Funnel Triage

Optimize the biggest leak first. Conversion math compounds — fixing 48pt beats fixing 4pt.

How to apply it

Sort steps by absolute drop, not percentage of step. Fix the leak with the largest impact.

Funnel Triage in practice: a concrete example

Fix the 48-point leak, not the 4-point one

A checkout loses 5% at the cart, 8% at shipping, and 48% at account creation. It is tempting to polish the small friction you just noticed, but account creation is where the money actually leaks. Sort the steps by the absolute number of people lost, attack the biggest drop first (here, add guest checkout), and the same effort moves far more revenue. Conversion is multiplicative, so the largest single leak dominates the result.

Common Funnel Triage mistakes

Chasing the friction you personally noticed rather than the step the data shows is bleeding the most users.
Ranking steps by their percentage drop in isolation instead of the absolute count of people lost, which is what compounds.
Optimising a late step while a much bigger earlier leak means most users never reach it anyway.

When Funnel Triage doesn't apply

Triage assumes you can measure the funnel. Early on, with little traffic, the numbers are too noisy to trust and qualitative research finds the real blockers faster. And a leak that looks small today can be worth fixing first if it sits on a step you are about to send far more traffic to.

Practise it on a real challenge

Activation Funnel — find where the users leak

Funnel: 100% land → 80% sign up → 70% verify email → 22% complete profile → 18% activate.