Analytics
Analytics show whether your flows are working - views, completion, and where people drop off - so you measure activation instead of guessing. This page defines every metric.
Two views
- Workspace analytics (the Analytics page) - totals and trends across all your flows.
- Per-flow analytics (a flow's Analytics) - the numbers for one flow, including per-step drop-off.
The metrics
| Metric | What it counts |
|---|---|
| Starts / views | How many times a flow was shown (each display). |
| Unique users | How many distinct people saw it (a person seen twice counts once). |
| Completions | How many times a flow reached its end (a tour finished, a checklist fully done). |
| Completion rate | Completions ÷ starts - the share that finished. A unique-user version (unique completers ÷ unique starters) shows the truer "share of people who complete." |
| Dismissals | How many times, and at what point, people closed a flow. Early dismissals signal friction. |
| Per-step drop-off | For a multi-step flow, exactly which step loses people - the step to fix first. |
A completer is always counted as a starter too, so the unique completion rate can never exceed 100%.
Per-type metrics
| Flow type | Extra metrics |
|---|---|
| Tour | Per-step views and drop-off across the sequence. |
| Checklist | Per-item completion - which task stalls people. |
| Announcement | Dismissals and call-to-action clicks. |
| Tooltip | Views and dismissals (a tooltip has no multi-step sequence). |
| Survey | Responses, response rate, and how far people got through the questions. A survey has no completion step, so it reports answers rather than a funnel - see Surveys and NPS. |
Date range and comparison
Pick a preset range (last 7 / 30 / 90 days, and more) or a custom range. Turn on compare to see the same metrics against the previous period, so you can tell whether a change moved the numbers. The workspace view also shows a daily trend line and breakdowns by plan and conversion.
While an A/B test is running on a flow, the range list gains Current A/B test and selects it by default, so the per-version figures cover that run and not whatever came before it. Pick any other range to look wider - useful after a restart, when the current run is deliberately counting from zero.
Surveys in the workspace view
If any survey has been answered in the range, the workspace view adds a Surveys & NPS panel: how many people answered, and a single NPS blended across every NPS question in the workspace. The blend is worked out from the underlying promoter and detractor counts, not by averaging each survey's score, so a survey with five responses does not weigh as much as one with five hundred.
In the flow comparison table a survey's Completions and rate columns show people who answered and the response rate. A survey records no completion event, so the funnel columns would otherwise read as zero.
Export
Export analytics to CSV for your own reporting or to combine with other data. A survey's own answers export separately, from the Answers panel on its analytics page - one row per person, one column per question.
How it's collected
The runtime sends small, batched beacons for events (start, complete, dismiss, item-complete) as they happen.
Metrics are keyed to each end user (anonymously by default, or by the id you send with
identify), scoped to your workspace, and never shared.
Use per-step drop-off to iterate: find the step where completion falls, rewrite or reorder it, republish, and watch the number. Fixing the single worst sticking point usually moves activation the most. To be sure a rewrite is what moved it - rather than a busier week - run the two versions against each other with an A/B test.
Related
- A/B testing - run two versions of a flow and see which one wins.
- Targeting - decide who sees a flow in the first place.
- Runtime API -
identify, and the events behind these numbers.
Last updated: Sep 2, 2026