Find the Constraint

Spend enough time around factories and you stop believing in general improvement. A manufacturing line has exactly one station that sets the pace for the whole thing. Speed up any other station and you've accomplished nothing — you've just piled up more work in front of the slow one. Until you fix the bottleneck, nothing else you do matters.

The same thing is true of almost every system I've worked on, including the ones made of people.

One thing is always the limit

A data pipeline has a stage everything waits on. A company has a function that everything else backs up behind. A team has the one person every decision routes through. At any moment there's a single constraint deciding how fast the whole system can move, and it's usually not where the noise is. The loudest problem and the binding one are rarely the same.

This is freeing once you accept it, because it means most of the work you could do doesn't matter right now. One thing does.

Why we optimize everything else

We avoid the constraint because the constraint is usually hard, political, or boring. It's the senior person who won't delegate. The approval step no one wants to touch. The legacy system everyone's afraid of. So instead we improve the things that are easy and visible, and feel busy while the system runs exactly as slow as it did before.

I've done this myself — tuned the fast part of a pipeline because it was satisfying, while the actual bottleneck sat untouched because fixing it meant a conversation I didn't want to have.

The discipline

So before I improve anything now, I make myself answer one question: what is the single thing setting the pace here? Not the worst thing. Not the thing that annoys me most. The thing that, if it got faster, would make the whole system faster — and then I check that nothing else can move until it does.

Find that one thing and pour your effort there. Everything else is motion, and motion isn't the same as progress.