The most honest conversation I’ve had with a customer was the one when they were leaving.
In a previous role, I ran exit programs as part of my Director of CX roles. Any customer was free to go — we never tried to trap anyone — but before we closed the relationship, we’d sit down and ask a handful of straight questions. What did you like? What didn’t work? Why are you leaving? And if you could change one thing, what would it be? We’d make one honest case for staying, and if they’d already decided, we respected it and parted well.
And we did part well. Customers often told us afterwards how much they valued being asked — and that goodwill opened doors later, in candid feedback and opportunities we’d never have gotten if we’d let them slip away quietly.
Those conversations were gold. They were the clearest, least defended insight we ever got — no survey politeness, no hedging, just people telling us plainly what we’d gotten right and where we’d lost them.
And we were having them at the worst possible moment: on the way out the door.
That’s the thing I’ve carried with me since. The questions were never wrong. The timing was. Everything we asked at the exit, we could have asked while the relationship was still ours to protect and save — and done something with the answers.
Why the insight always arrives late
Most organizations are built to notice a customer at two moments: when they sign, and when they’re about to leave. In between, the relationship runs quietly in the background, and the next real checkpoint is the renewal — usually a scramble that starts sixty or ninety days out, when someone realizes the number is at risk.
By then, you’re not gathering insight. You’re performing an autopsy. You can learn why it died; you can’t change the outcome.
And the uncomfortable truth is that churn is rarely a surprise. Look back at a customer who left, and the signals are almost always there — a drop in usage, a champion who moved on, slower responses, a renewal conversation that got vaguer each quarter. Someone could have seen them. The problem isn’t that the signals don’t exist. It’s that no one was watching for them while there was still time to act.
Customer health is a continuous audit, not a renewal event
The shift is simple to say and hard to operationalize: stop treating customer health as something you check before a renewal, and start treating it as something you audit through the relationship.
I use the word audit deliberately. An audit isn’t a survey you send and a score you file. It’s a regular, deliberate look at whether the relationship is still healthy — and, when it isn’t, what has moved. In practice, that’s a health status on every customer — red, amber, green — reviewed on a regular cadence, not dusted off when the contract clock starts ticking. It’s unglamorous, and that’s the point. The value isn’t in the color; it’s in the fact that someone looks, regularly, and acts on amber before it becomes red. It also means never assuming green will stay green. Protecting a relationship takes real, ongoing work — and it’s some of the most valuable work you’ll do.
But a RAG status is only as good as the signals underneath it. And here most teams make the same mistake: they track what’s easy to measure rather than what predicts leaving. A satisfaction score collected once a year tells you less than a quiet, steady decline in how often a customer logs in, who still turns up to the review calls, or whether the person who championed you is still in the building. The signals that matter are usually behavioral and relational, not survey-based — and they often sit in different systems, owned by different teams, which is exactly why no one connects them.
Drift runs in two directions
When a customer’s health slips, the instinct is to look at the customer. Something changed on their side. And often it did. A business model is a bet on a particular fit between what you offer and what the market needs, and that fit decays when the forces around it shift. Customers are no different. A customer who’s drifting isn’t a random event; something underneath the relationship has moved, and the useful question is what.
Usually, it’s one of a few things on their side. Their needs or goals have changed — the problem you were hired to solve isn’t the problem that matters to them now. Their economics have changed — budgets tightened, priorities moved, and spend that was easy to justify last year is under scrutiny this year. The market has changed — a competitor caught up, or a cheaper way of doing the job appeared. New technologies and regulations have arisen.
But here’s the part most companies never audit: the customer isn’t the only one who drifts. We do too.
Somewhere between winning a customer and keeping them, the way we treat them quietly changes. The attentiveness that won the deal fades once the deal is done. Response times slacken. The senior person who was all over the pitch is three accounts removed by year two. The brand made a promise — fast, personal, dependable, whatever it was — and the day-to-day experience drifts a little further from that promise every quarter, without anyone deciding it should. No single moment is a failure. The gap just widens, from our side as much as theirs.
This is the harder audit, because it turns the question on us. Are we still delivering what won them? Does the experience still match the promise our brand makes? Are we treating this customer the way we did when we were trying to earn them — or the way you treat someone you’ve stopped trying to impress? Brands rarely lose customers in a dramatic failure. They lose them by slowly becoming something other than what the customer signed up for, and never noticing the distance.
So a real customer audit checks both directions at once. Has the customer drifted, and why? Have the forces around them shifted? And — the uncomfortable one — have we drifted: in our service, in our attention, in how faithfully we keep the promise that won them? Miss either side and your read is half the picture. A customer can leave because their world changed, because yours did, or because the two quietly drifted apart while everyone assumed the relationship was fine.
Catching it while you still can
The point of auditing both kinds of drift isn’t to produce a tidier report. It’s that each cause has a different fix, and you can only choose the right one if you catch it early.
If the customer’s needs have genuinely moved on, the honest answer might be a different solution, or a graceful goodbye — but on good terms, with the door open. If their economics have tightened, the conversation is about value and flexibility, now, not a panicked discount at renewal. And, if a competitor has caught up, you need to know while you can still respond. And if the drift is yours — if you’ve quietly stopped delivering what won them — that is the most fixable cause of all, and the one you will never fix if you are only ever looking at the customer.
Every one of those is a live conversation while the relationship is healthy enough to have it. None of them is available to you at the exit, when the decision is made, and all you are collecting is the reason.
The goal isn’t a better exit interview
I’m glad we ran those exit programs. We learned things, we parted with goodwill, and now and then someone stayed. But the real lesson wasn’t about how to run a good exit. It was that we’d built a system that only asked its sharpest questions once the answers could no longer change anything.
The goal isn’t a better exit interview. It’s never needing one — because you were auditing all along: watching the customer, watching the forces around them, and watching your own drift, and you saw the gap widening while there was still a relationship to save.
Churn is rarely a surprise. It’s a series of signals someone could have caught, acted on while it still mattered, and didn’t. The organizations that keep their customers aren’t the ones with the best save motion at renewal. They’re the ones auditing all year — honest enough to check not just whether the customer has changed, but whether the organization has.
No comments yet.