Inigo AnguloContact
Work01How I think02Lab03Writing04
Writing

Instrument the gaps, not the stages: funnel analytics that find the real leak

Inigo AnguloNovember 4, 20266 min readConversion and funnels

Every analytics setup I have inherited counts people at each step of the funnel. Product page, cart, checkout step one, payment, confirmation. The numbers are right and the picture is wrong, because the leak is almost never inside a stage. It is between stages, in the moments the stage counters cannot see.

What stage counters miss

Rebuild the funnel per user

The fix is a query, not a tool. Take every user who reached the final step in a period, and walk their events backwards: how many checkout entries, how many hours between the first and the last, which step they left from each time. Then do the same for users who did not convert. The gap between the two populations is your leak, and it is usually not a step. It is a re-entry with no memory, a price that changed between visits, a form that cleared.

Three dimensions make the query useful: entries per converter (if the median is above one, re-entry is your biggest surface), time to convert (if it is hours, the session model is lying to you), and exit step on the first visit (which tells you where to put the memory).

Then instrument what you found

Only now add events, and add them for the gaps: a checkout_reentry event with the step and the hours since last exit, a cart_restored event, a price_changed_since_last_visit event. These are the events that will tell you whether the fix worked, and they are the ones no default setup ships with.

What this looked like twice

At GetYourGuide it became ten experiments and +11% checkout conversion. The winners all closed the gap: an honest hold timer, two named steps with the order summary persisting between them, and the math shown at the total. At a co-living platform the same query found the drop between application and booking was people returning after the roommate conversation, and the work that followed took booking conversion up 34%. Different products, same blind spot. The ten experiments are written up separately.

A note on tools

Any analytics product can do this if it exports raw events; most cannot do it in their funnel UI, which is why the funnel UI keeps telling you payment is the problem. Session recordings are the fastest way to see a re-entry with your own eyes; the per user query is how you know how many there are.

If your funnel has a leak you cannot find: how a CRO engagement runs. The case: GetYourGuide checkout redesign.

Kit

Copy the checklist: Checkout audit checklist, with a prompt to use it and its changelog.

Working on a funnel like this?

Finding the leak and proving the fix is what the conversion work looks like. Explore a conversion opportunity → Or email me.