Meta Conversions API Setup: Recovering Lost Signal After iOS 14

The iOS 14 Problem and Why It Still Matters in 2025

Apple's App Tracking Transparency (ATT) framework, rolled out in iOS 14.5, required users to opt-in to cross-app tracking. Approximately 75% of iOS users opted out. Combined with Safari's Intelligent Tracking Prevention and GDPR cookie consent restrictions in Europe, the Meta Pixel lost the ability to observe 40–60% of conversions on many e-commerce stores.

The result: underreported ROAS, broken audience building, and an algorithm optimising toward an incomplete picture of who was actually converting. Campaigns that appeared to be performing at 2× ROAS were often actually performing at 3–4× — or vice versa.

Meta's Conversions API (CAPI) is the solution — and it's still under-implemented in 2025 despite being available since 2020.

How Meta Conversions API Works

The Meta Pixel fires from the browser — if a user has opted out of tracking, the browser blocks it. The Conversions API fires from your server — it's not subject to browser privacy controls.

By sending conversion events (purchases, leads, add-to-cart) directly from your server to Meta's API, you restore signal that the browser Pixel is missing. The key is sending both browser and server events and deduplicating them so Meta doesn't double-count.

Data flow:

1. User purchases on your site
2. Browser Pixel fires a Purchase event (may be blocked by iOS/Safari)
3. Your server simultaneously sends a Purchase event to the Conversions API with the same event_id
4. Meta deduplicates based on matching event_id — the event is recorded once, from whichever source (or both) successfully transmitted it

Implementation Methods

There are three main ways to implement CAPI:

1. Native platform integrations (easiest): Shopify, WooCommerce, and most major platforms have native CAPI integration. In Shopify: Settings → Customer Events → Add app pixel → Meta Pixel (with CAPI toggle). This handles deduplication automatically and is the right choice for most stores.

2. Server-side GTM (most flexible): Server-side Google Tag Manager routes browser events through a server container before sending to Meta. More technical to set up but gives full control over data enrichment and can serve multiple destination APIs (Meta, Google, TikTok) from one server container.

3. Direct API integration (most powerful): Custom server code sending events directly to Meta's Graph API. Maximum control over event data, timing, and enrichment — but requires developer implementation and maintenance.

Event Match Quality: The Key Metric to Optimise

Event Match Quality (EMQ) is Meta's score for how confidently it can match your server events to Facebook accounts. Higher EMQ = more conversion events successfully attributed = better algorithm optimisation.

Parameters that improve EMQ (in order of impact):

1. Email address (hashed SHA-256)
2. Phone number (hashed)
3. First name + Last name
4. IP address
5. User agent
6. Facebook browser ID (from _fbp cookie)
7. Facebook click ID (from fbclid URL parameter)

For logged-in users at checkout, you'll have email, name, and often phone — EMQ scores of 7–8/10. For anonymous visitors, you'll have IP and user agent — EMQ of 3–4/10. The difference in attribution quality is significant.

Measuring the Impact and Setting Expectations

After CAPI implementation, measure the impact by comparing:

• Event Match Rate: In Events Manager, the percentage of Pixel events matched to a Meta account. Should increase from 40–60% to 70–85% with CAPI.
• Reported conversions: Total conversions reported in Ads Manager should increase — these are real purchases that were previously invisible.
• Audience size: Custom audiences built from purchase events will grow — better audience building feeds better lookalike models.

In our CAPI implementations across 25+ e-commerce accounts, the average reported ROAS improvement after full CAPI setup is 18–32%. This doesn't mean campaigns suddenly got better — it means we're now measuring them correctly.