest, then fix what reality reveals.

The same principle applies to every part of running a solo business — pricing, contracts, how you respond to customers, what you promise in a proposal. The temptation is to keep things vague because vague feels flexible. But vague usually means you're deferring a decision that will be harder to make later, under pressure, when someone is waiting on the answer.

Get precise now, while it's quiet, while you still have room to think.

The cleanup is never the exciting part. But the thing that works correctly is always worth more than the thing that almost works.