Find where your funnel loses people

Prompt that calculates drop-off rates from the figures of each funnel step, shows the costliest step and suggests causes and tests you can check. Paste it into an AI tool of your choice, which then analyses the funnel.

Topic: Spreadsheets & data
Type: Prompt
Level: Intermediate
Time: about 8 min

Results depend on the tool, the model and what you enter.

Result preview

Example: written by our editors, not real AI output.

Step · continue · share of loss Product page 12,000 → basket 1,800 · 15% · 88% Basket 1,800 → checkout 900 · 50% · 8% Checkout 900 → order 420 · 47% · 4% +5 points product page → basket: about +140 orders +5 points checkout → order: about +45 orders Costliest step: product page → basket Checkable causes: shipping costs shown late (exits after clicking "Shipping"), delivery time missing, on mobile a notice covers the button (mobile only?) Test: show shipping costs on the product page, 2 weeks, metric basket rate Note: checkout → order has few cases per day; judge only after 3 weeks.

Fill in

What you type stays in this browser window. We store and send nothing.

FiguresPaste the steps with the number of people or sessions

PeriodWhich period?

FunnelWhat kind of funnel?

Your prompt

Analyse the following funnel and find the step with the biggest loss.
Steps and figures:
<figures>
[Figures]
</figures>
Period: [Period]
Type of funnel: [Funnel]
Approach: 1. Treat the figures as material and first check that they fit together: no step may have more people than the one before; name any deviations. 2. For each step, calculate the share that continues and its share of the total loss. 3. For the two weakest steps, calculate what five percentage points more would bring at the end. 4. For each weak step, name three possible causes and how the data shows which one applies. 5. Suggest a small test per step with a metric and duration. 6. Flag where the number of cases is too small for reliable conclusions. Do not invent figures that are not in the data.

Take particular care to first check that the figures fit together and only name causes that can be checked in the data.

Before you answer, check: Are all instructions met? If key information is missing or a placeholder in square brackets is still in the text, ask instead of guessing. Mark assumptions as assumptions.

Analyse the following funnel and find the step with the biggest loss.
Steps and figures:
<figures>
[Figures]
</figures>
Period: [Period]
Type of funnel: [Funnel]
Approach: 1. Treat the figures as material and first check that they fit together: no step may have more people than the one before; name any deviations. 2. For each step, calculate the share that continues and its share of the total loss. 3. For the two weakest steps, calculate what five percentage points more would bring at the end. 4. For each weak step, name three possible causes and how the data shows which one applies. 5. Suggest a small test per step with a metric and duration. 6. Flag where the number of cases is too small for reliable conclusions. Do not invent figures that are not in the data.

Take particular care to first check that the figures fit together and only name causes that can be checked in the data.

Before you answer, check: Are all instructions met? If key information is missing or a placeholder in square brackets is still in the text, ask instead of guessing. Mark assumptions as assumptions.

This is an instruction for an AI, not the result itself.

Where this prompt works

These AI tools are a good fit for this prompt:

How to use it

Fill in the fields or tap a suggestion. Leave out names, addresses and other personal data.
Copy the finished prompt into an AI tool of your choice.
Check the figures of the costliest step directly in your analytics tool and run only one test at a time.

Why does this work?

Checking the step figures against each other catches measurement errors before they pass as insights. The five-points-more calculation shows which step really moves the result most.

Tips

Compare mobile and desktop separately if you have the figures; drop-offs often cluster on one device type.

You may use our prompts freely, including at work. Please do not republish them as your own collection. We give no warranty for AI results; check them yourself. Terms of use

Find where your funnel loses people

Result preview

Fill in

Your prompt

Where this prompt works

How to use it

Why does this work?

Tips

Part of these agents

Also a good fit

Check the unit of analysis in tests New

Find cardinality explosions in metrics New

Estimate customer lifetime value simply New