Why pixel-only tracking is no longer enough
A browser pixel fires when a page loads and the browser cooperates. In 2026, those two conditions are increasingly unreliable. Safari's Intelligent Tracking Prevention limits cookie lifespans to 7 days. Firefox blocks third-party cookies by default. iOS users who opt out of tracking produce zero pixel events on Meta's side. Chrome's Privacy Sandbox changes the rules further.
The result is a growing gap between what actually happened and what your pixel recorded. For some DTC brands running heavy Meta spend, pixel-reported conversions are 20 to 40 percent below actual. That gap makes optimization signals unreliable and ROAS calculations meaningless.
The dual-dispatch architecture
Dual-dispatch means every conversion event fires twice: once from the browser pixel to each platform's standard endpoint, and once from your server directly to each platform's CAPI endpoint. The server-side event includes data the browser cannot: first-party customer ID, email hash, phone hash, and order value from your Shopify backend.
The two events are deduplicated by the platform using an event ID you set. That event ID must be identical in both the pixel and the server event or the platform will count the same purchase twice, which inflates your reported conversions even further than before. This is the single most common error in CAPI implementations.
Platform differences you need to know
Each of the six platforms has different CAPI implementations, rate limits, and required parameters. Meta is the most mature and requires email hash, phone hash, and external ID for optimal match rates. Google's Enhanced Conversions works via Google Tag Manager or the API and focuses on email matching against signed-in Google accounts.
TikTok Events API and Pinterest Conversion API both require SHA-256 hashed PII and have stricter rate limits than Meta. Snapchat and X have less mature server-side implementations but still improve signal quality meaningfully for brands with active spend on those platforms.
Match rates and what they tell you
After implementing dual-dispatch across all six platforms, the metric to track is event match rate. This is the percentage of server-side events the platform can match to a logged-in user. Meta typically achieves 60 to 80 percent match rates with email hash alone. Adding phone hash and external ID pushes this above 85 percent for most DTC brands.
Match rates below 50 percent suggest data quality issues: inconsistent email formatting, missing hashes, or sending events for anonymous sessions that never collected any identifying information. The fix is always upstream. Improve your data collection, not your CAPI implementation.
What changes after you go live
The first thing you will notice is that your reported ROAS drops. This is expected and correct. The platform is now deduplicating events it previously counted multiple times from the pixel. A reported Meta ROAS of 4.8 becoming 3.1 after CAPI is not a crisis. It is the real number.
The second thing you will notice is that your campaign optimization improves. Platforms optimize toward conversion signals. Better signals mean the algorithm has better information. Most brands see CPAs improve by 10 to 20 percent within four to six weeks of clean dual-dispatch data flowing through.
---