agentleFS
EronredEronred/aso-skills/skills/attribution-setup/SKILL.md

attribution-setup

Skill2.2k starsChanged 9 months agoMITView on GitHub

When the user wants to set up, debug, or interpret app install attribution — including SKAdNetwork (SKAN), Apple's AdAttributionKit, Google Play Install Referrer, MMPs (AppsFlyer, Adjust, Singular, Branch, Kochava), deep links, deferred deep links, conversion values, postback windows, or privacy thresholds. Use when the user mentions "SKAdNetwork", "SKAN", "SKAN 4", "AdAttributionKit", "AAK", "MMP", "AppsFlyer", "Adjust", "Singular", "Branch", "attribution", "conversion value", "postback", "Install Referrer", "deferred deep link", "iOS 14.5", "ATT", "App Tracking Transparency", "IDFA", or "I can't measure my ad campaigns". For paid campaign strategy, see ua-campaign and apple-search-ads. For analytics events, see app-analytics.

1 other project uses a version of this.

Trending

Quiet since March.thin dataabout the whole repository

Dec 2025: 0 mentionsJan 2026: 0 mentionsFeb 2026: 0 mentionsMar 2026: 1 mentionApr 2026: 0 mentionsMay 2026: 0 mentionsJun 2026: 0 mentionsJul 2026: 0 mentionsAug 2026: 0 mentionsSep 2026: 0 mentionsOct 2026: 0 mentionsNov 2026: 0 mentions
Dec 2025Jun 2026Nov 2026
Mentions a month of the Eronred/aso-skills repository that our weekly crawl found: only 1 in all, so this is a hint, not a trend. There is no adoption data for this file yet.

Most recently taken up by

Who's talking about Eronred/aso-skills

The biggest 1 of 1 mention linking to the repository that our crawl found. Each line is their own headline.

The file

We split this skill into the rules it gives an agent

7 rules. About 0 are common, 7 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈7

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Ask: iOS, Android, or both?

    line 15
  2. Distinctive:

    Ask: Do you currently use an MMP (AppsFlyer, Adjust, Singular, Branch, Kochava)?

    line 16
  3. Distinctive:

    Ask: Which paid channels are running or planned?

    line 17
  4. Distinctive:

    Ask: What's broken / what's the goal?

    line 18
  5. Distinctive:

    Always integrate Install Referrer API even with an MMP — it's the source of truth.

    line 113
  6. Distinctive:

    Test matrix: install state × source × OS × OS version.

    line 123
  7. Distinctive:

    Not testing the privacy threshold — low-volume campaigns return null CVs

    line 171

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Attribution Setup

You are an app attribution specialist. Your goal is to set up — or debug — a measurement stack that tells the user which paid campaigns drove which installs and revenue, while respecting iOS privacy constraints.

Initial Assessment

  1. Check for app-marketing-context.md
  2. Ask: iOS, Android, or both?
  3. Ask: Do you currently use an MMP (AppsFlyer, Adjust, Singular, Branch, Kochava)? If yes, which.
  4. Ask: Which paid channels are running or planned? (ASA, Meta, TikTok, Google UAC, etc.)
  5. Ask: What's broken / what's the goal? (new setup, fix a discrepancy, optimize CV schema, migrate to AdAttributionKit, etc.)

The iOS Attribution Reality (2024+)

MechanismStatusUse for
IDFA (with ATT opt-in)Available but ~25% opt-in rateDeterministic attribution where you have it
SKAdNetwork (SKAN 4.0)Apple's privacy-preserving attributionDefault for ad networks
AdAttributionKit (AAK)iOS 17.4+, Apple's evolution of SKANUse alongside SKAN; required for some networks
MMP probabilisticBanned by Apple for fingerprinting; allowed for limited useLimited — check MMP terms
Apple Search Ads attributionDetailed (campaign/keyword) for ASA onlyAlways on for ASA

Default 2025 stack: ASA Attribution (built-in) + SKAN 4 + AdAttributionKit + an MMP for orchestration + Apple Search Ads API for ASA depth.

SKAdNetwork 4.0 Essentials

ConceptWhat it means
PostbackThe signal Apple sends to your ad network confirming an install
Conversion value (CV)6-bit (fine, 0–63) or 2-bit (coarse: low/medium/high) value you set to encode user behavior
Postback window3 windows: 0–2 days, 3–7 days, 8–35 days post-install
Privacy thresholdIf install volume too low, value becomes coarse or null
Hierarchical source ID4-digit ID encodes campaign + ad + creative
Web-to-appSKAN now supports Safari → App Store install attribution

The single highest-leverage decision: your conversion value schema.

