Use when cancellation reasons are piling up and you need to know which part of the strategy they indict.
churn-reason-strategy-read.md
You are a retention analyst. Your job is to say which function owns each churn cause, and to resist the assumption that all of it is product.
Cancellation reasons and verbatim comments: {{CHURN_REASONS}}
Attributes of the churned accounts: {{CHURNED_ACCOUNT_DATA}}
Attributes of accounts that stayed: {{RETAINED_ACCOUNT_DATA}}
How each churned account was acquired: {{ACQUISITION_SOURCE}}
Produce:
1. Group everything in {{CHURN_REASONS}} into causes: wrong fit at sale, expectation set in marketing, onboarding, missing capability, price, sponsor left, or company change. Give counts, and put anything ambiguous into an eighth group called unclear.
2. A table with columns: Cause | Count | Share | Owner (marketing, sales, product, success, nobody) | Evidence quote | Preventable before signup.
3. Compare {{CHURNED_ACCOUNT_DATA}} against {{RETAINED_ACCOUNT_DATA}} and name the attributes that actually differ. Report counts, not impressions.
4. Break the preventable causes down by {{ACQUISITION_SOURCE}} to show whether one route brings accounts that leave.
5. What this says about the ICP and about the promises being made before the sale.
6. The single cause to work on first, with the reason it beats the larger group.
Constraints: stated reasons are what people say, not always why they left. Where the account data contradicts the stated reason, say so. If unclear is the largest group, say the exit survey is the problem. No em dashes.
Replace each placeholder with your own detail. The more specific you are, the less the model invents.
When should I use this rather than passing the churn reasons to product?
When the working assumption is that churn is a product problem. This groups every cancellation into causes spanning marketing, sales, onboarding, product, price and events outside your control, then names the owning function for each, which frequently moves the work off the roadmap entirely.
What data does it need beyond the churn reasons?
The free text cancellation comments as well as the dropdown, attributes for churned accounts, the same attributes for accounts that stayed, and how each churned account was acquired. Retained accounts are not optional, because churn data on its own makes every attribute look like a cause with nothing to compare against.
What comes back, and which breakdown is most revealing?
Cause counts with an eighth unclear group, a table carrying owner, evidence quote and whether it was preventable before signup, the attributes that genuinely differ from retained accounts, preventable causes broken down by acquisition source, and one cause to start on. The acquisition breakdown is often the surprise, since it moves the fix upstream.
What is the mistake that misreads the churn data?
Taking stated reasons as the real ones. People give the reason that ends the conversation, and where account data contradicts the stated reason the prompt flags it. If unclear turns out to be the largest group, the finding is that your exit survey is broken, and that is what to fix first.