The Unsexy Work of Systems

There’s a moment in every serious project where the fun stops and the systems begin. The prototype worked — it was delightful, it demos well, the demo is a hit. And then the demo needs to work every time, for everyone, at any hour, and the delightful prototype turns out to be a beautiful seed that needs a mountain of unglamorous infrastructure to grow.

This is the moment most projects fail, and it’s not a technical failure. It’s a taste failure. The team’s taste was calibrated for the demo, and the systems don’t feel like the kind of work good taste should be spent on.

What systems actually are

A system is a promise you don’t have to think about. The log that’s always there. The schema that doesn’t lie. The deploy that doesn’t terrify. The alert that fires only when something is actually wrong. The name for a thing that the whole team calls by the same word.

None of it is clever. All of it is agreed. A system is a solved argument — the team took a decision, wrote it down, and moved on to something better than re-arguing it.

The unsexy work is the work of taking decisions out of the system so the remaining decisions can be made well. Everywhere you leave ambiguity, you leave a tax: someone re-derives the answer, someone contradicts the answer, someone’s career is spent defending the wrong answer because nobody ever wrote the right one down.

Boring as a compliment

When an operations person says your system is “boring,” it is the highest possible praise. It means nothing is on fire, nothing needs babysitting, and the nightly batch job will finish before anyone wakes up. Boring is what reliable feels like from the inside.

I’ve learned to chase boring. When I’m tempted by clever — a new queue, a new event-sourcing layer, a microservice that exists because the last one was embarrassing — I ask what the boring version is, and I build that instead. The boring version is never the wrong version. It’s the version that lets me sleep, which frees my working hours for the work that actually needs them.

The payoff is compounding

Systems are the only engineering investment I know that pays compound interest. A good naming convention saves five minutes a day forever. A good deploy pipeline saves an hour a week and a panic a month. A good incident process turns a crisis into an agenda item.

The compounding works the other way too. Every piece of duct tape is a debt with interest. Every undocumented decision is a future meeting. Every special case is a trap for the person who joins next year.

The teams I admire most are not the ones with the cleverest architecture. They’re the ones whose architecture is so settled, so agreed, so unremarkable that they spend their energy on the product instead of on the plumbing. That’s the whole game. The system is not the point — the system is what makes the point possible.

The unsexy work is the work that makes everything else sexy. I’d rather do the unsexy work than enjoy the credit for work that can’t actually ship.