Onboarding funnel analytics
See where users move forward and where they fall away. Every guide reports how many started, completed, or dismissed it — and a step-by-step funnel shows exactly where people drop off.
Start to finish
Started, completed, dismissed, and completion rate at a glance.
Per-step funnel
Spot the exact step where users drop off.
Act on it
Tighten the weak step and watch completion climb.
Live in three steps.
Publish a guide
Ship it to your users.
Watch the funnel
See completion and drop-off per step.
Improve
Fix the leaky step and re-measure.
Finding the step that kills your tour.
A tour with 60% overall completion is not a 60% tour — it is usually four strong steps and one broken one. The step funnel is how you find the one.
Read the headline first
Reports shows started, completed, completion rate, and dismissed. Say 500 started, 290 finished — respectable, but the funnel knows more.
Walk the funnel
Step 1 holds 95%, step 2 holds 91%, step 3 drops to 55%. There is your defendant — one bar in the chart, not a vague sense that "the tour underperforms".
Diagnose the drop
Step 3 is anchored to a button only admins see, and its fallback is "wait". Non-admins stare at nothing and leave. Change the fallback to skip — or target the tour at admins.
Confirm the fix
Republish and watch the same funnel: step 3 now holds 90%, overall completion jumps. The loop from symptom to fix to proof took an afternoon, not a quarter.
What to expect: Drop-off becomes a named, fixable step instead of a mystery. The usual culprits, in order: an anchor users don't have, a form asking too much, a tour two steps too long.
When not to use funnels & completion.
No pattern fits every job. Here’s where we’d point you at something else — sometimes ours, sometimes not.
Proving business outcomes
The funnel measures guide progression, not revenue. "Completing the tour correlates with retention" needs your product analytics — export the events to your warehouse and join them there.
Judging single-step guides
A banner has no funnel to walk — started, clicked, dismissed tells the whole story. Funnels earn attention on multi-step flows where the drop can hide.
Tiny samples
A funnel over 14 users is noise wearing a chart. Let a guide accumulate a few hundred starts before redesigning around its shape — or run an A/B test when you need a real verdict.
What counts as a completion?
Reaching the final step, or any button with the explicit complete action — so a three-step flow with an early "Done" button counts honestly. Dismissals are tracked separately, per step, which is what makes the funnel diagnostic.
Where does the funnel data come from?
The SDK emits step-level events as users move through guides; the analytics pipeline aggregates them per flow. No setup on your side — every published guide has a funnel from its first view.
My guide shows all zeros — broken?
Almost always one of three things: the flow is not published, the SDK is not installed on the pages where it should fire (the Install page shows detection), or the audience/trigger never matches. Check in that order.
Starts are lower than I expected — why?
Audience, trigger, and frequency all gate reach: a "once per user" flow will not re-fire for people who already saw it, and rate limits may defer guides in busy sessions. Reports plus the configuration tab usually explains the gap in a minute.
What is a good completion rate?
Depends on the surface: interactive tours that advance on real clicks routinely beat passive ones; welcome modals convert high; unsolicited surveys run lower. The honest benchmark is your own funnel last month — that is the number to beat.
See it on your own app.
Ship your first guide in minutes. Free to start — no credit card.