SkillChurn Save Playbook
Turn a list of churn signals into a risk-tiered save playbook: each customer scored from the signals you list, ranked by score, with fixed save actions per tier and a win-back draft for the highest-risk customers.
Works with: Claude Code, Codex, Cursor, Muse · Version 1.0.0 · Needs: Your churn signals: churn.txt (Product line, then one Customer block per customer with Signal lines from the fixed vocabulary and an optional Note line).
Preview
Sample output
An example of what this Skill produces. Sample content, shown for illustration.
Churn Save Playbook report — sample Customers: 4 | Gaps: 0 | Errors: 0 | Warnings: 0 Tiers: Acme Corp 65 Critical, Beta Labs 45 High, Gamma Studio 20 Watch, Delta Co 10 Low. Acme gets a personal reach-out within 24 hours and a win-back draft naming its two signals; Beta gets a check-in email and a usage-review call offer. Gamma gets a nudge and a re-check in 14 days; Delta is monitored. Gaps: None.
What you get
- A risk-tiered playbook: every customer scored, tiered, and ranked by score
- Fixed save actions per tier — Critical gets a personal reach-out within 24 hours
- Win-back drafts for Critical and High customers, naming only your listed signals
- Every gap, input error, and warning named in the report
About this Skill
Reads one plain-text file you write — churn.txt, with a Product line plus one block per customer (Customer / Signal / Note lines) — and produces save-playbook.md: a tier table sorted by score, highest first, plus one save plan per customer. Signals come from a fixed vocabulary with fixed weights (cancelled-renewal 40 down to usage-down-slight 10); tiers are Critical 60+, High 35–59, Watch 15–34, Low under 15. Local-only, standard library only; save actions are fixed per tier and a win-back draft names only the signals you listed — never a guessed reason, price, or discount — and a customer with no signals is a gap, not a zero-risk proof. The output is a plan for your own outreach: this skill connects to nothing and sends nothing.
Try it
Turn these churn signals into a risk-tiered save playbook — who do I contact first, what do I do, and what do I say?Good fit
- Deciding who to contact first when several customers show warning signs
- Turning a pile of churn signals into one ranked plan with concrete next steps
- Checking a half-written churn list for exactly what is still missing
Not for
- Sending mail — drafts only, through your own mail tool
- Predicting who will leave — a score ranks contact order, it is not a churn verdict
- Guessing a reason, a price, or a discount — the save offer is yours to decide after you ask
How to install
- Copy the install prompt into your agent:
Read https://skillu.dev/install/agent.md, then install the Churn Save Playbook skill (churn-save-playbook). - Or download the Skill package from this page after launch and place its folder in your agent’s skills directory.
- Call it with a slash command, e.g.
/churn-save-playbook, and describe what you want in plain words.
FAQ
How is a customer's score calculated?
Each listed signal adds its fixed weight — cancelled-renewal 40, payment-failed 35, no-login-30d 30, champion-left 30, usage-down-half 25, downgrade 25, support-unresolved 20, trial-ending 20, complaint 15, seats-unused 15, usage-down-slight 10. Duplicate signals count once. Tiers are Critical 60+, High 35–59, Watch 15–34, Low under 15.
Why do only Critical and High customers get a win-back draft?
A draft is a personal message that names real signals. Watch and Low customers get a nudge or monitoring instead, so a short note does not pretend a problem is bigger than the signals show. The draft itself names only the signals you listed — it never guesses a reason or a discount.