churn-autopsylisted

Find why customers leave before deciding how to keep them: build a ledger of every customer who left from the person's own data, score the five suspects (selection, promise, delivery, experience, economics) against that ledger, and rank the causes by how many customers each one explains. Use whenever someone says "clients keep leaving", "100% churn", "customers do not come back", "retention is bad", "nobody renews", "hamare clients ruk nahi rahe", "clients chhod ke ja rahe hain", "is it our quality or our UX", "help me find the root cause of churn", or asks for a retention plan, a loyalty offer, a win-back campaign or a discount before the cause is known. Also use when an owner arrives already certain of the cause ("it is the dashboard", "the team is slow", "our reports are confusing"). Stand down when the ask is a prompt (noguess owns it), a plan for a business with no customers yet (business-blueprint owns it), or a fix to one named thing one customer complained about.
albertpikkop/noguess-prompt-skill · ★ 0 · AI & Automation · score 72

Install: claude install-skill albertpikkop/noguess-prompt-skill

# Churn Autopsy ## What this makes Two files in the person's own folder, in this order, never the other way round: ```text CHURN-LEDGER.md one row per customer who left, from the data. Every silent cell is [PENDING] CHURN-AUTOPSY.md NOTES, FACTS, FINDINGS: the five suspects scored, causes ranked by count, the owner's own theory with its count, one experiment, the next gaps ``` Speak in plain words, short sentences. Answer in the language the person wrote in; the file headings, the suspect names and [PENDING] stay in English. No em-dashes in text you write. ## The one law: no cause before the ledger An owner who asks why customers leave already has an answer in mind. Quality. The team. The dashboard. An agent that agrees with it is a mirror, and a mirror is what they had before they asked. The customers who left are the evidence, one row each, and a cause counts only when it explains most of those rows. So the ledger comes first, every time, even when the person asks for the fix, the offer or the campaign in their first message. Why the order matters: a fix chosen before the count is a bet on one story. If the story is wrong, the fix costs money and the customers keep leaving, and the owner concludes that nothing works. If the story is right, the count would have shown it in an hour. ## Stand down - The ask is a prompt for another AI ("TCE this", "write the brief"): noguess owns it. - The business has no customers yet: business-blueprint owns i