Documentation
Website Banner Analytics
Measure website banner impressions, clicks, dismissals, unique visitors, click-through rate, and variant performance with SpryUI analytics.
Monthly allowance and paused analytics
Your plan's monthly impression allowance is shared across all projects and environments in your account. Accepted impressions count toward it; impressions from localhost and loopback addresses do not.
- When the allowance is reached, banners continue to display, subject to normal account access, project status, targeting and delivery settings.
- Recording pauses for all new analytics events, including impressions, clicks and dismissals. Existing analytics remain available within their retention period.
- The monthly usage count stops increasing while recording is paused. It does not measure additional banner displays during that time.
- Events during the pause are not queued or backfilled later, so reports will have a gap.
- Recording resumes when you upgrade to a plan with remaining allowance, or at the start of the next UTC calendar month. This reset is separate from your subscription payment date.
- Deleting old analytics does not restore consumed monthly allowance.
Core metrics
Impressions: count of render events.Clicks: clicks on actionable elements (anchors/buttons), not dismiss actions.CTR:clicks / impressions.Unique Users: distinct user/session keys on impression events.
Attribution and data source
Snippettext
Attribution modes:
- attributed: all variants have variant-attributed impressions
- mixed: some variants attributed, others message_rollup fallback
- message_rollup: no attributed impressions availableVariant rows use the selected variant ID when available, then fall back to message-level rollup.
Snippettext
coverage = variantImpressions / totalImpressions
where:
- variantImpressions: impressions with variantId
- totalImpressions: all impressions for the messageTrend and range behavior
- Dashboard analytics default to a recent reporting window.
- API reporting windows are capped to protect runtime performance.
- Impression trend series is derived from message and variant events.
- Variant table can still show message_rollup fallback when attribution is incomplete.
Confidence guidance in dashboard
- < 100 impressions: low confidence.
- 100-499 impressions: directional read only.
- 500+ impressions: sufficient for directional winner calls.