Match key
A match key is an identifier inside an event payload - a click ID, hashed email or phone number, IP address with user agent, or mobile advertising ID - that the ad platform uses to tie the event to a real user profile. Events rich in match keys get matched, attributed, and optimized; events without them are delivered but stay anonymous to the platform.
Which keys each platform wants
| Platform | Keys | Requirement |
|---|---|---|
| Snapchat | hashed_email, hashed_phone_number, hashed_ip_address + user_agent, hashed_mobile_ad_id | At least one is mandatory - an event without any of them is rejected with an error; passing all of them improves performance |
| IP address, click_id, email, phone, user agent, external ID, mobile IDs | Share as many as possible; IP address and click ID are strongly recommended | |
| Meta | user_data: em, ph, client_ip_address, client_user_agent, fbc, fbp, external_id | At least one user_data parameter per request; since Graph API v13.0 certain combinations are required |
| TikTok | hashed email and phone in the user object, ttclid click ID | Recommended for matching on Events API events |
The click ID deserves special treatment: it is the only match key that arrives already tied to an ad click, which is why Reddit and Snap both rank it first - but it is also the one most often lost between the landing page and the server event.
Normalize first, hash second
Hashed keys are SHA-256 digests of normalized values: email lowercased with dots stripped from the local part and everything after the plus sign removed; phone reduced to digits with country code; mobile IDs formatted case-sensitively per platform. Every digest is 64 lowercase hex characters. A value hashed without normalization - or double-hashed - does not error out; it simply matches nobody, and the loss is invisible until match rate drops.
The per-platform canonicalization differs enough that reusing one payload builder across networks silently degrades matching; the platform-specific rules are collected in CAPI user data hashing and normalization. The metric that tells you how well your keys are doing is match rate.