Churn Review Template
Prism Customer Success Handbook — Section 4.5
One template, not a process. The process is in 3.8 Churn & Offboarding; the quarterly review that consumes these is in 2.3 Operating Rhythm.
When to fill this in
- Enterprise and Mid-Market: every churn, individually. Completed within two weeks of the account closing, while the exit conversation is still fresh.
- SMB: aggregate at the quarterly review rather than one per account — individually reviewing every small churn costs more attention than it returns. Individual reviews for SMB only where the churn was surprising or the cause looks systemic.
The two questions this exists to answer
- What actually caused this? — not what the customer said caused it
- Was it preventable, and if so, what would have caught it?
Everything below serves those two. A post-mortem that produces a tidy narrative and no change is a story, not a review.
The template
Account
| Account | |
| Segment | |
| ARR lost | |
| Tenure | Months from handoff to close |
| Renewal date | |
| Date notice received | |
| Exit conversation held? | Yes / No — no response / Declined |
Timeline
Work backwards from the loss. The useful question in the exit conversation is when did you first start thinking about this, and what was happening then? — the answer is usually much earlier than the notice.
| Date | Event |
|---|---|
| Handoff, kickoff, milestone completions or stalls, escalations, champion changes, health movements, playbooks opened, renewal checkpoints, notice |
The point of the timeline is finding the moment it became likely — which is nearly always earlier than the moment it became visible.
Root cause
Primary cause (one, from 3.8 Churn & Offboarding)
- Never reached value — stalled in onboarding, never recovered
- Lost champion — successor never engaged, case never rebuilt
- Product gap — a capability they genuinely needed
- Service failure — escalations handled badly, feeling unheard
- Their business changed — clients lost, restructure, acquisition
- Price / value mismatch — probe: would they have paid it if value had been clear?
Contributing factors (any number)
Stated reason vs. root cause
| What they said | |
| What we concluded | |
| Why the difference | If there is one. "Price" is a symptom nearly every time — record what it was a symptom of. |
Preventability
| Preventable? | Yes / Partly / No |
| If yes — what would have caught it, and when? | A specific signal at a specific point, not "better engagement." |
| Did any signal fire? | Playbooks opened, health movement, escalations — and was it acted on? |
| If a signal fired and was missed, why? | Buried in triage, deprioritized, misread, or genuinely ambiguous |
| If no signal existed, what would have produced one? | Feeds trigger design at the quarterly review (2.3 Operating Rhythm) |
Distinguishing "we lost them" from "they left" is the point of this section. Some churn isn't preventable. A function that records every loss as a failure can't learn from the ones that were — and demoralises itself besides. Only preventable churn should generate a process change.
What we'd change
| Change | Where it lands | Owner |
|---|---|---|
| Trigger, playbook, weight, process page, product |
Or: "Nothing — this was not preventable and the process worked." That's a legitimate and honestly common conclusion. Manufacturing a lesson from an unpreventable loss produces process changes that solve nothing and add friction.
Loop closed
- Sales notified with root cause (2.5 Cross-Functional Interfaces)
- Product informed, if a product gap (2.8 Voice of Customer & Product Feedback)
- Account marked churned; data exported; end-client transition completed (3.8 Churn & Offboarding)
- Champion contact retained for future re-engagement
- Root cause categorized for quarterly aggregation (1.8 Metrics Library)
Reading these in aggregate
The individual review matters less than the pattern across several. At the quarterly system review (2.3 Operating Rhythm), read them together:
Cluster by cause. Three "never reached value" churns in a quarter is an onboarding problem, not three account problems. That's a change to 3.6 New Customer Onboarding, not three lessons learned.
Cluster by tenure. Losses concentrated in year one mean onboarding or expectation-setting. Losses at year three mean value erosion or relationship decay — different fix entirely.
Cluster by segment. SMB churn concentrated in one cause suggests the tech-touch model has a specific hole worth patching with content (2.9 Customer Education & Enablement).
Check the preventable ratio. If nearly everything is recorded as unpreventable, the reviews aren't honest. If nearly everything is preventable, either the process has real gaps or the reviews are over-attributing blame — both worth knowing.
Check whether signals fired. Churn where a playbook opened and went stale is a capacity problem (2.4 Account Prioritization Framework, 3.0 Playbook Index). Churn where no signal existed is a model problem (1.6 Health Scoring Methodology). These need opposite responses.
What makes a bad churn review
- "Reason: price." Recorded without probing. It will mislead every aggregate read afterward.
- A narrative with no change. Coherent story, no trigger adjusted, no page updated, nothing different next quarter.
- Blame. Of the customer, of Sales for a bad-fit deal, of Product for a missing feature. Any of those may be true and still be useless unless it converts into something that changes — a handoff field, a routed feedback item, a disqualification rule.
- Written from memory months later. The timeline is the most valuable section and the first thing to decay. Two weeks is the window.
Related pages
- 1.8 Metrics Library — churn by root cause, preventable vs. not
- 2.3 Operating Rhythm — the quarterly churn and system review
- 2.5 Cross-Functional Interfaces — notifying Sales
- 2.8 Voice of Customer & Product Feedback — routing product-gap causes
- 3.8 Churn & Offboarding — the playbook this template closes out