Jump to

Share directly to

Product

A funnel tells you where users quit; three replays tell you why

Pair one funnel step with a small set of session replays to turn a drop-off number into a diagnosed, fixable product problem.

OXVO

OXVO Team

Illustration of a three-step funnel connected to a replay filmstrip and a completion marker, representing measuring a drop-off and then watching replays to explain it.

A funnel can tell you that users abandon your onboarding at step three. It cannot tell you whether they left because the button was broken, the copy was confusing, or the permission state was never explained. That gap between "where" and "why" is where most product investigations stall.

OXVO gives you both halves: funnels in Insights measure where a journey loses people, and OXVO Sessions replays show what actually happened in a specific visit. Used together, they turn a drop-off chart into a diagnosed problem your team can fix.

Start with the number, not the recording

Replays are for depth, not for measuring scale. The right order is:

  1. Build a small funnel from the events your app actually sends — a start step, a completion step, and the milestones between.

  2. Find the largest drop in context: same app, same environment, a reporting window you trust.

  3. Only then open replays for visits that reached the failing step.

A few recordings can explain an experience, but they cannot tell you how common it is. Let the funnel size the problem; let the replays explain it.

What to look for in the replays

When you watch a small sample of drop-off visits, you are classifying, not just observing. Most drop-offs fall into recognizable patterns:

  • Loops: the user bounces between two screens, signalling confusion about what to do next.

  • Dead ends: a disabled or unresponsive control with no explanation.

  • Silent failures: a click with no visible state change, sometimes paired with an error in the timeline.

  • Hesitation: a long pause before an unavoidable choice.

Note the first moment the actual behavior diverges from the intended path. That moment, not the abandonment itself, is what the fix should target.

Keep the two tools in their lanes

The pairing works because each tool does what the other cannot:

  • The funnel answers "how many" and "where," and it keeps working after the fix ships so you can compare the same definition over time.

  • The replay answers "what did this visit actually look like," with the route changes, interactions, and technical context your recorder is configured to capture.

Resist two shortcuts: treating three replays as a measurement, or treating a funnel dip as a diagnosis. One describes scale, the other describes cause, and a decision needs both.

Close the loop after the fix

Ship the smallest change that addresses the divergent moment — clearer state, explained constraints, an obvious recovery path — then return to the same funnel with the same filters and window. If the drop persists, sample fresh replays and look again. The goal is not a one-time investigation but a repeatable habit: measure, watch, fix, measure.

Start with one funnel that matters to your business and five replays from its weakest step. That is usually enough to replace a debate with a fix.

Subscribe to get daily insights and company news straight to your inbox.