What died with third-party cookies
Third-party cookies allowed ad platforms to track users across websites without any cooperation from those websites. A user visiting your Shopify store and a user visiting a news site could be linked by a shared cookie from the same ad network. This cross-site tracking is what made retargeting work at scale and what powered platform attribution models.
Safari has blocked third-party cookies since 2017. Firefox followed. Chrome's full deprecation, combined with Google's Privacy Sandbox changes, means the last major browser holding the old model together is gone. The data infrastructure of 2018 to 2022 DTC advertising no longer exists.
First-party identity resolution
First-party identity resolution means you are matching users across sessions using data they gave you, not data a third party tracked without their knowledge. The primary signals are email address, phone number, and customer ID from Shopify. When a customer creates an account, checks out, or subscribes to your email list, you capture an identifier you can use.
The match rate across sessions depends on how often customers identify themselves. A DTC brand with 40 percent returning customer rate and strong email capture at checkout might achieve 55 to 65 percent first-party identity coverage. That leaves 35 to 45 percent of sessions genuinely unidentifiable with certainty. Probabilistic methods fill part of that gap.
Probabilistic identity stitching
Probabilistic methods use device signals, IP address patterns, and behavioral fingerprints to connect sessions without a deterministic identifier. The output is a confidence score, not a certainty. A session that matches on device type, browser, IP subnet, and behavioral timing patterns to a known customer might receive an 82 percent confidence score. You should treat that differently from a 95 percent deterministic match.
Layer Insights applies three tiers of confidence and reports them separately. Tier 1 is deterministic. Tier 2 is high-confidence probabilistic above 80 percent. Tier 3 is medium-confidence probabilistic between 50 and 80 percent. Anything below 50 percent is treated as anonymous. Revenue attribution in reports reflects these tiers with appropriate uncertainty ranges.
What this means for your retargeting
Retargeting audiences built on third-party cookies were cheap to create and broad in reach. First-party retargeting audiences built on email lists and customer IDs are smaller but far more accurate. A retargeting campaign to your actual purchasers using Meta's customer list upload will perform differently from a pixel-based retargeting audience.
The trade-off is scale versus precision. First-party audiences are typically 30 to 60 percent smaller than equivalent cookie-based audiences. But the purchase intent signal is stronger because you are reaching people who actually bought from you, not people who vaguely matched a behavioral profile.
Building your identity resolution infrastructure
The foundation is a clean customer data layer in Shopify: consistent email capture at checkout, account creation incentives, and a single customer ID that persists across sessions. Layer this with server-side event tracking that sends that customer ID alongside every purchase event to your CAPI endpoints.
On top of that, implement a consent-gated data collection strategy. Customers who opt into marketing give you permission to use their data for personalization and attribution. Customers who opt out should be tracked anonymously. Your consent rates directly affect your identity resolution coverage, which directly affects your attribution accuracy.
---