Match quality

An ad platform can only optimize against a conversion it can tie back to a person who saw an ad. Send a purchase with nothing but an IP address and it may match nobody; send the same purchase with a click ID and a hashed email and it almost certainly matches.

Match quality is a 0–100 score summarizing how much matchable signal each conversion carried when it left for a destination.

How it’s calculated

Each identifier a platform accepts carries a weight. The score is the summed weight of the identifiers actually present, divided by the platform’s summed weight for click ID + email + phone + IP + user agent, expressed as a percentage.

That denominator is the point. It’s the realistic best case for a well instrumented web conversion — not the sum of every field the platform will accept. Fields outside the denominator (a browser ID, an external ID, a full postal address) still add to the numerator, which is why a very complete event can push the score high.

Weights per platform

Each platform values identifiers differently, so the same event scores differently at each destination.

The dashboard bands scores as Excellent, Good, Fair, and then red.

Improving it

The weights tell you exactly where to spend effort, and it’s the same short list everywhere.

• Collect email at conversion. Worth 30 at every destination, and it’s the identifier that survives longest — a click ID expires, an email doesn’t. See Identity resolution.
• Collect phone where it’s natural. Another 25, and it matches people whose email you don’t have.
• Keep click IDs flowing. 35–40 points, and the highest-confidence match there is. Don’t strip query strings in redirects, and let click backfill attach earlier clicks to conversions that arrive without one.
• Send server-side events with real context. IP and user agent are only 8 each, but they’re free from a browser event and frequently missing from backend ones.
• Fill the address fields for Meta. Zip, city, state and country add up to 17 at Meta alone.