Match rate by platform: Meta, TikTok, Snap, and Google compared
A brand running across Meta, TikTok, and Snap came to us with the same complaint most multi-platform teams have.
Their match rate looked fine on Meta. TikTok felt unstable. Snap was showing modeled conversions they did not understand. Google was a black box.
They assumed the problem was campaign structure. It was not. Each platform was receiving different signal quality because each platform weights identifiers differently, reports differently, and penalizes signal gaps differently.
Understanding match rate on one platform does not tell you what is happening on the others.
Why match rate is not the same number across platforms
Every ad platform tries to connect your conversion events to real users in its system. But each one uses different identifiers as primary signals, applies different matching logic, surfaces the result under a different name, and degrades in different ways when the signal is weak.
A brand sending identical event data to Meta and TikTok will likely see different match rates on each. That is not a reporting inconsistency. It reflects how each platform's AI is built and what it needs to learn effectively.
The practical consequence: fixing match rate is not a single fix. It is four separate conversations, one per platform. The starting point is understanding how server-to-server tracking improves matching on each one.
Meta: Event Match Quality
Meta calls it Event Match Quality, or EMQ. It is scored on a scale of 1 to 10 inside Events Manager, not expressed as a percentage.
EMQ reflects two things simultaneously: how often Meta can connect an event to a user, and how rich the identifier data attached to each event is. A score of 7 or above is generally where optimization starts to work properly. Below 6, campaign learning is being compromised.
What Meta weights most heavily
Email address is Meta's strongest identifier for web conversions. It needs to be lowercased, whitespace-trimmed, and SHA-256 hashed before transmission. The normalization step matters more than most teams realize. A customer who enters their email with a capital letter at checkout and a lowercase version in their Meta account becomes two different people if the hashing is applied before normalization.
The fbclid, Meta's click identifier, is the second critical signal. It ties a conversion directly back to an ad click. When fbclid is preserved through the full conversion journey and passed server-side, match quality improves significantly. When it gets stripped by redirects, landing page transitions, or iOS Safari's link decoration stripping, that direct match path breaks.
Phone number is a strong supplementary signal, particularly for markets where phone-based account registration is more common. In the GCC, phone number often outperforms email as a matching identifier because of how users register on Meta's platforms.
What weak EMQ actually costs you
Meta's AI needs matched events to build and refine its audience models. Low EMQ means the AI is making targeting decisions from a smaller, less reliable sample of your actual buyers. Lookalike audiences become less accurate. Advantage+ campaigns have less to learn from. Retargeting pools shrink.
The damage accumulates over time. A campaign running on poor signal for several weeks builds a model that is increasingly misaligned with your real customer base.
Where to check it
Events Manager > your pixel or CAPI source > Overview tab > Event Match Quality score per event type.
TikTok: match rate percentage
TikTok surfaces match rate as a direct percentage in Events Manager, which makes it the most transparent of the four platforms on this metric.
The number you want to see is above 70%. Below 50%, TikTok's optimization engine starts visibly struggling. The algorithm cannot build reliable audience segments, video completion signals become a stronger optimization proxy than conversion signals, and CPAs drift upward without a clear campaign-level explanation.
What TikTok weights most heavily
TikTok's matching relies heavily on email and phone, but the identifier hierarchy shifts depending on the market. In Southeast Asia and the GCC, phone number tends to outperform email because TikTok account registration in those regions is more phone-centric. Sending both is always correct. Sending only email in a phone-primary market is a quiet, consistent match rate drain.
TikTok also uses its own click identifier, ttclid. Like fbclid on Meta, it creates a direct link between an ad click and a conversion event. Unlike fbclid, it is less likely to be stripped by browser privacy features, but it still needs to be captured at click time and passed through to the server-side event.
What weak match rate actually costs you on TikTok
TikTok's algorithm is fast-learning but sensitive to signal gaps. When match rate is low, TikTok defaults to optimizing on engagement signals rather than conversion signals. That means it finds users who watch and interact with your content, not necessarily users who buy.
The consequence is volume without conversion. CPM stays stable. Clicks look fine. ROAS drops.
