Almost everything we measure about a customer is a report on the past. How they felt about the last interaction. Whether last quarter went well. What they did in the product last month.
Churn isn't in the past.
Churn is a decision about the future, and customers make it based on what they believe is going to happen next. Which means the most useful thing you can know about a customer is what they believe about next quarter, and almost nobody asks.
The Question
How confident are you that [we / our product] will help you [the specific thing we promised you] over the next [time frame]?
That's The Confidence Question™. It's as simple as it looks, and I think the simplicity is what puts people off. The simplicity is the point.
Three things make it work.
It's forward-looking, so it's a leading indicator rather than a lagging one. It's about them rather than about you, which is why response rates are so much better than anything else I've put in front of customers. And anyone can answer it. An end user and a CFO will both have a view, and their answers mean different things, which is a feature rather than a problem. A no from an end user is usually friction. A no from an executive is usually outcomes.
It also fits next to whatever you're already running. If you like your satisfaction survey, keep it and add this. If you'd rather this replace it, that works too. I don't much care, and the question doesn't need the argument.
What It's Built to Catch
The failure this question exists to surface has a specific shape, and it's worth naming.
A customer says yes when they aren't sure.
They're not lying. They're being agreeable, or they don't want to look like they weren't paying attention, or they genuinely don't know what they don't know yet. Call it an uncertain affirmation. A yes with nothing behind it.
Nobody challenges it, because it was a yes and a yes was what you wanted. So it stands. And then it compounds. Uncertainty turns into poor execution, poor execution turns into poor results, poor results turn into frustration, and frustration turns into blame.
When a customer comes back angry at the end, it's almost always an uncertain affirmation that nobody challenged at the beginning.
Asking about confidence directly makes the uncertain affirmation expensive to give. It's much harder to say "yes, completely confident" than it is to nod along.
And when someone does say they're fully confident, hold them to it out loud. Ask them to tell you the moment that stops being true. Most people will, and now you have a customer who reports their own drop instead of one who quietly goes dark.
Anything Less Than 100% Is a Red Flag
The scale exists for the customer's benefit. It gives them room and it feels like a normal thing to be asked.
Behind the scenes there are only two buckets. Fully confident, and not.
The first is fine. The second needs intervention, because a customer who isn't fully confident you'll help them get what they came for is telling you something is wrong. And they're telling you now, rather than at renewal.
Watch for the deflection. Someone will say they never give a perfect score. Don't let it go. Try this instead: I'm glad you're confident, but a 10 means you're fully confident we can help you do what you need to do, the way you need to do it. What's keeping you from a 10?
Nine is not a rounding error. Nine is a customer telling you about something they haven't said out loud yet.
You Can't Diagnose From the Number
The number triggers the intervention. It doesn't tell you what the intervention is.
A seven might be entirely your fault. A three might be something happening in their world that has nothing to do with you and will still sink them as a customer. Scores don't sort by cause. So don't design your response off the score, design it off the conversation the score earns you.
Two questions do most of the work once you're in that conversation. What's the reason for the gap, and is it on our side or theirs? And what needs to happen to close it, on our side or theirs?
Keep It at the Account Level
Average it across your whole base and you've built yourself a vanity metric. It'll trend, it'll look reasonable on a slide, and it will tell nobody what to do on Monday.
The value is per account, where it's actionable. One name, one number, one intervention.
On benchmarking, I'm the wrong person to ask. I'm not interested in the average, because the companies worth learning from are usually the outliers those reports strip out to keep the curve clean. And in this model a fat middle isn't a healthy distribution, it's a queue of customers who aren't sure about you.
What Confidence Is Actually Measuring
Confidence is a reading on whether the customer can see the value yet, which is a different thing from whether the value is being delivered.
That distinction matters most in businesses where delivery is slow or invisible. Long implementations. Compliance work. Anything whose payoff is a disaster that doesn't happen. In those, there's no realized value to be satisfied about for a long time, so backward-looking measures have nothing to work with. Confidence has plenty.
It also gets at the thing customers actually fear. They hate unknowns, surprises, and repeating themselves. An uncertain affirmation is all three waiting to happen.
Start Asking It
Put it in a QBR. Send it in an email. Ask it on a call. Put it in the product. Add it alongside whatever you already send.
Don't overthink it. You're asking a customer whether they believe you'll help them get what they came for.
If the answer is anything short of yes, you just found the work.
