The most expensive assumption you're making

Last time I argued that your product isn’t your business — that the durable value sits in the operating system underneath it. This piece is about the single most expensive way that plays out in practice.

It’s an assumption. A specific one. And it’s so natural, so reasonable, that most people don’t even notice they’re making it:

We already know what our customers’ problem is.

You probably do know something. You’ve been doing this for years. You talk to customers. You have instincts earned the hard way. But knowing something about your customers’ problem is not the same as understanding it — and the gap between the two is where good companies quietly build the wrong things.

Assumption feels like expertise

That’s what makes it dangerous. When you assume you know the problem, it doesn’t feel like a guess. It feels like experience. It feels like exactly the kind of judgment you’re supposed to have.

So you skip a step. You move straight from “I know the problem” to “here’s the solution” — more features, a new service line, a faster version of what you already do. And because you’re skilled, the solution is usually competent. It’s just aimed slightly wrong. You’ve optimized the wrong thing, beautifully.

The cost shows up later and indirectly: features nobody adopts, a service that’s hard to sell because it answers a question customers weren’t really asking, effort spent polishing a symptom while the actual cause sits untouched.

Real understanding lives in your operating system, not your intuition

Here’s the good news, and it’s the throughline of everything I write: you are probably already sitting on the evidence.

Your processes and your data quietly record what customers actually do — as opposed to what you assume they want, or even what they tell you they want. The truth is in the patterns:

• Where they churn. The moment people leave tells you what they couldn’t get.
• What they contact support about. Repeated questions are a map of where your offering and their reality don’t line up.
• Where the funnel leaks. The step where people abandon is the step where the real problem lives.
• What they actually use. Usage is the most honest customer interview you’ll ever run.

The old line is that customers wanted “a faster horse.” But if you’d watched how they actually traveled — the data on their journeys — you’d have seen the real job: get somewhere reliably, without the animal getting tired. The assumption (“they want horses, but quicker”) points you at breeding. The evidence points you at the engine.

The pattern, in one small example

A company is convinced its customers want more features, so it keeps building them. Adoption is mediocre, and they assume the features just aren’t good enough yet. But the data tells a different story: most new customers abandon in the first ten minutes, before they ever reach a feature at all. The real problem was never feature depth. It was the first ten minutes.

No amount of expert intuition surfaced that — because intuition was busy being confident. The process data surfaced it in an afternoon.

The method: follow the number to the problem

This is just the numbers-first habit applied to customers. Don’t start with the solution. Start with a number that’s behaving oddly — a churn rate, an abandonment point, a support category that won’t go away — and follow it back to the process that produces it. The real problem is almost always waiting at the end of that trail, and it’s almost never the one you assumed.

Your takeaway

Here’s a question worth sitting with this week:

What do I believe about my customers that I’ve never actually checked against data?

Write down two or three of those beliefs — the ones that feel too obvious to question. Then check one of them against what your systems actually record. Either you’ll confirm it, and now you know instead of assume, or you’ll find a much better problem to solve.

Next in the series: the function in your business that holds most of this evidence — and why you’ve probably been treating it as a cost to cut rather than the intelligence centre it actually is.

---

Part of a four-part series on the operating system beneath a business:

1. Your product is not your business
2. The most expensive assumption you’re making — you’re here
3. Finance is not a cost centre
4. What intelligence actually buys you

← All posts