It's Live: The Science of Customer Bonding

The report is live. The Science of Customer Bonding — the real retention behavior of more than 1.3 million customers across roughly 70 B2B SaaS companies, and what it says about who actually stays. If you’ve been following the last few weeks, this is where you can read the parts I’ve been holding back.

Here is the whole thing in one line: three things build customer loyalty — the product, the fit, and the results — and Customer Success has spent a decade watching three others that don’t: satisfaction, health scores, and usage.

That’s the reset. Take it a piece at a time.

What works. Product: does it produce a compelling result at all? That single fact explains 88% of the variation in who bonds — far more than the customer’s size or budget. Fit: are you selling to the customers it’s actually for? Right-fit customers bond around 92%; wrong-fit customers collapse. Results: did those customers actually reach a result? Reach a measured result and they bond almost completely — a ~99% core.

What doesn’t. Satisfaction: the happiest customers leave soonest. Health scores: at one company, accounts a success team rated “Great” churned no better than the ones rated “Poor.” Usage: the heaviest users churned as fast as the lightest, because activity was never the goal — the result is.

Three things build customer loyalty — the product, the fit, and the results. And Customer Success has spent a decade watching three others that don’t: satisfaction, health scores, and usage. The entire CS dashboard, and none of it forecasts who stays.

Everything in the report reduces to one number: the bonded fraction — the share of your customers who are with you for good. It’s the only retention number that can’t be faked by a good quarter or a long contract, and it’s the one a churn rate can’t show you, a renewal rate can mask for years, and a satisfaction survey will never find.

The whole report is one question you can’t answer from your current dashboard: how many of your customers are actually bonded? It’s the number a churn rate can’t show, a renewal rate can mask for years, and a survey will never find — and almost no company knows its own.

Which is the honest problem with everything I’ve written for a month. It’s easy to say “measure your bonded fraction.” It has been genuinely hard to do. You can’t get it from a survey; you have to read it out of behavior — the shape of retention over time — and until recently that took a data team and a project nobody had time for.

So we built the thing that does it. You upload a CSV of your customers — when each started, when each left — or connect Salesforce or HubSpot. Out of that alone we can measure your bonded fraction (your PMF Score), the customer type in your base that actually bonds, and where you sit against companies genuinely like you. Then I take you through it. I’ve read a lot of these curves, and what they say about a business is rarely what you’d guess from the top-line number.

You don’t measure the bonded fraction with a survey. You read it from behavior — the shape of retention. This is where you stop taking my word for it and see your own number.

I’ve spent the last few weeks arguing that the metrics on your dashboard are pointed the wrong way, and that there’s one underneath them that tells the truth. This is where you stop taking my word for it and see your own number.

Read the report, then run yours. Everything I’ve said gets a lot more personal the moment it’s your base we’re looking at.