Analyze Your Product Funnel and Find the Biggest Drop-Off to Fix First
Paste your funnel step counts and get a conversion table, the largest leak flagged with reasoning, and one testable fix per step with an expected lift range.
You getFunnel table, biggest leak and fix table
Prompt
Act as a senior product analyst who runs conversion experiments.
Your job is to analyze the product funnel I will paste right after this prompt. It will list each step and the number of users who reached it. If the time window, the product type, or the order of steps is unclear, first ask me up to 3 short numbered questions in one message and wait for my answers.
Rules:
1. Show every calculation to one decimal place: step conversion, users lost, and cumulative conversion from the first step.
2. Flag the largest leak by both percent lost and users lost. If they differ, pick one to fix first and explain why in 2 sentences.
3. Give exactly one testable fix per step transition. Each fix needs a hypothesis, a concrete test, and one success metric.
4. Give each fix an expected relative lift range, clearly labeled as your assumption. Do not cite industry benchmarks or statistics.
5. For the top fix, calculate how many extra end-of-funnel users the low and high end of the range would produce.
Output format:
- Funnel table: Step | Users | Step conversion | Users lost | Cumulative conversion
- Biggest leak (3-4 sentences)
- Fix table: Transition | Likely cause | Test | Success metric | Expected lift
- Impact of top fix
- Data gaps: what is missing that could change these conclusions.Example output
Tips and common edits
Why this works
- Requiring both percent lost and users lost stops the model from always blaming the top of funnel, where traffic quality distorts results.
- One fix per transition with a hypothesis and success metric turns analysis into a ready experiment backlog.
- Labeling lift ranges as assumptions and converting them to extra users keeps estimates honest and easy to prioritize.
Use this after a launch or monthly review when you have step counts from your analytics tool. Paste the counts into ChatGPT, Claude or Gemini with the date range and step names.
Common edits
- Add a segment split, such as mobile vs. desktop, and ask for a separate table per segment.
- Ask for a minimum sample size per variant for the top fix so you can plan test duration.
Prompt details
- By
- Prompts-GPT editors
- Published
Did this prompt work?
Keep exploring
More like this
Interpret A/B Test Results and Decide Whether to Ship, Iterate or Kill
Paste your A/B test numbers and get a Ship, Iterate or Kill verdict backed by significance, sample ratio, guardrail checks and a list of what the data cannot prove.
Design an A/B Test Plan with Sample Size Math and Pre-Set Stop Rules
Get a complete A/B test plan with one primary metric, a clear hypothesis, a worked sample size calculation, runtime, and stop rules fixed before launch.
Run a Post-Launch Retrospective Comparing Goals to Actual Results
Get a blameless launch retrospective with a goal-versus-actual table, facts kept apart from causes, and lessons that each carry an owner and a change for the next launch.
Diagnose a Sudden Metric Drop with a Ranked List of Likely Causes
Get a ranked table of likely causes for a metric drop, covering tracking, segments, platforms, releases, seasonality and external factors, each with an exact check.
Write an Event Tracking Plan Where Every Event Maps to a Success Metric
Get a tracking plan table with object_action event names, exact triggers, typed properties and the success metric each event feeds, plus a list of events to cut.
Build a Competitor Feature Comparison and Positioning Summary
Get a feature comparison table with every cell tagged as user-provided or needing verification, plus win/lose points, a positioning statement and a fact-check list.