Skip to main content
Overall Browser Coverage is the percentage of your Meta Pixel (browser) events that Meta matched and deduplicated with a Conversions API (server) event using a given dedupe key. Meta calculates it. Upstack imports and averages it.

​
Formula

Overall Browser Coverage = AVERAGE ( Meta’s overall_browser_coverage_from_dedupe_key percentage ) across the imported rows in your selection

​
How it differs from Browser Dedupe Coverage

Meta reports three numbers for each event name and dedupe key (such as event_id, external_id, or fbp): The first two ask “does the key exist on the event?” This one asks “did Meta actually pair the browser event with a server event?” A browser event can carry an event_id and still not be deduplicated if no server event arrives with the same value. Meta also notes that this percentage is incremental for each dedupe key. The first key covers most events, and each further key adds only the extra events it matched.

​
How Upstack calculates it

  1. Upstack pulls Meta’s Dataset Quality data for your dataset and stores one row per event name and dedupe key each time it collects.
  2. The metric takes a plain average of the stored percentages in your date range and filters. It doesn’t weight by event volume.
Because the per-key values are incremental, mixing keys in one average gives a number that’s hard to read. Look at it per dedupe key before drawing conclusions.

​
Worked example

Meta reports these values for Purchase on one collection: Averaging the Overall Browser Coverage rows gives 47%, which looks alarming. Read per key, the picture is good: 91% of browser purchases were deduplicated by event_id, and external_id matched another 3%, for 94% in total. The real gap is the 7 points between 98% (key present) and 91% (key matched), browser events that carried an event_id but weren’t paired with a server event.

​
How to interpret it

  • Key present but coverage low: browser events have the key, but matching server events are missing or use a different value. Check that the browser and server copies of each event send the same event_id.
  • Low on one event name only: look at how that event is sent server-side. Other events being fine rules out a site-wide problem.
  • Don’t add a second Conversions API sender to raise coverage. See Low rate of Pixel events covered by Conversions API for why that causes duplicates instead.

​
FAQ

Browser Dedupe Coverage is the share of browser events that include the key. Overall Browser Coverage is the share Meta actually deduplicated against a server event using that key.
A common cause is an average that mixes dedupe keys. Each key after the first reports only its incremental coverage. Look at the event_id rows on their own.
No. Meta calculates it in the Dataset Quality API. Upstack stores and averages Meta’s figures.
See all EMQ metrics →