Free iOS 14 / ATT Attribution Loss Calculator

What changed in April 2021

iOS 14.5 introduced App Tracking Transparency (ATT). Apps that wanted to use the IDFA — the cross-app device identifier ad platforms relied on — had to ask permission via a system-level prompt. Roughly 75 percent of users declined. Without IDFA, Meta, Google, TikTok, and every other platform with a mobile pixel lost the deterministic 1:1 click-to-conversion match they used to have. They replaced it with statistical models, but the models are imprecise and over-report by 20 to 40 percent on most accounts.

The math, top to bottom

ios_conversions       = total_conv * ios_share
opted_out             = ios_conversions * opt_out_rate
pixel_dependent_loss  = opted_out * pixel_dependency
revenue_lost          = pixel_dependent_loss * aov

with CAPI / server-side:
recoverable           = pixel_dependent_loss * 0.40    // 40% recovery
remaining_loss        = pixel_dependent_loss * 0.60

The 40 percent recovery rate is an industry average. In practice, accounts with strong first-party data capture (email at every signup, hashed identifiers, login flows) recover up to 50 percent. Accounts running mostly cold prospecting with minimal first-party identifiers recover closer to 25 to 30 percent.

Why pixel-attribution dependency varies

Not every conversion in your account depends on the iOS-affected pixel chain. If a customer clicks a Google search ad, lands on your site, fills out a form, and you call them back to close, the attribution is mostly first-party (form submit happens on your domain, server-side data captures the lead). Pixel dependency is closer to 30 percent for that flow. Compare to mobile app installs from Meta ads with no email signup, where pixel dependency is 90 percent. The default of 65 percent fits typical paid-acquisition ecommerce.

What CAPI actually does

The Conversions API sends conversion events from your server directly to Meta (or Google, etc.) instead of relying on the browser pixel. Three benefits: (1) bypasses iOS browser restrictions and Safari ITP, (2) captures conversions even when the user disabled cookies or tracking, (3) sends richer data (email hash, phone hash, click ID) that ad platforms can match against their own user data to attribute. Setting up CAPI is a one-time engineering task, usually a few days of work, and recovers 30 to 50 percent of the otherwise-lost attribution data.

What CAPI does not solve

CAPI is not a silver bullet. It still requires user identifiers (email, phone, hashed login) to match conversions to ad clicks. For purely anonymous prospecting traffic, CAPI provides limited improvement. Privacy regulations (GDPR, CCPA, the new state laws) constrain what identifiers you can collect and pass. And SKAdNetwork, Apple's privacy-preserving alternative, only covers app installs, not web conversions, and reports aggregated data 24 to 48 hours later.

The full mitigation stack

• CAPI / server-side events: 30-50% recovery for known users
• SKAdNetwork: app install attribution only, partial coverage
• First-party data capture: every form, every signup, hashed and ready for CAPI matching
• Multi-touch attribution: stitches the journey for users you do see across multiple sessions
• Marketing Mix Modeling: top-down regression to fill gaps no deterministic method can

Layered together, accounts can hit 75 to 85 percent attribution coverage, up from ~50 percent on pixel-only setups. The remaining 15 to 25 percent is genuinely lost to ATT and won't come back.

A worked example

You run an ecommerce site with 2,000 monthly conversions at $120 AOV. Half your conversions come from iOS users (1,000). 75 percent opt out of tracking (750). Your pixel-attribution dependency is 65 percent (488 conversions). At $120 AOV, that is $58,500 of monthly revenue your reports under-attribute. $702K per year. Set up Meta CAPI and recover 40 percent of that ($23,400 monthly, $281K yearly). The remaining $421K per year is the real ceiling on your attribution accuracy unless you add multi-touch attribution and first-party data capture on top.