Open Analytics to review searches, shop results and seller actions. Use the recorded period and source when you share a result with a retailer.
Choose the reporting period
Select 7d, 30d, 90d or 12mo in the toolbar. The default is 30 days. The retailer filter changes the top-retailer list only; it does not filter every panel or the CSV export.
Read the coverage note on each panel. Raw-event panels can cover less than the selected 90-day or 12-month period. A longer selection does not restore events beyond retention.
Server daily counts use UTC dates. Compare equal-length periods with the same brand, market and source. A panel with a different scope is not the denominator for your rate.
Understand the measures
| Measure | What it counts |
|---|---|
| Widget answers | Server answers to widget searches, including zero-result answers. |
| Anonymous searches | Server answers in the hosted brand, product and campaign locator. |
| Shops shown | The full admitted answer population across widget and locator results. |
| Shops shown out of stock | Those shown shops with an out-of-stock observation. |
| Observed seller actions | Recorded directions, call and retailer-website clicks. |
Server answers do not require the same consented browser events as click reports. A combined count is unavailable when a required source lacks coverage. Unavailable is not zero.
The attribution guide explains sources, rate denominators and campaign arrivals. These measure definitions do not imply that an unavailable source has been measured.
Export a report
- Choose the reporting period and check the panel's scope.
- Choose Export CSV for the recorded click events in that period.
- Keep the dates and source with the exported file.
- Use Download value report (PDF) in the estimated reach panel when the button is available.
The event CSV holds timestamps, action types, retailers and UTM fields. A PDF reach estimate is not a sales ledger. A directions click is not a confirmed visit, and a phone click is not a completed call.
For a failed read, retry the request. For partial coverage, report the covered interval. Never replace a missing measurement with zero.
Category benchmark
The category benchmark compares published medians for a category and market. Its action rate is action clicks divided by result views. It is not proven view-to-action attribution or measured sales.
| Publication rule | Requirement |
|---|---|
| Minimum sample | At least five paying suppliers contribute to each measure. |
| Dominance limit | No supplier has 50 percent or more of result views for views and rate, or locations for coverage. |
| Exclusions | Demo tenants and non-paying tenants do not contribute. |
| Withheld result | A withheld measure publishes no figure. A withheld edition publishes nothing. |
Observed seller actions are calls, directions and clicks to the seller's website from the widget. The edition record retains the period, cohort size and per-measure samples for review.
Read related reports
Demand and Weekly pulse provide demand and week comparisons. Weekly pulse uses Stockholm weeks. If a source cannot represent that exact interval, its count is not measured.
A quiet week requires measured zeroes. A failed request, missing source or incomplete interval cannot establish no activity.