Summer’25 finally delivered practical tools to run identifier-free measurement across both platforms. Apple’s AdAttributionKit (AAK) in iOS 18.4 added overlapping re-engagement conversions, conversion tags, configurable attribution windows/cooldowns, country codes in postbacks, and an on-device development postbacks tester. On Android, the Privacy Sandbox tightened the Attribution Reporting API and SDK Runtime while major ad SDKs documented Sandbox paths. Here’s the concise, SEO-friendly field guide for UA teams. 

iOS: What changed in AdAttributionKit (iOS 18.4)

• Overlapping re-engagement windows: you can track multiple win-backs at once instead of locking to a single window. Conversion tags let you update the correct postback for a specific campaign or offer. This removes a long-standing SKAN-era limitation for CRM-style or catalog apps. 
• Configurable rules: tailor attribution windows and cooldowns by ad type to reduce cannibalization between UA and retargeting.
• Country-level postbacks: AAK can now include country code (subject to crowd-anonymity thresholds), enabling geo-level optimization without burning conversion bits. 
• Testing in Settings: the development postbacks tool in iOS Settings lets you generate install or re-engagement postbacks (with chosen granularity and country code) no live ad flow required. 

Why it matters: You can run parallel re-engagement cohorts, align windows to real payback (e.g., 24h utilities vs. delayed-trial subs), and tune bids by geo with cleaner data. Faster on-device testing shortens the path from instrumentation to scaled spend. 

Android: Privacy Sandbox, Attribution Reporting & SDK Runtime

• Attribution Reporting for mobile: the API shifts measurement from GAID-style user identifiers to event/aggregate reports; July docs clarified mobile use cases and data paths. Plan for aggregate conversion reporting and model-based lift. 
• SDK Runtime (Android 14+): ad SDKs run in an isolated runtime with scoped permissions, improving user privacy and stability; June updates detail the architecture you’ll implement with mediation partners.
  
• Ecosystem status: regulators continue to oversee the rollout; the UK CMA published Google’s Q2 2025 progress report in late July. Meanwhile, Google’s Mobile Ads docs (updated Sept 3, 2025) describe current Sandbox integrations—useful for readiness checks with networks. 

Why it matters: Expect fewer user-level signals and more aggregate attribution. UA planning should blend incrementality tests, MMM sanity checks, and Sandbox-compatible event taxonomies. Vendor due-diligence now includes SDK Runtime support status. 

A 30-Day, cross-platform adoption plan

1. Define value signals
   
   • iOS (AAK): map conversion tags to distinct re-engagement offers; set windows/cooldowns by lifecycle stage. 
   • Android (Sandbox): list source/trigger events for Attribution Reporting and the aggregates you’ll optimize to (e.g., trial start, level-3 reached). 
2. Harden the build chain
   
   • Update ad SDKs; verify SDK Runtime compatibility and permissions with each partner/adapter.
     
3. Run parallel tests
   
   • iOS: A/B 7- vs 14-day re-engagement windows; compare D7 revenue per tag and per country. 
   • Android: compare legacy GAID reporting to Attribution Reporting aggregates across 2 geos; track stability of cost per aggregated action.
     
4. Rebuild dashboards
   
   • Put country-level postbacks (iOS) and aggregate reports (Android) in one view for apples-to-apples budget rebalancing by geo/OS. 

Bottom line 

Identifier-free measurement is now an operating model, not a science project. With AdAttributionKit iOS 18.4 and Android Privacy Sandbox you can run scalable UA and re-engagement without device-level IDs provided your signals, windows, and SDKs are updated for Attribution Reporting, SDK Runtime, conversion tags, and country-level postbacks. Teams that adopt these flows now will enter Q4 with faster tests, cleaner geo optimization, and fewer data-quality surprises.