Diagnose Why a Specific Client Churned

You are a head of accounts at a digital marketing agency running a churn post-mortem to extract durable lessons.
Agency: [AGENCY_NAME] — [SERVICES]
Churned client file:
[CLIENT_FILE]
Include: client name, tenure, MRR at churn, scope, stated churn reason from exit call, full timeline of last 6 months (calls, deliverables, results, escalations, leadership changes), CSAT/NPS history, results delivered vs promised at sale.
Diagnose the true root cause of churn — distinguishing it from the polite stated reason — and identify which earlier moments could have changed the outcome. Output should sharpen process, not blame people.

- Compare the stated churn reason to the timeline evidence — flag any contradiction
- Identify the earliest "intervenable moment" where a different action would likely have saved the account
- Agency-specific failure modes to test against: misaligned KPIs at sale, results below industry baseline, POC change handled poorly, reporting that didn't tell the story, scope creep eroding margin and quality, AM bandwidth, leadership not looped in
- Separate findings into: what the client said, what likely actually happened, what the agency could have done differently, what to change in process
- End with one falsifiable hypothesis the team can test on similar accounts

Sections: 1) Snapshot. 2) Stated reason vs evidence. 3) Timeline of intervenable moments (table: date | event | what should have happened). 4) Root cause (1 paragraph). 5) Process changes (3 bullets). 6) Hypothesis to test next quarter.

• [AGENCY_NAME] — Your agency name
• [SERVICES] — Services offered
• [CLIENT_FILE] — Full file on the churned account: tenure, MRR, scope, stated reason, 6-month timeline, NPS/CSAT history, results vs promises

Agency: Northbeam — SEO + content. Client: Vella Apparel, 11mo tenure, $9k MRR, full-funnel SEO + 4 blog posts/mo. Stated reason on exit: "budget reallocation." Timeline: Mo 1-3 strong onboarding, ranked 12 commercial terms top 20. Mo 4: original POC promoted, replaced by junior coordinator. Mo 5-6: 2 monthly reports went unread per email open data, no QBR booked. Mo 7: CSAT dropped 8 → 6. Mo 8: scope cut from 4 to 2 posts. Mo 10: invoice paid 21 days late. Mo 11: exit. Results vs promise: hit ranking targets, missed traffic target by 30%.

1) Snapshot: $9k MRR, 11mo, hit ranking goal, missed traffic goal by 30%.

2) Stated reason: "budget reallocation." Evidence contradicts — invoice paid late suggests budget pressure, but the silence after POC change is the bigger signal. True reason is closer to "new POC never bought in, no one rebuilt the case."

3) Intervenable moments:
| Date | Event | What should have happened |
|---|---|---|
| Mo 4 | POC promoted | Re-onboard new POC within 14 days, exec-to-exec intro |
| Mo 5 | Reports unread | Switch to 15-min Loom + 1-page exec summary |
| Mo 7 | CSAT drop | Trigger save play, not noted |
| Mo 8 | Scope cut | Should have been a red alert, treated as routine |

4) Root cause: agency did not detect or respond to the POC transition. Every later signal flowed from that miss.

5) Process changes: (a) auto-trigger re-onboarding playbook on any POC change, (b) replace static PDF reports with 1-pager + Loom, (c) treat any scope reduction as a Red save trigger.

6) Hypothesis to test: clients who get a re-onboarding session within 14 days of a POC change retain >2x longer than those who don't.

• Never run the post-mortem with only the AM in the room — bring delivery and a neutral exec
• Always test the stated reason against payment history and CSAT — words and money rarely lie together
• Capture one hypothesis per post-mortem and actually test it; the file is worthless if it just gets archived