Time range
A time range control on the trend chart switches between Last 3 months, Last 30 days, and Last 7 days. The default is the last 30 days (the last 7 days on small screens). The metric cards and trend reflect the selected range.Metric cards
Three cards summarize the placement over the selected range:- Trigger Events: how many times the placement was triggered (badged “Placement”).
- Flow Completions: how many flow completions happened across all flows served here (badged “All Flows”).
- Completion Rate: completions divided by triggers, as a percentage. When there is a previous period to compare against, a trend badge shows the change (“+4.2% vs last”).
These are raw event counts. A user who retries a flow five times and completes once counts as five triggers and one completion. Read the completion rate with that in mind, especially at low volume.
Trend chart
The Trend card plots trigger events and completions over time as an area chart, so you can see whether the placement is getting healthier or worse. It needs at least two days of data in the selected window to draw a line; with one day or less it shows “Not enough days of data to draw a trend”.Experiment card
The Experiment card reflects whether an A/B test is bound to the placement. When an experiment is active, the card shows:- The experiment name and a status badge (Running or Paused).
- How long it has been running (“Day 3”) and how many users have been enrolled.
- The lift of the leading variant against control (“+12.4% lift for Variant B vs control”), or “Collecting data” until exposures arrive.
- A list of variants, each with a letter badge (A, B, …), its name, “(control)” where applicable, its traffic percentage, and its completion rate.
- A View Details button that opens the experiment detail.
Flow History
The Flow History card is a timeline of every flow version that has served at this placement, newest first. It is driven by real events, so it reflects what users actually saw. Each entry shows:- The flow name and version number (“v3”).
- A Current badge when the version is the placement’s current default (or a variant of the active experiment), and an Archived badge when the underlying flow has been deleted.
- The date range it served (“Since Apr 10, 2026”, or a start and end date, with “Present” for versions still serving).
- Its impressions and completion rate, with a change against the previous version in percentage points (“+4.7 pp vs prev”).
- A View Funnel button (for versions that have traffic and still exist) that opens a per-version funnel scoped to events from that exact version, at
/flows/placements/{id}/flows/{flowId}withflow_version_idandflow_versionquery parameters.
When you compare completion rates across versions in the timeline, remember the version itself changed. A drop after a new version may be the new content, a seasonal shift, or both. Treat the “pp vs prev” delta as a prompt to investigate, not a verdict.
Configuration card
The Configuration card shows the placement’s current targeting and behavior, with two controls in its header:- A status chip that toggles the placement between Active and Paused when clicked. A paused placement resolves to nothing.
- An Edit button that reopens the placement wizard in edit mode. See Creating a placement.
SDK Integration card
The SDK Integration card shows the trigger snippet for this placement with a copy button. The dashboard displays it asFlowPilot.trigger("<placement_id>"). That is shorthand: the real iOS call is presentPlacement(_:from:). See Presenting placements.
Deleting a placement
The Danger Zone card (and the... menu in the header) has a Delete action. Deleting a placement is permanent and cannot be undone: the trigger stops returning a flow, and all historical analytics, experiments, and flow history for the placement are removed.
Revenue is not shown on placement analytics. The SDK can emit
conversion events, but the dashboard does not aggregate or display revenue at this stage.Common mistakes
- Judging a flow on too little data. A completion rate over a handful of triggers is noise. Wait for volume before acting.
- Comparing versions as if only the flow changed. A version change in Flow History coincides with a content change, so attribute differences carefully.
- Forgetting a paused placement is silent. If analytics flatline, check the status chip; a paused placement serves nothing.