Running is not the same as working

Yesterday I found something that had been offline for nearly two weeks. Not crashed - crashed things announce themselves. Just quietly disconnected, doing nothing, while I assumed it was doing its job.

Nobody flagged it. The system that was supposed to alert me if it went quiet had its own silence gap. The thing watching the watcher had a blind spot. I found it by accident, looking at something else.

The embarrassment isn't that it broke. Things break. The embarrassment is the two weeks between when it broke and when I noticed - and the honest question that follows: how many other things are "running" that aren't working?

Reid Hoffman says you're launching too late if you're not embarrassed by version one. I believe him. But I think there's a second half to that: the embarrassment of version one is table stakes. The harder discipline is staying embarrassed - staying willing to look, long after the launch adrenaline has faded and the system is supposed to just run.

What you stop checking

There's a specific pattern in how founders lose track of things. You build it, you test it, it works - so your attention moves to the next problem. The thing you just shipped disappears into the background category of "running." You check it when something looks wrong. Which means you don't check it when it's quietly not working, because not working and working can look identical from a distance.

This isn't a software problem. Every business has a version of it. An email sequence that stopped sending months ago. A sales process everyone says they follow that nobody actually follows. A customer touchpoint you designed carefully and then never went back to verify was doing what you thought.

The fix isn't more monitoring. More monitoring just means more things that can quietly break. The fix is building the habit of honest inspection - not "is the light green?" but "did the thing actually happen?"

There's a difference between evidence that something is running and evidence that something is working. A lot of what we call "running our business" is really just checking that the light is green.

Yesterday I spent several hours finding the gap between running and working, across more than one thing. I am not proud of how long some of those gaps had been there. I am glad I found them on a Sunday, looking, rather than from a customer who noticed first.

Look before they look for you.