| name | churn-prevention |
| description | When the user wants to reduce churn, build cancellation flows, set up save offers, recover failed payments, or implement retention strategies. |
| allowed-tools | Read, Write, WebSearch, WebFetch, AskUserQuestion |
| model | sonnet |
Churn Prevention & Retention
Expert knowledge for reducing customer churn through proactive signals, cancellation flow design, dunning sequences, and win-back campaigns.
Churn Taxonomy: Know What You're Fighting
Treating all churn the same produces generic solutions that fix nothing.
Voluntary churn: Customer actively decides to leave.
- Value-not-realized: They never got the outcome they bought the product for
- Needs changed: Their situation changed (company downsized, project ended)
- Competitor switch: A competitor offered something compelling
- Price objection: They don't believe the price is worth it anymore
Involuntary churn: Customer leaves due to payment failure.
- Card expired
- Insufficient funds
- Bank fraud block
- Card lost/stolen
Silent churn: Customer stopped using the product but hasn't cancelled yet. Subscription still active, but they are a churner waiting to happen.
Involuntary churn often accounts for 20-40% of total SaaS churn. Fix it before building complex save flows.
Churn Signals: Early Warning System
Build a customer health score from behavioral signals before implementing retention tactics.
High-risk signals:
- Login frequency dropped >50% vs their first 30 days
- Key feature usage (the one that correlates with retention) declined
- Support tickets opened about billing or "how do I cancel"
- Last session was >14 days ago for a daily-use product
- Decision-maker contact changed at the account
- Renewal date is within 30 days and QBR hasn't been scheduled
Health score framework (simple version):
Health = (Login frequency score × 0.3)
+ (Key feature usage score × 0.4)
+ (Expansion signals score × 0.2)
+ (Support sentiment score × 0.1)
Score 0-100. Red < 40, Yellow 40-70, Green > 70.
Trigger proactive outreach at Yellow. Don't wait for Red.
Proactive Retention Playbook
Stage 1: Value Realization (Days 1-30)
Goal: Ensure they hit their first "aha moment."
- Onboarding sequence with activation milestones
- Success manager check-in at day 7 for paid accounts
- In-app prompts when key feature not yet used
Stage 2: Engagement Maintenance (Days 31-90)
Goal: Build habits that make cancellation feel costly.
- Weekly usage digest emails ("Here's what you accomplished")
- Feature discovery campaigns for unused features
- Customer success call for accounts with declining usage
Stage 3: Renewal Defense (30 days before renewal)
Goal: Preemptively address objections before they become cancellations.
- Health check for all accounts renewing in 30 days
- Proactive outreach for Yellow/Red accounts
- ROI report showing value delivered since last renewal
Cancellation Flow Design
The cancellation flow is not about trapping users — it's about understanding why they're leaving and offering the right intervention at the right time.
Cancellation Flow Architecture
Cancel button click
↓
[Pause / Downgrade offer — show alternatives]
↓
[Reason selection — 4-6 options]
↓
[Targeted save offer based on reason]
↓
[Confirmation + offboarding]
Reason-Based Save Offers
| Stated Reason | Save Offer |
|---|
| Too expensive | Downgrade tier, pause account, annual plan discount |
| Not using it enough | Pause account for 1-3 months, usage coaching |
| Missing a feature | Roadmap preview, workaround, escalation to PM |
| Found a competitor | Competitive comparison, differentiation reframe |
| Business circumstances changed | Pause, downgrade, extended payment terms |
| Technical problems | Immediate escalation to technical success |
Pause > Cancel
Offering a 1-3 month pause recovers a significant portion of would-be churners. They are not lost — they are in a holding state. Paused accounts reactivate at materially higher rates than full churns.
Save Offer Principles
- Time the offer correctly. Show the save offer after they've stated their reason, not before. Premature offers feel desperate.
- One offer, not five. Multiple offers signal desperation and dilute perceived value.
