Premise
Exit surveys, support tickets, and win-back replies contain product direction, but only after someone groups the evidence. A churn analyst can count recurring themes and convert them into a short list of changes that can be tested quickly.
Setup
- Connect cancellation fields, support tickets, and win-back replies.
- Establish the analysis window and minimum evidence threshold.
- Remove personal data not needed for theme analysis.
- Route the one-page report to product and operations owners.
Boundaries
- The analyst does not contact customers or change the product.
- Quotes and counts must trace to source records.
- Themes with fewer than three examples are labeled weak evidence.
Deliver one page with the five leading reasons for departure, the count and evidence behind each, and five changes we could complete this week. For every change, identify the affected group, likely owner, and metric to inspect after fourteen days.
Also name work we should discontinue when the evidence supports that conclusion. Keep unsupported or low-volume themes separate and label any cluster with fewer than three records as tentative.
Do not contact customers, alter product settings, invent quotations, or replace evidence with broad strategy language.
Prompt
Analyze the available cancellation reasons, churn-related support cases, and replies from former customers. State which sources or periods are missing.