Meta Ads & iOS 14: How to Fix Conversion Tracking

In April 2021, Apple introduced the ATT (App Tracking Transparency) prompt in iOS 14.5 — "Do you want to allow tracking?". Over 75% of users tapped No. Overnight, the Meta pixel lost access to roughly half of its signals. ROAS on Meta dropped 30–50% for nearly every advertiser. Three years on, solutions exist — but most advertisers still aren't using them, just as they aren't using the Custom + Lookalike audience structure that has been adapted to the new constraints. This guide walks through all of it.

What iOS 14 Actually Broke

Before iOS 14, the Meta pixel tracked everything — clicks, site visits, add-to-cart events, purchases — tied to a user ID. Attribution was nearly perfect. After iOS 14, for users who decline tracking:

• The pixel no longer sees a user ID — only aggregated numbers
• The attribution window was reduced from 28 days to 7 days
• The number of aggregated events is capped (max 8 per domain)
• Conversion tracking has a 24–72 hour lag (modelled data, not real-time)

The result: the ROAS you see in Meta Business Manager is dramatically understated. True ROAS is typically 30–60% higher than what the pixel reports.

Solution 1: Conversions API (CAPI)

CAPI is server-side tracking. Instead of sending the signal through the browser (where iOS can block it), it is sent directly from your backend to Meta. Apple cannot block server-to-server communication.

How to Install It

Three options, ordered by complexity:

• Shopify, WooCommerce, Magento: native CAPI integration — enable it in 5 minutes
• WordPress: a plugin such as PixelYourSite Pro handles it out of the box
• Custom backend: build the API call directly in your code (2–3 hours of developer time)

The rule: CAPI works alongside the pixel, not instead of it. Both send signals and Meta deduplicates them. This way you cover both in-browser tracking and the server-side fallback.

What You Get

• 20–30% more conversions attributed to your ads
• A stronger signal for the algorithm, resulting in better-optimised campaigns
• No data loss for iOS users who declined tracking

Solution 2: Aggregated Event Measurement (AEM)

Meta's answer to iOS 14 was AEM — a system of up to 8 prioritised conversion events per domain. You need to configure them manually in Events Manager: list your conversions, rank them by importance, and Meta will optimise in that order.

A standard ecommerce setup:

• Purchase (highest priority)
• InitiateCheckout
• AddToCart
• AddPaymentInfo
• ViewContent
• AddToWishlist
• Subscribe
• Lead

If you have fewer than 8 real events, ranking 4–5 key ones is perfectly fine. But without a configured AEM, a campaign cannot optimise toward purchases — only toward ViewContent, which is nearly impossible to tie to revenue.

Solution 3: UTM Tracking + GA4 as Your Truth Source

Because Meta no longer has full visibility, add UTM parameters to every ad and track conversions independently in GA4. A solid UTM structure:

• utm_source=facebook (or instagram)
• utm_medium=paid
• utm_campaign=campaign_name
• utm_content=creative_name
• utm_term=audience_name

GA4 then shows you real conversions by campaign — often 20–40% higher than what Meta reports. This becomes your truth source for budget decisions.

Solution 4: Marketing Mix Modeling (for Larger Brands)

For brands spending $20k+ per month on ads, MMM is an approach that looks at overall business outcomes (revenue, leads) and uses statistical modelling to determine how much each channel contributed. Tools like Triple Whale, Northbeam, or custom modelling in Excel using historical data all work here.

Metrics to Watch Instead of In-App ROAS

• Blended ROAS (total revenue / total ad spend across all channels)
• Customer Acquisition Cost (CAC) from GA4 attribution
• New customer revenue (new buyers only, not repeat purchases)
• Marginal ROAS (how much an additional $1k in spend actually moves the needle)
• Brand search volume trend (Google Trends, GSC) — an indirect signal

iOS 14 changed Meta Ads permanently. Brands that adapt their measurement gain a clearer picture — and often discover that Meta still performs extremely well; the pixel was just lying to them. Brands that ignore the change are working with fundamentally flawed data and gradually losing ground to competitors who have adapted. (See also how to allocate your budget across Reels, Stories, and Feed instead of letting Meta decide for you.)