- Set expectations for the offer. "We'd like to offer you 2 months at 50% off" not "we have something special for you" (builds resistance).
- Don't offer what you can't sustain. Discounts offered in cancellation flows get remembered by customers and expected on renewal.
Dunning: Recovering Failed Payments
Involuntary churn from failed payments is 100% recoverable if you have a good dunning sequence.
Dunning Sequence Design
Day 0 (failure): In-app banner + email "Payment failed — update your card to keep access"
Day 3: Soft reminder email — "Your card ending in XXXX failed. Update to avoid service interruption."
Day 7: Firm email — "Action required: Your account will be suspended in 3 days"
Day 10: Suspension with grace period — account suspended, data preserved
Day 20: Final email — "Your data will be deleted in 10 days. Download or reactivate."
Day 30: Deletion or archival per your data retention policy
Smart retry schedule (for card errors):
- Day 0: First attempt
- Day 3: Retry (cards often work again after banks clear temporary blocks)
- Day 7: Retry
- Day 14: Final retry before suspension
Dunning email copy principles:
- Lead with the consequence, not the process ("Your access is at risk" not "Payment processing issue")
- Make the CTA one click to card update page — never send them to login first
- Show specifically what they'll lose
- Show what they won't lose (their data is safe)
Card Update Optimization
Dunning emails must link directly to a pre-filled billing update page. Friction = lost recoveries.
- Auto-fill account email
- Accept all major card types
- Support Apple Pay / Google Pay for mobile
- Show saved card last-4 for context
Win-Back Campaigns
Churned customers who left on good terms are your best leads. They already know the product.
Win-Back Timing
- 30 days post-churn: Too soon. They just left. Exception: involuntary churn where you just fixed the issue.
- 60-90 days post-churn: Prime window. Especially if they churned due to circumstances (budget, project ended).
- 6 months post-churn: Second window. Use if competitor or product gaps were the reason — circumstances may have changed.
Win-Back Offer Hierarchy
- New capability: "Since you left, we launched [feature they asked for]"
- Special offer: 2-3 months at a reduced rate to re-establish value
- Free audit/review: Offer a free account audit or strategy session
- Social proof: "[Customer like them] just [achieved outcome] — come see how"
Win-Back Email Sequence (3-touch)
Email 1 (60 days): "A lot has changed since you left" — product updates relevant to their stated reason
Email 2 (74 days): "We want you back — here's an offer" — time-limited discount or extended trial
Email 3 (88 days): "Last chance — offer expires in 48 hours" — urgency close
NPS Correlation to Churn
NPS (Net Promoter Score) is a leading indicator of churn if used correctly.
- Detractors (0-6): High churn risk. Trigger immediate account review. Do not wait for renewal.
- Passives (7-8): Moderate risk. Focus on converting to promoter via feature education.
- Promoters (9-10): Low churn risk. Focus on expansion and referral programs.
NPS survey cadence:
- 30 days after signup (first impression)
- 90 days (value realization check)
- Every 6 months thereafter
Do not use NPS as vanity metric. The score alone is useless. The follow-up action on detractors is where the retention value lives.
Common Rationalizations
| Rationalization | Reality |
|---|
| "Churn is just part of SaaS — you can't prevent it" | 30-50% of churn is preventable with proactive signals and targeted interventions. |
| "Our cancellation flow is annoying customers — let's make it one-click" | One-click cancel removes your ability to save anyone. Friction in cancellation, when done right, recovers 10-30% of cancellations. |
| "Win-back campaigns don't work" | Win-back rates are consistently 20-40% when timed correctly and paired with a relevant offer. They already know your product. |
| "Failed payments are IT's problem, not marketing's" | Involuntary churn is a revenue problem. Marketing should own dunning email copy and recovery rates. |
| "We don't have enough data to build a health score" | Login date + feature usage + support tickets = enough for a basic health score. Start simple. |
Verification