Lumi saw a 170% increase in app installs and a 50% reduction in CPI on Snap after improving signal quality and delivery. Directionally, the same pattern applies on TikTok: when the platform can match more events to real users, it finds more of those users efficiently.
Where to check it
TikTok Ads Manager > Assets > Events > your event source > Match Rate column.
Snap: matched vs modeled
Snap is the most distinctive of the four in how it communicates signal quality. Instead of a score or a percentage, Snap separates conversions into two categories: matched and modeled.
Matched conversions are events Snap connected to a real user profile with sufficient confidence. Modeled conversions are events where Snap did not have enough signal to match directly and instead estimated based on probabilistic patterns.
Most teams see modeled conversions and assume they are fine. They are not the same as matched conversions, and treating them as equivalent overstates true signal quality.
What Snap weights most heavily
Snap's matching is heavily email and phone dependent, with phone being particularly strong given Snap's mobile-first user base. Snap also uses its own click parameter, which needs to be preserved from click through to conversion.
One common Snap-specific issue: phone number formatting. Snap requires E.164 format, meaning the full international format including country code, with no spaces or punctuation. A UAE phone number that arrives without the +971 prefix, or with spaces, fails to match even if the underlying number is correct.
What a high modeled ratio actually costs you
When modeled conversions dominate your Snap reporting, two things happen. First, your reported ROAS becomes partially estimated rather than directly measured. The numbers can look healthy while actual performance is softer than it appears. Second, Snap's algorithm has less clean signal to optimize from, which makes audience targeting less precise over time.
The distinction between matched and modeled is Snap's way of telling you how much of your signal it trusts. Pay attention to it.
Where to check it
Snap Ads Manager > Assets > Events > your pixel or CAPI > Conversion breakdown by matched vs modeled.
Google: Enhanced Conversions and diagnostic health
Google does not surface a match rate number in the way the other three platforms do. Instead, it shows conversion diagnostic status and Enhanced Conversions coverage, which together tell you whether your signal is flowing correctly.
Enhanced Conversions is Google's mechanism for recovering signal lost to cookie deprecation and consent restrictions. It works by sending hashed first-party data, email address primarily, alongside standard conversion tags. Google then uses that data to match conversions back to signed-in Google accounts.
What Google weights most heavily
Email is the dominant identifier for Enhanced Conversions. Unlike Meta, where phone number is a meaningful supplementary signal, Google's matching infrastructure is more heavily built around email because of how Google account authentication works.
Gclid, Google's click identifier, is also critical. It links a conversion directly to a specific ad click and campaign. When gclid is captured correctly and passed to the conversion event, Google can attribute and learn from that conversion with high confidence. When it is lost, Google falls back to probabilistic modeling.
What weak Enhanced Conversions coverage actually costs you
When Enhanced Conversions is implemented but coverage is incomplete, Google's Smart Bidding has less first-party data to calibrate against. Target ROAS and Target CPA strategies become less accurate because they are working from a signal set that does not reflect the full conversion picture.
The symptom is familiar: bid strategies that behave erratically, CPAs that drift without a clear cause, and Performance Max campaigns that allocate budget in ways that are hard to explain.
Where to check it
Google Ads > Goals > Conversions > Diagnostics tab. Look for Enhanced Conversions status, coverage percentage, and any flagged implementation issues.
What multi-platform signal management actually requires
Running across Meta, TikTok, Snap, and Google means maintaining signal quality against four different measurement systems, four different identifier hierarchies, and four different ways of telling you when something is broken.
Most teams check one platform when performance dips. They fix what they find there. They do not check the others.
The right approach is to monitor all four simultaneously, with platform-specific benchmarks for each. A drop in Meta EMQ, a rise in Snap modeled conversions, a TikTok match rate below 60%, and incomplete Google Enhanced Conversions coverage can all be happening at the same time, in the same account, for different reasons.
Signal quality is not a campaign setting. It is infrastructure.
And infrastructure that is not being watched is infrastructure that is quietly degrading.
If you are running across Meta, TikTok, Snap, or Google and have not audited your signal quality recently, it is worth doing before your next budget cycle.
