ads-apple

ads-apple provides a structured, deep-analysis framework for advertisers running Apple Ads (formerly Apple Search Ads, ASA) to promote mobile apps on iOS. It solves the problem of evaluating an Apple Ads account holistically rather than metric-by-metric: it takes account exports or pasted metrics, identifies which placement types are active, runs each applicable check as PASS, WARNING, or FAIL, calculates an overall ASA Health Score from 0 to 100, and produces a findings report with an action plan. The skill reflects Apple's 2025 rebrand and recent platform changes such as Creative Sets deprecation and the Maximize Conversions bidder.

The analysis is organized into weighted areas. Campaign Structure (25%) checks bottom-, middle-, and top-of-funnel design, including brand, competitor, and category campaigns, plus keeping Search Match ad groups isolated from Exact Match. Bid Health (20%) reviews CPT (Cost Per Tap), TTR (Tap-Through Rate) benchmarks, conversion rates, and the Maximize Conversions auto-bidder (GA February 26, 2026) with its Target CPA, recommended daily budget of at least 5x target CPA, and two-week learning period. Custom Product Pages (15%) covers CPP variant testing (up to 70 per app as of October 2025) and asset alignment to keyword themes. Attribution & MMP Health (15%) covers AdAttributionKit, SKAN/AAK dual attribution, MMP integration (AppsFlyer, Adjust, Branch, Singular) and ATT opt-in. Budget Pacing (10%) and TAP Coverage across the four placement types (Search Results, Search Tab, Today Tab, Product Pages) round out the checks.

The skill targets mobile app advertisers and ASA/UA managers who need to audit iOS ad accounts, diagnose underperformance, or plan optimizations. It is triggered by terms like Apple Ads, Apple Search Ads, ASA, App Store ads, AdAttributionKit, and view-through attribution. A stated limitation is that Maximize Conversions currently optimizes only for installs, not post-install events such as trials, subscriptions, or ROAS. The skill is marked user-invokable: false and expects account data to be supplied for analysis.