How Many of Your Customers Will Never Leave?

Last week I said there’s a single number underneath retention that a churn rate, a renewal rate, and a satisfaction survey all miss. Here it is.

Every metric you already track answers the same question: how many customers are still paying? There’s a better one: how many are never going to leave? Those are not the same question, and the gap between them is the whole game.

Retained today isn’t retained for good

A customer can be “retained” and still be on the way out — paying through a contract, coasting on inertia, quietly evaluating alternatives. Another customer is bonded: the product has woven itself into how they operate, and leaving would cost them something they aren’t willing to pay.

The first is retained today. The second is retained for good. Only the second one builds anything that lasts — and you cannot tell them apart from a churn rate, because this month they are both still paying.

The number hides in the shape, not the rate

You can only tell them apart from the shape of retention over time. Follow a company’s customers from the day they sign up and plot the share still there. The line falls as the leavers leave — and then, if a durable core exists, it stops falling and flattens onto a floor.

The height of that floor is the share of customers who are bonded. I call it the PMF Score — the one retention number that can’t be faked by a good quarter or propped up by a long contract, because it’s defined by who is still there long after both are gone.

Every customer base is really three populations

• Bad-fit — the product was never going to serve them. Gone fast.
• Still-leaving — paying now, but sliding toward the door.
• Bonded — the core that stays for good.

In the average company: ~6% bonded. ~92% still leaving. Most of what a healthy retention rate measures is customers who simply haven’t left yet.

Across a market, it splits in two

Measure that bonded fraction across a whole market and something strange happens. Companies don’t spread evenly from “bonds no one” to “bonds everyone.” They pile up at the two ends.

62% of companies bond under 10% of their customers for good. ~25% bond a deep, durable core — more than half. Almost nothing lands in between.

That near-binary shape is not a coincidence. It’s what you’d see if bonding were driven by one condition: does the product change something the customer can’t easily give up, or not. When it does, a core bonds for good. When it doesn’t, essentially none do.

So there are only two kinds of companies

One has a bonded core and builds on itself. The other has none, and has to re-buy its growth every year. They can post the identical churn rate, renewal rate, NRR, and NPS. The only number that tells them apart is the one nobody measures.

If you’re an investor, the single most important thing you can determine about a company is which of the two it is — and it’s invisible to the metrics diligence relies on. A healthy churn rate on a hollow base is a treadmill dressed as an engine.

If you’re building the company, your entire job reduces to one thing: build and expand the bonded core. Everything else — satisfaction, headcount, discounting — matters only to the degree it moves that number.

There’s a lot more underneath this

What actually builds the core (it isn’t customer size — it’s 88% about the product), the customer type already bonding inside your own data, and the segment that looks most loyal while being the least bonded of all.

That’s the report: The Science of Customer Bonding.

→ Read The Science of Customer Bonding

Start here

You don’t need the report to begin. Start with the only question that matters, the one your dashboard can’t answer: how many of your customers are actually bonded?

Almost no company knows. The ones that find out stop steering by the average and start steering by the shape.

— Greg