Churn Reason

Churn — customers leaving the product — is easy to count. Six of last month's 200 accounts ended their contracts, so that's six. Why those six left, though, lived for a long time only in free-text notes a rep typed at the end of a call.

Those notes read like this: "budget cut internally," "the person who used it left the company." Each one has its own sad story, and after twenty of them you still don't know which problem is the biggest. Nothing in the pile can be sorted, so nothing in it can be ranked.

So the meeting always stalled in the same place. The CEO asks why customers left last quarter, the rep recounts the two or three accounts still fresh in memory, and those two or three are usually the loudest or the most recent. Engineering can't decide whether to build anything on that, and the same meeting runs again next quarter.

The churn reason is a field that makes these records countable. Instead of a sentence, the rep picks one of five or six predefined options. Twenty free-text notes produce no ranking; twelve of twenty landing on the same option decides what gets worked on next quarter on the spot.

The rep picks one predefined option the moment a contract ends, and those picks get tallied every quarter. Three steps.

1. Define the options first. Price, missing features, not using it, champion left, budget cut, moved to a competitor is enough. Past ten options, two reps file the same situation under different labels and the counting stops meaning anything; under three, half of everything lands in "other."
2. One option per account. Allowing several makes the per-option totals exceed the number of accounts, so no percentage works. When causes overlap, pick the last one that actually ended the contract. A team whose budget shrank and dropped this tool while keeping three others belongs under not using it rather than budget cut.
3. Tally by quarter. Count the money those accounts paid alongside the number of accounts. Price often ranks first by count while missing features ranks first by revenue, because small accounts are sensitive to price and large ones to what the product can and can't do.

Two inputs fill the field. One is the answer the customer picked in the cancellation survey; the other is what the rep heard in months of conversation. When the two disagree, the rep picks one and writes a single line in the notes field explaining the choice. The option exists to be counted, the note exists to be reread later, so both stay.

Accounts cut off by an expired card — involuntary churn — get counted separately. Those customers didn't leave because they disliked the product, and mixing them in pushes payment failure to the top of the ranking and hides what matters.

Collected reasons flow to three places: what to build next, which customers to hold on to now, and how to adjust pricing and contract terms.

1. Product priorities. If 7 of 12 departures cite missing features, the notes field tells you which features those 7 meant, and that goes on next quarter's list. The option sets the order; the note supplies the content. Without the labels nobody knows which notes to open first, and without the notes the labels never become a task.
2. Renewal forecast. Whatever the departed accounts had in common in their final months gets applied to the accounts still here. If every account labeled "not using it" had its monthly active users cut in half over the last three months, an account showing that pattern now gets flagged as a red account.
3. Pricing and terms. Price concentrated among small accounts means a cheaper entry plan is missing; price concentrated among large ones means nobody built a use case worth that amount. The same label points to opposite work depending on where it came from.

Set it beside the voice of customer (VoC) and one more thing appears. When the same theme shows up both in complaints from people still using the product and in the endings written by people who left, that complaint demonstrably reaches contract termination. A theme that generates plenty of complaints but never appears as a churn reason is noisy rather than fatal.

"Missing features is the top churn reason for the second quarter running. I think it has to go on the roadmap."

A ranked number being used as the argument for changing engineering priorities.

"File this one as not using it, not price. They hadn't logged in for two months."

Classify by what actually happened rather than copying down the answer the customer gave.

"Pull the payment failures out of the churn reason tally."

Only customers who chose to leave belong in the ranking.

At Pagedeck, Sofia walks through every contract that ended in the quarter and settles the labels. The finished table goes to Mia, who lays the count ranking and the revenue ranking side by side and drafts next quarter's work list. When the two rankings disagree, Grace decides which one to follow — count in a quarter where small free-plan teams dropped off, revenue in a quarter where an enterprise account left.

One quarter price jumped to the top. Opening the notes field showed that six of the eight cases were a version of "only two people use it but we have to buy ten seats," and the fix ended up being finer seat increments rather than a lower price.

The most common mistake is taking the stated reason as the real one. It came straight from the person leaving, so it looks authoritative, and doubting a departing customer feels rude. But people on the way out pick the answer that is quick and least awkward, and price wins that contest almost every time. Most "too expensive" answers don't mean the price is high; they mean the account never used it enough to justify the price. Record them literally and you end up cutting the price 20 percent, and six months later revenue is down while churn is unchanged. Putting the last three months of usage next to the label filters out most of this.

A second mistake is treating the act of collecting reasons as churn reduction. A tally that grows every quarter feels like management. But by the time the table gets filled in, that customer is already gone. On teams where the table never changes the next quarter's work list, the top reason sits in the same spot two years later.