Diagnose why customers leave: classification, root cause per segment, 10-step retention playbook, 3 experiments, early-warning signals.
Your monthly churn is 6.8% and three months ago you told yourself you'd fix it 'after the launch'. The launch shipped. Churn is now 7.1%. You have a Notion doc called 'why people leave' with three bullets and one of them is just the word 'pricing'. The retention playbooks online assume you have a CS team, a product analytics stack, and 18 months of cohort data. You have none of that. This tool diagnoses what's actually killing your retention from the signal you do have — top cancel reasons, segment data, last few churn-call summaries — then classifies the churn (product-fit, pricing, activation, competitor, lifecycle), root-causes per segment, gives you a ranked retention playbook based on impact × effort, and three specific experiments you can run in the next 30 days with success criteria written in advance. Plus the early-warning signals to instrument so you stop being surprised next month.
Brief used: “B2B SaaS, AI meeting notes for sales teams. MRR $52k, monthly churn 6.4% (logo) / 5.2% (revenue). Top 3 cancel reasons from exit survey: (1) 'didn't end up using it enough' (44%), (2) 'team uses Gong/Chorus' (28%), (3) 'too expensive for what we got' (18%). Segments most affected: solo SDRs at sub-50-person companies (~70% of churners), and one-off self-serve trials that converted but never onboarded the team. Last 5 churn interviews (summary): 3 of 5 said they never got past the first 2 weeks; 1 said they used it but moved to Gong when their company expanded; 1 said the AI summaries were 'too generic to act on'. Mostly self-serve at $99/mo. No CS team yet.”
# Churn Diagnostic — [Company] ## Headline Your 6.4% monthly logo churn is dominated by activation failure, not pricing or competitor loss. The single biggest lever in the next 60 days is your first-14-days onboarding for solo SDR signups at sub-50-person companies. Pricing complaints and Gong/Chorus loss are real but secondary. ## Churn Classification | Category | Weight | Evidence | |---|---|---| | **Activation failure** | 50% | 'Didn't use it enough' + 3/5 interviews never made it past week 2. This is your dominant churn mode. | | **Competitor (lifecycle)** | 22% | Gong/Chorus mentioned by 28% of exit survey + 1/5 interviews — but specifically when the customer's company crossed a size threshold. This is upgrade-churn, not loss-to-competitor. | | **Pricing perception** | 18% | 18% cite pricing — but pricing complaints almost always co-occur with activation failure. 'Too expensive' usually means 'didn't get value', not 'value too low for price'. | | **Product-fit** | 8% | 'Summaries too generic' from 1 interview. Real but small. | | **Lifecycle (natural)** | 2% | Companies that wound down or pivoted away from outbound. Unrecoverable. | ## Root Cause by Segment **Segment A — Solo SDRs at sub-50-person companies (~70% of churn).** This segment signs up alone, never invites a teammate, and the product loses its 'shared notes' wedge. Week 1 usage is mostly first-meeting-only. Week 2 they forget to install the Zoom side and the friction wins. They churn at the renewal point of month 2-3. **Segment B — Self-serve trials that converted without team onboarding (~15% of churn).** This segment paid (often with a personal card) and then never got the rest of the team on. They use it alone for 4-6 weeks, can't justify the budget, and churn at first invoice or first manager-conversation. **Segment C — Mid-sized SaaS that hit a growth threshold (~12% of churn).** Genuine competitor loss to Gong/Chorus. These customers grew past the 'solo SDR' use case into a 'sales leadership wants conversation intelligence' use case. You don't compete here yet — and may not want to. ## 10-Step Retention Playbook (Ranked by Impact × Effort) 1. **Force a 'second user invite' in onboarding** — make the first action after signup an invite-a-teammate step. Impact: high (kills Segment B). Effort: low. 2. **Day-3 activation email + in-app nudge for users who haven't recorded a meeting** — concrete: 'install the Zoom integration'. Impact: high. Effort: low. 3. **Define 'activated' = 3 meetings recorded + 1 teammate invited in 14 days. Instrument and alert.** Impact: high. Effort: medium. 4. **First-meeting wow moment** — the first transcript needs an obvious useful artifact (action items, next-step email draft). Impact: high. Effort: medium. 5. **Save-call script for cancel form** — when self-serve users hit cancel, offer a 30-min audit call. Impact: medium. Effort: low. 6. **Pricing page clarity for solo vs team** — solo at $49, team at $99/seat. Re-segment your $99 customers. Impact: medium. Effort: medium. 7. **Reposition for solo-SDR-at-small-company** — your dominant user is also your dominant churner. Sharpen the value prop for them specifically. Impact: medium. Effort: high. 8. **Win-back sequence for Segment B at 45 days post-churn** — they may have rejoined a new company. Impact: low-medium. Effort: low. 9. **Concede Segment C consciously** — don't try to compete with Gong on conversation intelligence. Document this as policy. Impact: low (frees energy). Effort: low. 10. **Charge annually for team plans** — pricing pre-commit reduces month-3 churn for the small subset of team-converted accounts. Impact: low at current scale. Effort: medium. **Do this month: 1, 2, 3, 4. Stop everything else.