War-game a churn survey rewrite by testing whether question scope, routing owners, and analysis SLAs still produce decisions when cancel volume spikes and product teams want a single root cause from a checkbox list.
Churn surveys fail when "insight" means a stacked bar without a named reader, when incentives dual-count the same cancellation as both saved and lost, and when free-text replies live only in an export nobody opens. The review needs one written purpose for the survey, one owner who can change a question, and a documented consequence when analysis misses the SLA.
Frame the rewrite against real cancel paths
Collect the current cancel flows, survey placements, question bank, incentive rules, CRM fields, and the last ninety days of response rates with sample free-text themes. Mark which answers already map to a product or billing action. Name exclusions such as involuntary churn and fraud cancels. If a mobile cancel path skips the survey, show how that omission skews the dataset.
State the decision options: ship the rewrite, cut specific questions, or hold until routing and analysis owners are named. Include the fields that compute completion so two reviewers can reproduce the same quality label.
Challenge each question as an operating cost
- Which question still has no owner who will act on the answer within a fixed SLA?
- Which free-text prompt invites legal or abuse risk without a moderation path?
- Which incentive changes the cancel reason distribution enough to poison the sample?
- Which multi-select allows customers to pick every cause and erase priority?
- Which follow-up route can reopen billing disputes without a support handoff?
Run the rewrite in Pingpong with lifecycle, product, support, and research seats. Walk a price-driven cancel, a quality complaint with screenshots, a competitor switch, and a request to add five new checkboxes mid-quarter. For each case, start from the documented survey purpose. Ask who can add a question and which dashboard proves the answer was read. Any step that depends on an unnamed person becomes a rewrite condition.
Replay one prior cancel cohort under the proposed questions. If the cohort would still leave free text unread while the team cites the survey in roadmap debates, revise the routing before publishing.
Pair this rewrite with the lifecycle marketer seat and a winback campaign review. Preference honesty after cancel often needs an email preference center stress test. Adjacent churn work also lives in a churn playbook review and the war-game decisions hub.
Publish the survey only when each kept question has a named reader, an SLA, and a sample of how the answer changes a decision.