How a conversion funnel works, with the maths
A funnel describes the steps people take from first hearing about you to buying, and how many drop out at each step. Because the steps multiply, a small change at one step changes the final number more than you might expect.
Funnel calculator
Enter how many people start, then the percentage who move on at each step. Leave a step blank to skip it.
The steps of a typical funnel
Names vary, but most funnels for a small online business look like this:
- Visit: someone lands on your site or page.
- Interest: they sign up, start a trial or request a quote.
- Activation: they use the thing, open the quote, or book a call.
- Purchase: they pay.
The step rate is the share of people at one step who reach the next. The overall rate is the share of the people at the top who reach the bottom.
The formula
overall rate = rate 1 × rate 2 × rate 3
Worked example (made-up numbers). A fictional online pottery class gets 10,000 visits. 4% sign up for the free taster, 25% of those attend it, and 10% of attendees book the paid course.
- Sign-ups: 10,000 × 0.04 = 400
- Attend: 400 × 0.25 = 100
- Book: 100 × 0.10 = 10
- Overall rate: 0.04 × 0.25 × 0.10 = 0.001, which is 0.1%
Which step should you improve?
Because the rates multiply, raising any one step by the same proportion has the same effect on the final count. In the example, raising any single step by a fifth (4% to 4.8%, 25% to 30%, or 10% to 12%) gives 12 bookings instead of 10.
What differs is how hard each change is. Compare two options measured in percentage points:
| Change | Calculation | Bookings |
|---|---|---|
| None | 10,000 × 0.04 × 0.25 × 0.10 | 10 |
| Sign-up 4% → 5% | 10,000 × 0.05 × 0.25 × 0.10 | 12.5 |
| Booking 10% → 11% | 10,000 × 0.04 × 0.25 × 0.11 | 11 |
| Double the visits | 20,000 × 0.04 × 0.25 × 0.10 | 20 |
One percentage point at a low-rate step is a large relative gain (4% to 5% is a quarter more). The same point at a higher-rate step is a smaller relative gain. Look for the step where a realistic fix produces the biggest relative jump.
How to measure it by hand
- Pick one time window, such as a calendar month, and count each step within it.
- Count people, not page views, where you can.
- Write the counts in a column and divide each by the one above it.
- Watch for small numbers: a step with 12 people can swing wildly from month to month by chance.
Common mistakes
- Adding rates together instead of multiplying them.
- Comparing a funnel from one month with sign-ups from another.
- Chasing more traffic when a broken step loses most of it.
- Declaring a winner from a handful of conversions. Use the A/B test guide to check sample sizes.
Questions people ask
What is a good conversion rate?
It depends heavily on the product, the price and where visitors come from, so a single benchmark is not meaningful. Compare your funnel with itself over time.
Does the calculator store my numbers?
No. It runs in your browser and keeps nothing.