Fix the foundation before you scale anything
The work you avoid at the start doesn't disappear — it just waits to cost you more later.
Yesterday I spent a long stretch fixing something that should have been right from the beginning. A timing assumption, buried deep in a piece of work I run, had been wrong for weeks. Not catastrophically wrong — the thing was still running, still producing output. But wrong in a way that made every downstream number quietly unreliable. Every decision built on top of it was built on sand.
The fix itself took less than a day. The cost of not fixing it earlier? Harder to calculate. Weeks of data I can't fully trust. Decisions I'll need to revisit. A subtle anxiety that had been sitting at the back of my mind, unaddressed, while I kept building new things on top of the problem.
That's the trap. When something is mostly working, the pull to keep moving forward is enormous. You tell yourself you'll come back to it. You have customers to think about, features to ship, revenue to chase. The foundation isn't glamorous. Nobody sees the foundation.
The debt you don't acknowledge
There's a version of this in every business, technical or not. The invoicing process you jury-rigged in month two and never revisited. The pricing model that sort of works but has a flaw you've been papering over. The onboarding flow you know confuses people but hasn't caused enough visible damage yet to force action.
These aren't small things waiting patiently. They're compounding. Every new customer you add, every new process you layer on top, every new hire you make — all of it inherits the flaw. You are scaling the problem.
Knuth said premature optimization is the root of all evil, and he was right. But there's a mirror image that doesn't get named as often: deferred correction is its own kind of evil. The difference between the two is whether you actually know something is wrong. Optimizing before you understand the problem is waste. Ignoring a problem you've already identified is something else. It's a choice, and it has a price.
I fixed the timing bug. I cleared the bad assumptions. I rebuilt on something I can actually trust now. It wasn't exciting work. Nobody will notice it happened. But this morning, for the first time in a while, I'm not carrying that low-grade unease about the numbers underneath everything else I'm doing.
Sometimes the most important thing you ship is invisible — because it's the thing that makes everything else true.