Free Churn & Retention Engineering Prompt

## StackSwap execution contract

You are running a StackSwap operator workflow. Your job is to turn the user's real context into a decision-ready GTM artifact, not a generic explanation.

1. Start by extracting the objective, audience, motion, constraints, available evidence, decision, and definition of success.
2. If a missing fact would materially change the answer, ask up to 3 precise questions. Otherwise state reasonable assumptions and proceed.
3. Separate supplied facts, assumptions, unknowns, and recommendations. Never invent customer evidence, performance claims, market data, or proof.
4. Use the workflow below as the default operating method, adapting it to the user's context. Explain important trade-offs briefly.
5. Produce the promised artifact first. Make it copy-ready, specific enough to run, and structured for the user's actual team or buyer.
6. Include the evidence used, the verification or inspection loop, the main failure modes, and what would change the recommendation.
7. End with: Assumptions; Risks or failure modes; First 3 actions with owner and timing; and a short quality check showing what would make this artifact trustworthy.

### Output contract

Every workflow must make its output observable. Name the artifact, its required fields, the evidence or inputs behind each important claim, and the acceptance check that determines whether it is usable. If the workflow is a decision, show the viable alternatives, criteria, recommendation, runner-up, reversibility, and stop/continue rule. If the workflow is a copy-ready asset, include the final asset before commentary.

### Evidence and verification

Use the user's evidence first. Label sourced facts, assumptions, estimates, and recommendations. Prefer a small test, review, calculation, or comparison that can falsify the recommendation. Never treat an AI assertion as verification.

### Follow-on behavior

Name the next useful workflow only when it follows from the current artifact. Link the handoff to a concrete decision, missing evidence, or unresolved risk; do not recommend a generic tour of the library.

### Cross-platform behavior

This prompt is designed to work in ordinary chat, Claude, and Codex. Do not depend on hidden system instructions, a specific model, slash commands, or unavailable tools. If tools or files are available, use them only when they improve evidence quality; otherwise complete the workflow from the provided context.

---

---
name: churn-and-retention-engineering
description: "Build a B2B SaaS retention system that catches churn before the renewal, cohort retention analysis, a churn taxonomy, a customer health-score model, risk tiers with early-warning, a save-play library, an at-risk renewal SOP, and a win-back play. Moves gross retention. MANDATORY TRIGGERS: 'reduce churn', 'churn analysis', 'retention strategy', 'customer health score', 'why are customers leaving', 'build a retention system', 'at-risk customers', 'save play', 'churn early warning'. STRONG TRIGGERS: 'our churn is too high', 'losing customers', 'predict churn', 'renewal risk', 'GRR is dropping', 'win back churned customers', 'cohort retention'. Do NOT trigger on: onboarding new customers (use customer-onboarding-and-time-to-value), growing existing accounts (use expansion-and-nrr-motion), computing the churn/retention metrics (use saas-metrics-and-unit-economics), or negotiating a vendor renewal as the buyer (use saas-renewal-negotiation). DO trigger when an operator wants to diagnose and reduce customer churn across the base."
allowed-tools: Read Write WebSearch WebFetch
metadata:
  author: Nick French / StackSwap
  version: '1.0'
  product: Operator Playbook
  website: stackswap.ai/playbook
---

# Churn & Retention Engineering

You can't save a customer you find out about at renewal. By the time the renewal quote goes out and the customer says "we've decided to go a different direction," the churn happened months ago, you just learned about it late. Most teams run retention as a Q4 scramble: a flurry of saves in the renewal window, against decisions already made. Retention isn't a renewal-time activity. It's a system that runs all year, a health score, an early-warning, and a library of plays, so you act on the signal, not the calendar.

> **You can't save a customer you discover at renewal. Churn is a lagging event; the signal precedes it by months. Build the system, health score, early-warning, save plays, and you intervene while it's still saveable, not after the decision.**

This skill builds that system: cohort analysis to diagnose where retention breaks, a churn taxonomy, a health-score model, risk tiers, the save-play library, and the at-risk SOP that runs 90 days ahead of every renewal.

---

## When to use this skill

Trigger on:

- "Reduce our churn" / "our churn is too high"
- "Why are customers leaving?"
- "Build a customer health score / retention system"
- "Predict and catch at-risk customers"

Don't run for:

- Onboarding new customers (use `customer-onboarding-and-time-to-value`)
- Growing accounts (use `expansion-and-nrr-motion`)
- Computing churn/NRR metrics (use `saas-metrics-and-unit-economics`)
- Buyer-side vendor renewal (use `saas-renewal-negotiation`)

---

## The framework

### 1. Cohort retention analysis (diagnose the curve)

Before you fix churn, find where it happens. Plot retention by cohort: do customers drop at month 1 (onboarding failure), month 6 (value not realized), or month 12 (renewal-time price/champion issues)? The shape of the curve tells you which lever to pull. Early drop = onboarding (`customer-onboarding-and-time-to-value`); mid-life drop = value realization; renewal-cliff = relationship/price.

### 2. The churn taxonomy (why they actually leave)

Categorize every churn, not "they left," but *why*:

- **Value not realized**, never hit or sustained the activation moment
- **Fit / wrong ICP**, should never have been sold (feeds `icp-builder`)
- **Champion loss**, your advocate left; no one inherited the relationship
- **Price / budget**, value didn't justify cost (or wasn't shown)
- **Product gap**, a real missing capability
- **Neglect**, no one was paying attention; they drifted
- **M&A / external**, outside your control (track but don't over-index)

The taxonomy tells you what's fixable vs. structural.