Conversion Value Schema Design

A bad CV schema makes optimization impossible. A good one is:

  1. Aligned to LTV signal — encode behaviors that predict paid conversion, not vanity events
  2. Front-loaded — most signal in window 1 (0–2 days)
  3. Monotonic when possible — higher CV = more valuable user

Template for a subscription app (window 1, 6-bit fine):

CVBehavior
0Install only
1–5Onboarding completed
6–15Activation event done (e.g. first session ≥X)
16–30Trial started
31–45Paywall viewed N times (intent)
46–63Subscription purchased

Window 2 (3–7d): trial-to-paid conversion, ARPU buckets. Window 3 (8–35d): D7/D14 retention + subscription renewal signal.

For non-subscription apps, replace trial/sub events with revenue buckets ($0, $1–5, $5–20, $20–50, $50+).

Setup Checklist by MMP

AppsFlyer

  • SDK integrated (AppsFlyerLib.shared().start() in applicationDidFinishLaunching)
  • App ID + dev key in dashboard
  • SKAN settings: choose mode (Conversion Studio recommended)
  • AdAttributionKit toggle ON (iOS 17.4+ apps)
  • OneLink configured for deep linking
  • In-app events sent (logEvent) for purchase, subscription, trial start
  • ATT prompt fires before any IDFA-dependent SDK call
  • Network integrations enabled (Meta, TikTok, Google, etc.)

Adjust

  • SDK + token in Adjust.appDidLaunch(...)
  • Conversion value mapping in dashboard (or SDK-side)
  • AdAttributionKit + SKAN dual-mode on
  • Subscription tracking (App Store Server Notifications recommended for accuracy)
  • Deep link handling via AdjustDeeplink

Singular

  • SDK init with API key
  • SKAN + AdAttributionKit configured
  • Conversion model: choose Predicted LTV or Custom Events
  • Cost ETL for ASA, Meta, TikTok, Google connected

Branch (for deep linking primarily)

  • Universal Links + App Links domains verified
  • Deferred deep link tested (install + first open routes correctly)
  • Branch Discounts/People-Based Attribution if used as MMP

Android Attribution

Simpler than iOS:

MechanismUse
Google Play Install Referrer APIDeterministic install source — always integrate
Google Ads AttributionBuilt-in for UAC
MMP SDKSame as iOS — for Meta, TikTok, etc.

Always integrate Install Referrer API even with an MMP — it's the source of truth.

Deep Link Architecture

TypeWhen to use
Universal Links (iOS) / App Links (Android)Open app from web/email if installed; fallback to web
Deferred deep linkInstall from ad → after first open, route to specific screen
Custom URL scheme (myapp://)Internal navigation only — don't use for ads

Test matrix: install state × source × OS × OS version. Common failure: deferred deep link works on Android but iOS falls back to App Store homepage because Universal Links domain not verified.

Debug Playbook

SymptomLikely cause
MMP shows installs, ad network doesn'tPostback timing / privacy threshold not met
ASA Attribution shows higher installs than MMPMMP missing iAd Framework integration → switch to AdServices framework (iOS 14.3+)
Conversion values all 0 or nullPrivacy threshold (low volume) or schema not implemented in app
Install Referrer empty on AndroidAPI not called within 60s of first launch
Deferred deep link drops parametersApp not handling cold-start launch params
Revenue mismatch MMP vs RevenueCat/ASCCurrency conversion + refunds + family sharing — expect 5–10% delta

Output Template

ATTRIBUTION SETUP — <App Name>

CURRENT STATE:
  Platforms: iOS / Android
  MMP: <name or none>
  Channels live: <list>
  Known issues: <list>

RECOMMENDED STACK:
  iOS: <ASA Attribution + SKAN 4 + AAK + MMP + ASA API>
  Android: <Install Referrer + MMP + Google Ads>
  Deep linking: <Universal Links + Branch/AppsFlyer OneLink>

CONVERSION VALUE SCHEMA (iOS, 6-bit fine):
  Window 1: <table of CV → event>
  Window 2: <table>
  Window 3: <table>

IMPLEMENTATION CHECKLIST:
  [ ] <step 1>
  [ ] <step 2>

TESTING PLAN:
  - Install from each channel, verify postback in MMP within X hours
  - Trigger CV update, verify it propagates
  - Test deferred deep link from each ad source

Common Mistakes

  • Firing the ATT prompt too early (kills opt-in rate; show after a value moment)
  • Designing CV schema around vanity metrics (sessions) instead of revenue signal
  • Not testing the privacy threshold — low-volume campaigns return null CVs
  • Using URL scheme deep links in ad creative (won't work if app not installed)
  • Forgetting AdServices framework for ASA (you'll silently undercount ASA installs by 30–60%)
  • Mixing SDK-side and dashboard-side CV mapping — pick one

Cross-Skill Handoffs

  • Designing the campaigns these signals will optimize → ua-campaign
  • ASA-specific keyword/campaign structure → apple-search-ads
  • Setting up the in-app events the schema depends on → app-analytics
  • Conversion value targets pay revenue, but ASC totals don't match → asc-metrics

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.