Replace AppsFlyer: faster cold start

Q: How do I measure cold start before / after?

iOS: Xcode Organizer's Launch Time metric (aggregate from real users) or Instruments → App Launch template (synthetic). Android: `adb shell am start -W /. ` or Play Console's Vitals → Startup time. Measure before removing the old SDK, after, and compare on the same device class.

Q: Does Respectlytics block the main thread on init?

No. `Respectlytics.configure(appKey:)` is synchronous but does only in-memory work (allocates the ring buffer). The network flush runs on a background dispatch queue / coroutine.

▸Example AppsFlyer call (the "before")

swift Respectlytics

import AppsFlyerLib

// ATT prompt path — required before AppsFlyer reads IDFA:
ATTrackingManager.requestTrackingAuthorization { _ in
    AppsFlyerLib.shared().start()
    AppsFlyerLib.shared().logEvent("af_purchase", withValues: [
        AFEventParamRevenue: price,
        AFEventParamCurrency: "USD",
        AFEventParamContentId: sku,
    ])
}

Heavy analytics SDKs do work at app launch — reading identifiers, initialising queues, network dispatch — that compounds visibly on lower-end devices. Respectlytics's SDK adds typically under 30ms to cold start, vs 100-300ms for Firebase Analytics's full initialisation chain.

☑Remove AppsFlyer cleanly

1
Remove the AppsFlyer SDK from your build (AppsFlyerFramework / af-android-sdk / react-native-appsflyer / appsflyer_sdk)
2
Remove AppsFlyerLib.shared().start() and logEvent(...) call sites
3
Re-check your Info.plist for NSUserTrackingUsageDescription — if no other SDK needs ATT, remove it (Apple flags apps that ship the key without code that calls ATTrackingManager)
4
Re-check your Android merged manifest for com.google.android.gms.permission.AD_ID and remove the corresponding <uses-permission> if no other SDK contributes it
5
Plan how you'll attribute installs without AppsFlyer — Apple SKAdNetwork + Google Play Install Referrer (both first-party, no SDK needed) cover most cases

⇋AppsFlyer vs Respectlytics — faster cold start

	AppsFlyer	Respectlytics
Typical cold-start contribution (p50)	— see tool note above	< 30ms
Initialisation work on launch	Reads IDFA/AAID, opens SQLite, spins up threads	Allocates ring buffer (RAM-only)
Number of background threads spawned	— typically 2-4	1
Synchronous I/O on init	— typical (SQLite open)	None

❓Frequently asked questions

How do I measure cold start before / after?

iOS: Xcode Organizer's Launch Time metric (aggregate from real users) or Instruments → App Launch template (synthetic). Android: adb shell am start -W <package>/.<activity> or Play Console's Vitals → Startup time. Measure before removing the old SDK, after, and compare on the same device class.

Does Respectlytics block the main thread on init?

No. Respectlytics.configure(appKey:) is synchronous but does only in-memory work (allocates the ring buffer). The network flush runs on a background dispatch queue / coroutine.

What's typical magnitude of improvement?

On a mid-range Android device, removing Firebase Analytics + AppsFlyer typically saves 100-300ms off cold start, depending on Google Play Services init state. On iOS the delta is usually 50-150ms. Effect is more pronounced on cold-start (uncached) than warm-start launches.

Does cold-start improvement actually affect business metrics?

Yes — first-session abandonment correlates with launch latency in published benchmarks. A 100ms improvement on the slowest deciles of your device distribution can show a measurable first-day retention lift.

⇢Related migration guides

Replace AppsFlyer to speed up cold start