“Teams completely lose track of who owns what,” he said, and he said it the way other people announce that their child has been accepted to a good school.
It was a Tuesday evening at a small gathering for the volunteer thing I do, the kind of evening where everyone drinks lukewarm coffee out of cups that bend when you hold them. Nobody in that room was supposed to be talking about software. But he had found out, somewhere between the coffee and the snacks, that I do marketing, and that turns out to be enough.
So he told me. And he was thorough. Tasks falling through the cracks. Two people building the same thing without knowing. A deadline missed because all four of them assumed one of the other three had it. He had examples, he had dates, he had a little name for the problem. He had talked to people, lots of people, and they had all nodded. Yes, this happens. Yes, it is annoying.
And everything he said was true. People do lose track of who owns what. I have sat in exactly those situations, and so have you, and so, in a small way, had the room we were standing in, where there were two bowls of identical crisps because nobody knew who had brought the snacks. But true is a very low bar. The world is stuffed with things that are true and broken and that no human being will ever pay a cent to fix. My bike has made the same noise for two years. The 17:42 to Amsterdam is broken in a way I could describe to you in some detail. The fact that something does not work is the easiest thing in the world to confirm, precisely because almost nothing fully works.
What he had described was the external problem, and that is a real and useful thing. It is the part that breaks out in the open, where you can point at it, put it on a slide, and get a room to agree that yes, that is broken. You need it. It is where every real problem first shows itself.
But it is also the surface, and he had mistaken reaching the surface for reaching the bottom. He had confirmed the thing was broken and decided the job was done. Broken and worth fixing are not the same sentence. He never went near the harder question, the one that does not feel nice to ask, which is whether this costs anyone enough that they would reach into a budget to make it stop. That answer does not sit on the surface. It sits underneath, in the layers a polite nod at your demo will never show you.
That is the trap. The visible part is the only one that says yes to you without being asked twice. The layers under it make you work, so they get skipped.
He was still going, on to the second version of the product, the one with the extra thing.
All of it mattered. Nobody was ever going to build software for it.
He had found the part you can point at. He never once looked under it.