Your Ideal Customer Is Hiding in Your Data

Two weeks ago I told you bonding is close to binary — you either build a durable core of customers or you don’t. The obvious next question is why: what decides which customers bond?

The first answer is one you can read before a customer ever logs in. It’s fit — not fit as a vibe or a persona slide, but fit as a measurable fact: is this customer the kind of buyer the product is actually built to serve?

You can see fit in the behavior

Take one company and split its customers by whether they matched its real profile. The ones who fit bonded almost completely. The ones who didn’t collapsed — no core at all.

Right-fit customers bonded into a ~92% core. Wrong-fit customers collapsed — about one in five left within three years. Same product, same onboarding, same success team. Opposite results — and the split was set before anyone logged in.

The biggest lever is pulled at the sale

That inverts how most teams spend their retention energy. We pour effort into saving accounts after they’ve started slipping. But the largest single lever on whether a customer bonds was pulled the moment you decided to sell to them. A wrong-fit customer isn’t a customer-success problem. It’s an acquisition decision that customer success then gets blamed for.

Weak overall doesn’t mean no fit

When a company’s overall loyalty number looks weak, the reflex is to conclude the product doesn’t have product-market fit. Usually that’s wrong.

In about one in four companies with a weak overall number, one slice of the base bonds strongly — real, durable product-market fit, hiding inside a mediocre average.

The product isn’t failing. The company is averaging the customers it fits together with the customers it should never have sold to, and reporting the blur.

Your ICP is a query, not a workshop

This is the single most actionable thing in the whole study. You almost certainly don’t need to reinvent your product to find fit. You need to find the customer type that already bonds, name it, and point acquisition there — and stop diluting the number, the roadmap, and your team’s attention with customers who were never going to stay. Your ICP is not a workshop exercise. It’s a query.

Fit shows up at the sale

And fit shows up earlier than most people think — even in how a customer buys. When we put company size and deal size head to head, raw size washed out and the commitment signal survived: serious customers stay longer, not bigger ones. A buyer who puts real money and real intent on the table is telling you they mean to get a result. That intent is the leading edge of fit — visible before the first login.

This week: stop guessing

Fit gets the right customer in the door. It doesn’t finish the job — a well-fit customer who never reaches a result still leaves, and that second lever is where most of the real work lives. But that’s next week’s argument. This week’s is simpler and more urgent: stop guessing at your ICP. It’s already written, in the retention behavior of the customers you’ve already had. Go read it.

The full picture — how much fit explains, what separates the bonders from the leavers, and the customer type everyone chases that turns out to be the least bonded of all — is in the report, The Science of Customer Bonding.

→ Read The Science of Customer Bonding

— Greg