A funnel answers “where do people drop out of this journey?”. It does not answer “why?” That is still Wonder work, but it tells you where to look first.
Build the funnel from intent
Define steps by what the user is trying to do, not by page names:
- Lands on a product page
- Adds to cart
- Starts checkout
- Enters shipping details
- Completes payment
Check that every step is tracked consistently across devices and that consent banners aren’t hiding a large share of users before you trust the numbers.
Rank by absolute loss
| Step | Users | Continue rate | Users lost |
|---|---|---|---|
| Product page | 100,000 | 8% | 92,000 |
| Add to cart | 8,000 | 55% | 3,600 |
| Start checkout | 4,400 | 70% | 1,320 |
| Shipping details | 3,080 | 78% | 678 |
| Payment complete | 2,402 |
Illustrative numbers. The biggest percentage drop is not always the best opportunity. Here the product page loses the most people by far, but much of that loss is browsing behaviour you can’t and shouldn’t eliminate. The checkout steps lose fewer people, but those people have shown strong intent. Weigh absolute loss, intent and how fixable the cause is.
Segment, carefully
Compare a handful of segments chosen in advance: device, traffic source, new versus returning, country. Look for large, consistent gaps (for example, mobile checkout completion far below desktop).
Avoid slicing into dozens of segments and celebrating the most dramatic one: with enough slices, some differences appear by chance alone.
Turn leaks into questions
Each leak becomes a research question for recordings, surveys or usability tests:
- “What are mobile users doing in the 40 seconds before they leave the shipping step?”
- “What would have made you complete your order today?” (exit survey on checkout)
Those answers become hypotheses.