** ## 3 Experiments for the Next 30 Days **Experiment 1: Forced teammate invite in onboarding** Hypothesis: forcing an invite step before the 'core action' will improve activation rate from current ~41% to >55%. Intervention: add a 'who else on your team should see these meeting notes?' step as required step 2 of signup; skippable but with friction. Success criteria: activation rate up 10+ points in 21 days for new signups, no >5pp drop in signup completion. Time bound: ship by day 5, measure days 5-30. **Experiment 2: Day-3 activation email for non-recorders** Hypothesis: users who haven't recorded a meeting by day 3 are 5x more likely to churn; a single targeted email recovers ~15% of them. Intervention: triggered email day 3 if no recording, headlined 'install the Zoom side in 90 seconds — here's how', with a single CTA. Success criteria: 15%+ of recipients record a meeting within 5 days of the email. Time bound: ship by day 3, measure days 3-30. **Experiment 3: Pricing re-segmentation page test** Hypothesis: $99/mo flat is overpriced for solo SDRs and underpriced for teams — a $49 solo / $99 team split reduces self-serve cancel rate at first invoice. Intervention: A/B the pricing page (50/50) for 21 days, route new signups to the new pricing. Success criteria: 30-day retention on new cohort >= existing cohort, with directional improvement in either conversion or first-invoice retention. Time bound: ship by day 10, measure days 10-31. ## Early-Warning Signals to Instrument Instrument these as events; alert weekly on customers crossing the thresholds: 1. **No meeting recorded in first 7 days post-signup** — predicts 75% churn within 60 days. 2. **Solo user (no teammate invited) at day 14** — predicts 60% churn within 90 days. 3. **No login for 10 consecutive business days** — predicts 50% churn within 30 days. 4. **Decline in meetings-per-week >50% week-over-week for 2 weeks** — late-stage churn signal, ~45% churn within 21 days. 5. **Support ticket containing pricing keywords in last 30 days** — predicts 25% churn at next renewal. 6. **Company crossed 75 employees (LinkedIn enrichment)** — predicts Segment C competitor evaluation; flag for outreach. ## Save-Call Script (when self-serve user hits cancel) 'Hey [Name] — saw you're cancelling. Quick question before you go: was it something specific that broke for you, or did the product just not stick? Either way, no pressure to stay — but if you'll trade me 15 minutes, I can either fix it for you or refund the last month if it didn't deliver. Pick whichever.' — sends 30-min link. ## Win-back Email (45 days post-churn, Segment B) **Subject:** Did you move companies? 'Hi [Name] — when you cancelled in [Month], you mentioned the team never quite got on. Two questions: (1) are you still at [Company]? (2) if you've moved, would meeting notes be useful at your new team? If yes to either, happy to spin you up fresh — first month on us. Reply yes / no / new-company.'
Static example — your run uses Claude live on your specific brief.
Post-PMF SaaS founders watching MRR plateau because new revenue equals churn, founders who just hired a CS person and have no idea what to point them at, anyone whose 'retention strategy' currently means responding to cancel emails, teams seeing healthy top-of-funnel but flat net new MRR. Not for: companies pre-PMF with churn under 2% (you probably have a different problem), or B2C consumer apps where churn dynamics are different.
A full churn diagnostic pack: (1) churn classification across 5 categories (product-fit, pricing, activation, competitor, lifecycle) with the % weight of each based on your inputs, (2) root cause analysis per affected segment — which customer types are leaving and why their reasons are different from each other, (3) 10-step retention playbook ranked by impact × effort (the top 3 are the ones to do this month), (4) 3 specific experiments to run in the next 30 days, each with hypothesis, intervention, success criteria, and time bound, (5) early-warning signals to instrument — the 4-6 in-product events that predict churn 14-30 days before it happens so you can intervene, (6) a save-call script and a win-back email tuned to the dominant churn reason. Markdown ready to paste into your CS doc.
New MRR looks healthy but net MRR is flat for 3 months. Diagnose whether you're losing accounts faster than you're winning them and which segment is the leak.
Investors will ask about churn. Get a doc that classifies it correctly, names the segment that's the problem, and shows the experiments you're running about it.
Your new CS hire starts Monday and you have no playbook. Hand them the 10-step ranked playbook + the early-warning signals to start tracking immediately.
You did the interviews and have notes. Turn them into a real diagnostic with classification and ranked actions instead of a Notion page of quotes.
The diagnostic still runs on exit-survey reasons + segment data, but the per-segment analysis will be coarser. The tool will name the 3-5 specific things to ask in your next 5 churn interviews to sharpen the diagnostic. Doing those interviews is itself one of the recommended actions.
Yes, with a caveat the tool calls out — directional but not statistically clean. At small N, the dominant churn class is usually one or two patterns, and the playbook focuses on those. Re-run the diagnostic at 100+ churned customers for tighter weighting.
Generic playbooks list 30 things. This ranks the steps for your actual data and tells you which 3 to do this month and which 7 to ignore until next quarter. The experiments come with success criteria written in advance so you don't fool yourself.
The experiments are hypotheses with kill criteria. Roughly 40-60% of well-designed retention experiments produce a measurable lift in early-stage SaaS; the rest are useful negative results. The tool deliberately picks high-leverage experiments where even a negative result is informative.
Then the playbook shifts — the top actions become positioning sharpening, feature gap analysis, and a stay/leave call before churn. The tool re-weights when 'competitor' is the dominant class instead of activation. Be honest in the input about the reasons.
Yes. You get an anonymous preview instantly with no signup. Drop your email and you unlock 3 full-length runs per month for Customer Churn Diagnostic — no credit card. Unlimited runs are $79 one-time, or $19/mo for every tool.
Paid ($79 one-time) unlocks unlimited runs for Customer Churn Diagnostic, longer outputs from Claude Sonnet, full exports, and priority generation. $19/mo unlocks every tool on JustNeeda.
Free runs render in-browser and can be copy-pasted. Paid unlocks copy-to-clipboard, Markdown, and plain-text exports — and history of every run tied to your account.
No. Every run hits Claude live with your specific input. We don't reuse outputs across users. Your input stays private to your session and account.