"Too expensive" is never why they left
It is the most common answer on every cancellation survey and every lost-deal note, and it is almost always false — because the customer already decided the price was acceptable when they bought.
Run an exit survey on any B2B product and one answer dominates: too expensive. It is near the top for almost everybody who publishes their data. The obvious response is to cut the price, or add a cheaper tier, or start discounting to save accounts at risk.
That response is nearly always wrong, and the argument against it is airtight:
Here's why it's never the reason. What did the customer do already? They came to your website, they looked around your homepage, they looked at your pricing page, they looked at competitors maybe. And they decided, yes, I'll buy this for this price. Which means they decided it was not too expensive.
Jason Cohen
The price did not change. They already accepted it once, with more scepticism than they have now. So what changed is the other side of the ratio — the value they believe they are getting. "Too expensive" is a statement about value, phrased as a statement about price, because price is the socially easy thing to say.
What they actually mean
"Too expensive" is a container for at least four distinct diagnoses, and they have nothing in common except the words:
- It did less than they expected. They believed it would do X. It does not, or not with their setup, or not without work nobody warned them about.
- They never got started. The value was real and they never reached it — the onboarding stalled, the data was never imported, the person who championed it moved on. They are cancelling something they never used.
- The thing it was for ended. A project finished. A campaign shipped. This is not a failure at all, and treating it as one distorts every metric you have.
- Someone above them is cutting costs, and yours is the line item with the weakest internal advocate. This is a champion problem, not a price problem.
Cutting the price addresses precisely none of these. It does, reliably, reduce revenue from everyone who was perfectly happy.
Cohen's framing of where the fault might sit is worth keeping, because it splits the fix:
Maybe they were wrong to expect that and you targeted the wrong segment. Maybe they were correct to expect it because you said so on the homepage and you're the one who didn't fulfill it.
Jason Cohen
Wrong segment, or overclaiming homepage. Two different fixes, and neither is a discount.
The one case where it is real
There is a version where price genuinely is the issue, and it is worth naming so this does not become dogma: when the price rises. If you have raised prices on existing customers and churn follows, the survey is telling the truth.
Even here the pattern is specific. Teams who run a regular annual increase — 5–10% every year, same month, announced upfront and never skipped — find the churn lands almost entirely on accounts that were never properly onboarded. Which makes the increase a diagnostic instrument rather than a loss: it identifies the accounts that were never getting value.
The failure mode is skipping years. One company skipped, then had to catch up with a roughly 20% jump, and "we definitely hit churn on that." Small and predictable beats large and occasional.
Instrument it properly
You cannot fix a reason you did not capture. The cancellation flow that produces usable answers:
- Multiple choice, never free text. Free text yields less, not more — most people type nothing, and the ones who type something write "too expensive". Structured options force a distinction.
- Offer a pause, not just a cancel. Far cheaper than re-acquiring the customer later, and it separates "done for now" from "done".
- Route by reason. Technical problem → open a real support conversation, with a human. Too expensive → offer a downgrade based on their actual usage. Not using it → that is an onboarding failure, and it should be routed to whoever owns onboarding.
- Send a founder email five minutes after cancellation asking for a reply. This reliably yields more truth than any in-product form, because it arrives after the decision is safely made and it is obviously from a person.
- Do not budget on win-backs. Roughly one in thirty respond — even to an offer of 100% off for two months. Win-back campaigns are not a retention strategy.
The same mistake in the sales pipeline
This generalises past churn. "Too expensive" is also the most common lost-deal note, and it fails there for the same reason: a prospect who reached a pricing conversation had already decided the number was plausible enough to keep talking.
What "too expensive" usually means on a lost deal is I did not believe it enough — insufficient proof, an unclear outcome, or a champion who could not make the internal case with what you gave them. The fix is proof and a better internal argument, not a discount. A discount, offered at the point of doubt, tends to confirm the doubt: it says the number was arbitrary, which invites the buyer to wonder what else is.
The three things to change
- Stop treating "too expensive" as a reason. Treat it as a symptom code that must be resolved into one of the four diagnoses above.
- Split your churn and your losses by whether the customer ever reached value. Someone who never activated is an onboarding number. Someone who used it happily for a year and left is a product or a lifecycle number. Averaging them together hides both.
- Before discounting anything, ask what changed. The price did not. Whatever did is the actual problem, and it is usually cheaper to fix than the revenue you were about to give away.
Keep reading
- Personalisation that works, and personalisation that costs you trust Using someone's first name is not personalisation — and the people you most want to reach read it as proof that a machine sent the email. Here is the line between the two.
- The only prospects worth emailing this quarter Most lost deals were never deals. They were people who agreed the problem was real, took the meeting, and then went back to the two things their job actually depends on.
- What to say when everyone already knows the problem Naming a pain your prospect discusses openly proves nothing — they hear it from every vendor in the category. There are two deeper levels, and the third one is where deals are actually won.