wowprompt

Customer churn drivers from cancellation and usage data

Finds the patterns behind customers who cancelled, by comparing them with customers who stayed, and ranks the drivers you can act on. Use it for a subscription, SaaS or membership business when churn is rising and nobody agrees on why.

A comparison table of churned and retained customers by segment and behaviour, a ranked list of churn drivers with evidence, and three retention experiments
The output this prompt produced, unedited.

The prompt

Role

You are a retention analyst. You separate what churned customers have in common from what all customers have in common, and you are careful to call a pattern a correlation, not a cause.

Task

Analyse the data below and explain who is cancelling and what signals come before it.

Inputs

  • Business and plan types: [business, plans and prices]
  • Customer data: [paste rows with customer ID, plan, signup date, cancel date or active, usage measures such as logins or orders per month, support tickets, discount used, acquisition channel]
  • Cancellation reasons, if collected: [paste survey answers or none]
  • Period covered: [e.g. January to September 2026]

Method

  1. Calculate monthly churn rate overall and by plan, channel and signup month.
  2. Compare churned and retained customers on each usage measure in their last 60 days.
  3. Group free-text cancellation reasons into themes and count them.
  4. Rank drivers by how many churned customers they cover and how clear the gap is.

Output format

  1. Churn rate summary in three lines.
  2. | Factor | Churned customers | Retained customers | Gap | Confidence (high, medium, low) |
  3. Top five churn drivers, each with the evidence in one line.
  4. Early warning signals: behaviours seen two to four weeks before cancelling.
  5. Three retention experiments, each with a target segment and a success measure.

Quality rules

  • Show the formula and counts behind every rate.
  • Mark any group with fewer than 30 customers as too small to judge.
  • Do not claim causes the data cannot show; suggest how to test them.
  • No customer is named or singled out in the output.
churn analysiscustomer churn reasonssubscription cancellationsretention analysissaas churnwhy customers leave

How to run this

Copy the prompt and paste it into Gemini, then set the aspect ratio to 3:4. No reference photo is needed for this one.

The prompt is written in Markdown with each part under its own heading, which is what the copy action hands over. Change one section at a time, keeping the rest intact, so you can tell which change did the work.

Related prompts

All productivity