The first guess is supposed to be wrong

A pile of projects that happen to make money isn’t a business, and from the inside the two look almost identical. Every project has a customer on it and every one has revenue attached.

I spent a stretch working alongside a software division built exactly that way. It kept producing little software products. I’d find another one every week or two: a spreadsheet somebody had been running for a year turned into a standalone application, something built for one customer turned into a thing a salesperson was now supposed to sell, an internal tool became enterprise. Each one arrived with code and some amount of infrastructure, which meant support, security, maintenance, anything you’d find in a platform. There were eight or ten of them. The company had the bandwidth to support two.

Software only pays well one way: build it once, sell it many times, so the cost of serving the next customer rounds down. With one customer it’s inverted. Say one of these products brings in $50,000 a year, and keeping it alive runs 400 hours: support, security patches, a dependency that moved, a customer who needs it to talk to something new. At $100 an hour that’s $40,000 of engineering to hold onto $50,000 of revenue, and the interesting number is the 400 hours. Now run it eight times. Thirty-two hundred hours a year, better than an engineer and a half, spent entirely on keeping yesterday alive. Little apps add up to a massive drain.

So why build the next one at all?

I figured the guy running the division liked building things, a hobbyist with a budget and an engineering organization. From a distance it looked like somebody screwing around with company money, and a few people I talked to had the same feeling.

It was possible. It was also wrong. I’d made it about a person’s character, which is the laziest place a diagnosis can come to rest.

Keep asking

I kept asking. Who owns this one? Where did that one come from? Why did a spreadsheet become an application? Is anyone actually selling it? When five people tell me five versions of the same event, the disagreement is the data.

Eventually I found someone who still carried a piece of the company’s software history. As it turns out, the person running the division had been told, by someone well above him, to go make money with software.

That was the whole thing. Kinda wild, in hindsight. He was doing what he’d been told to do.

One piece of information changed the meaning of everything around it. That’s the moment you’re working toward. Not certainty. Coherence.

The first guess isn’t the diagnosis

There’s a comfortable idea that good judgment means withholding judgment until you have enough information. Inside a real company it leaves you dead in the water for months. You’re never going to have all the information. People remember the same meeting three different ways, the CRM tells one story and the financials tell another, and half the decisions that matter were made verbally by people who don’t work there anymore.

So form an explanation early and hold it loosely. Then ask what else would have to be true if it were right. If he’s building pet projects because he enjoys it, I should find projects clustered around his own interests, no commercial process anywhere near them, and people who tried to stop him. Now go look. The hypothesis gives you somewhere to walk. The evidence tells you whether to keep walking.

The closest thing I know to this is gradient descent. You don’t need to know where the lowest point is before you move. You take a step in the direction that looks lower, see what you learned, adjust, step again. I’m forming opinions immediately and using the next question to try to kill them.

Companies are machines made out of people

As a portfolio strategy, what I was looking at was indefensible. Seen through the eyes of a man who’d been told to go sell software, it was the obvious thing to do. Leave the people out of the machine and perfectly rational behavior looks like nonsense.

That changes what you fix. If my first theory had been right, the fix was governance: control what he’s allowed to build. The real question sat one level up. What was the software business supposed to be? A software company? A place to grow products into businesses of their own? Something that existed to make the consulting business more valuable? Those are different companies. Nobody had written down which one this was, so people filled in the blank themselves, reasonably, with the last clear instruction they’d been given.

As far as I know, nobody ever answered it. I raised it. Then I moved on, and maybe I didn’t push hard enough.

Sometimes you find the wrong bottom

Gradient descent has a known flaw. You can land in a local minimum: follow the evidence downhill, arrive at an explanation that fits everything you’ve looked at, and still be wrong. Talk only to sales and the sales explanation gets more convincing every hour.

AI makes this easier to fall into. A language model takes whatever you hand it and builds a beautifully coherent answer, good enough to feel like magic and stop the investigation. It can’t interview the person nobody thought to include, and it doesn’t know about the instruction given verbally two years ago. So every so often you have to jump. Talk to the customer nobody suggested. Ask somebody junior what everyone senior takes for granted. Ask what would make this apparently irrational behavior rational.

Being wrong is cheap from outside

The advantage experienced people have is better guesses, and they’ve been wrong enough times that dropping a theory doesn’t feel like losing anything. I think it’s this. Let’s see. Nope. Then maybe it’s this. Let’s see.

That’s hard to do from inside your own company. Everyone in the building has a stake in the current explanation. They wrote it, or inherited it, or their budget was approved on the strength of it. So the first guess gets defended instead of tested, and the investigation stops at the first answer that lets everybody stay who they are.

The first guess is supposed to be wrong. Its job is to get you moving.