match-qualitylisted

# Match quality The platform scores how well it can tie each server event to a person. A low score means fewer of your conversions are attributed, which understates the channel and starves optimisation. It also means almost nothing at low volume. Both halves matter. ## What actually moves the score Not the count of parameters. Their determinism. | Signal | Why it ranks where it does | |---|---| | Click identifier from the ad click | Deterministic. One parameter, and the strongest one you can send | | Browser identifier cookie | Deterministic within the browser | | Your own stable customer ID | Strong, and it costs you nothing you do not already hold | | Email, phone | Strong when normalised correctly, useless when not | | Name, city, postcode, date of birth | Probabilistic. Fills gaps, rarely rescues a score | | IP address and user agent | Weakest. Present by default, do not count on them | **Adding three weak parameters rarely beats fixing one strong one.** If the click identifier is missing from server events, that is the whole task. ## The bugs that produce a populated field matching nobody These all pass inspection. The field is there, it has a value, it never matches. - **Double hashing.** A value hashed on the client and hashed again on the server produces a valid-looking digest of a digest. - **Hashing an empty string.** A missing field sent as a hash of `""` looks populated in every debugger and matches nobody. Omit the field instead. - **Normalising afte