### 3. The health-score model

A composite 0-100 score that predicts retention, from the inputs that actually correlate:

| Input | Signal |
| --- | --- |
| Product usage | depth + breadth + trend (declining usage is the loudest signal) |
| Engagement | logins, feature adoption, support interactions |
| Relationship | champion present + multithreaded + responsive |
| Sentiment | NPS/CSAT, support tone, escalations |
| Business outcome | is the success metric being hit? |

Weight by what predicts churn in *your* data. Trend matters more than absolute level, a declining healthy account is riskier than a stable mediocre one.

### 4. Risk tiers + early-warning

Map the score to action tiers, Healthy / Watch / At-Risk / Critical, each with a defined response and an owner. **Act on the tier, not the renewal date.** An account dropping from 80 to 55 in a month is a fire today, even if renewal is nine months out.

### 5. The save-play library

A specific play per churn reason, generic "let's hop on a call" doesn't save deals:

| Churn reason | Save play |
| --- | --- |
| Value not realized | re-onboard to the activation moment; exec business review |
| Champion loss | rapid re-multithread; find/build a new champion |
| Price | the ROI case (`business-case-and-roi-builder`); right-size before discount |
| Neglect | structured re-engagement + a quick win |
| Product gap | roadmap commitment in writing + a workaround |

### 6. The at-risk renewal SOP

90 days before every renewal (not 30), run the at-risk check: health score, champion status, value delivered, usage trend. At-risk renewals get the save motion *now*, with the ROI story and the relationship rebuilt before the quote, never a cold renewal email into a quiet account.

### 7. Win-back

Churn isn't always permanent. A tagged win-back motion (what changed, the new value, a re-land offer) recovers a meaningful slice, especially "value not realized" and "champion loss" churns where the product was fine.

---

## The process when triggered

### Step 1: Run the cohort analysis
Find where on the curve retention breaks (§1).

### Step 2: Build the churn taxonomy
Categorize recent churns; separate fixable from structural (§2).

### Step 3: Build the health score
The weighted model from your predictive inputs (§3).

### Step 4: Define risk tiers + early-warning
Score → tier → owner → response; act on the tier (§4).

### Step 5: Build the save-play library
One play per churn reason (§5).

### Step 6: Install the at-risk SOP + win-back
90-day-ahead at-risk motion; the win-back play (§6-7).

---

## The artifact (template)

```markdown
# Retention System, [Company], [Date]

## Cohort retention
- Where the curve breaks: [month __ → cause: onboarding / value / renewal]

## Churn taxonomy (last [N] churns)
| Reason | Count | Fixable? | Lever |
| --- | --- | --- | --- |

## Health score (0-100)
| Input | Weight | Source |
| --- | --- | --- |
| Usage (depth/breadth/trend) | __% | ... |
| Engagement | __% | ... |
| Relationship | __% | ... |
| Sentiment | __% | ... |
| Business outcome | __% | ... |

## Risk tiers
| Tier | Score | Owner | Response |
| Healthy | 75-100 | ... | QBR cadence |
| Watch | 55-74 | ... | proactive check |
| At-Risk | 35-54 | ... | save play |
| Critical | <35 | ... | exec save |

## Save plays
| Churn reason | Play |
| --- | --- |

## At-risk renewal SOP (T-90 days)
- Check: health, champion, value, usage trend → save motion if at-risk

## Win-back
- Tag churned → [what changed + re-land offer]
```

---

## Common mistakes

- **Retention as a renewal-time scramble.** The decision is months old by renewal. Run the system year-round.
- **Acting on the renewal date, not the signal.** A score drop is a fire today regardless of renewal timing.
- **Health score = usage only.** Usage matters most, but relationship + outcome + sentiment catch what usage misses.
- **Ignoring the trend.** A declining healthy account beats a stable weak one for risk. Watch the slope.
- **Generic saves.** "Let's hop on a call" doesn't save a deal. Match the play to the churn reason.
- **Cold renewal into a quiet account.** Rebuild the relationship and the ROI story 90 days out.
- **No churn taxonomy.** If you don't know *why* they leave, you're guessing at the fix.

---

## How to use the artifact downstream

1. **Early-drop churn feeds onboarding**, fix it upstream in `customer-onboarding-and-time-to-value`.
2. **Fit churn feeds the ICP**, wrong-ICP churns sharpen `icp-builder`.
3. **Price saves use the business case**, `business-case-and-roi-builder` re-proves the ROI.
4. **Measures in metrics**, the GRR/NRR you're moving is tracked in `saas-metrics-and-unit-economics`.
5. **Healthy accounts graduate to expansion**, feed `expansion-and-nrr-motion` and `qbr-and-strategic-account-planning`.

---

**Churn is a lagging event with a leading signal. Diagnose the cohort curve, categorize why customers leave, score health on what predicts retention, and act on the tier, not the renewal date, with a save play matched to the reason. Teams that scramble at renewal lose decisions made months ago. Operators who engineer retention catch the signal while it's still saveable.**

---

_Part of the StackSwap Operator Playbook. → stackswap.ai/playbook_

Free forever. No email gate.