AppsFlyer: Essence, Attribution, and Marketing Analytics Tools

What is AppsFlyer?

AppsFlyer is a cloud-based mobile attribution and marketing analytics platform founded in 2011. The system allows app developers to determine which advertising channels drive installs, registrations, and purchases. AppsFlyer processes over 12 billion installs annually.

The platform supports more than 8,000 ad networks and MMP partners. Data is collected via mobile SDKs, Server-to-Server API, and Apple’s SKAdNetwork integration. AppsFlyer is certified under ISO 27001 and SOC 2 Type II standards for data protection.

AppsFlyer’s key value is a unified dashboard for all advertising campaigns. Marketers can see cost per install (CPI), average revenue per user (ARPU), and return on ad spend (ROAS) for each channel in a single interface without manually collecting data from different sources.

Attribution Models in AppsFlyer

AppsFlyer supports several attribution models for installs and events. The primary model is Last-Click Attribution: the install is attributed to the last ad click before navigating to the App Store or Google Play. If a user saw an ad on Facebook, then on Google Ads, and installed the app — the install will be credited to Google Ads.

Last-Click and First-Click Attribution

The First-Click model attributes the install to the first click in the touchpoint chain. This model is useful for evaluating the effectiveness of top-of-funnel channels. AppsFlyer allows flexible configuration of the attribution window: from 1 day to 90 days from the time of click.

View-Through Attribution

View-Through (or impression-based attribution) accounts not only for clicks but also for ad views. If a user saw an ad banner, did not click on it, but installed the app within a specified window (typically 24 hours) — AppsFlyer can attribute the install to that impression. The View-Through window is configured separately from Click-Through.

Cohort LTV Analysis

Cohort LTV is an AppsFlyer report that shows revenue per user by day after install. For example, a cohort of users who installed the app on March 1st may show an ARPU of $2.50 on day 7 and $4.80 on day 30. This allows comparing traffic quality from different sources.

SKAdNetwork and Attribution without IDFA

SKAdNetwork (SKAN) is Apple’s framework for install attribution without using IDFA, launched in 2021. AppsFlyer was among the first MMPs to fully support SKAdNetwork. The transition to SKAN was driven by the iOS 14.5 privacy policy change, which requires explicit user consent for IDFA access.

SKAdNetwork operates on a postback principle: the ad network sends an encrypted click notification to Apple, and Apple sends back aggregated postback messages confirming the install. AppsFlyer receives these postbacks and decodes them for display in reports.

SKAdNetwork limitations include a postback delay of up to 48 hours, the absence of user-level data (aggregated only), and a maximum limit of 100 campaigns per app. AppsFlyer provides tools for optimizing work under these constraints, including SKAdNetwork Copilot — a recommendation system for selecting optimal campaign settings.

AppsFlyer Protect360 Fraud Protection

AppsFlyer Protect360 is a built-in mobile ad fraud protection system. The system uses machine learning to detect invalid installs: click flooding, install fraud, click spam, and device fraud. Protect360 checks each install in real time and assigns a suspicion rating.

According to AppsFlyer, Protect360 blocks up to 30% of fraudulent installs before they appear in reports. The system analyzes over 100 signals: click frequency from a single IP, time between click and install, anomalies in OS versions and device types. Marketers see a separate Fraud Score panel for each channel in the dashboard. Fraud detection operates in real time: each install is verified in milliseconds before being recorded in reports.

Types of fraud detected by Protect360 include: Install Fraud (inflating installs using emulators and device farms), Click Fraud (false clicks without actual store navigation), Click Spam (automated clicks with no install intent), and SDK Spoofing (falsifying SDK data). Each fraud type has unique patterns that the system identifies through supervised ML models trained on historical data from billions of installs.

Key Performance Indicators (KPIs) adjusted for fraud are displayed in a filtered view. AppsFlyer automatically recalculates CPI, ROAS, and LTV after excluding fraudulent installs, providing a true picture of ad campaign effectiveness. The Fraud Score for each traffic source is displayed in the dashboard with color coding: green (low risk, 0–5% fraud), yellow (medium risk, 5–15%), and red (high risk, over 15%).

Machine learning in Protect360 is trained on data from millions of installs and is continuously updated. Models are retrained every 24 hours to account for new fraud patterns. This allows the system to adapt to new types of fraud without manual rule updates.

Integrating AppsFlyer SDK into a Mobile App

AppsFlyer provides native SDKs for iOS, Android, React Native, Flutter, and Unity. Basic integration includes SDK installation, initialization with a dev key, and configuring delegates for link handling. The SDK automatically collects data on installs, sessions, and purchases.

Initialization on Android (Kotlin)

import com.appsflyer.AppsFlyerLib

class MainApplication : Application() {
    override fun onCreate() {
        super.onCreate()
        AppsFlyerLib.getInstance()
            .init("YOUR_DEV_KEY", null, this)
        AppsFlyerLib.getInstance()
            .start(this)
    }
}

Initialization on iOS (Swift)

import AppsFlyerLib

class AppDelegate: UIResponder, UIApplicationDelegate {
    func application(
        _ application: UIApplication,
        didFinishLaunchingWithOptions launchOptions: [UIApplication.LaunchOptionsKey: Any]?
    ) -> Bool {
        AppsFlyerLib.shared()
            .appsFlyerDevKey = "YOUR_DEV_KEY"
        AppsFlyerLib.shared()
            .appleAppID = "123456789"
        AppsFlyerLib.shared()
            .start()
        return true
    }
}

Integration via Flutter

import 'package:appsflyer_sdk/appsflyer_sdk.dart';

final appsFlyerSdk = AppsFlyerSdk(
    AppsFlyerOptions(
        devKey: "YOUR_DEV_KEY",
        isDebug: true,
        appId: "123456789"
    )
);

void main() {
    appsFlyerSdk.initSdk(
        registerConversionDataCallback: true,
    );
}

Marketing Analytics and Report Dashboards

AppsFlyer provides several built-in dashboards: Overview, Cohort LTV (revenue by cohort), Raw Data Report (raw data for export), and Aggregate (market benchmarks). Overview shows key metrics: installs, organic traffic, non-organic traffic, revenue, ROAS, and cost per install. All dashboards are available via web interface and through the REST API for integration with internal reporting systems.

Raw Data Report is a detailed data export at the individual install and event level. Data is available in CSV and JSON formats, as well as through the S3-exporter for automatic loading into your own data warehouse. Raw Data Report is essential for building custom dashboards in Tableau, Power BI, or Looker.

Aggregate is a comparative analytics tool that shows how your app’s metrics relate to market benchmarks. AppsFlyer aggregates data from over 12 billion installs and provides benchmarks by app category, region, and ad platform. Developers can see what CPI is considered normal for their category and how Retention Rate varies by geography.

AppsFlyer Pricing and Attribution Cost

AppsFlyer operates on a paid subscription model, with pricing based on monthly install volume. For small projects, a Starter plan is available with a limit of up to 5,000 installs per month. The basic plan (Pro) starts at around $2,000–3,000 per year and includes the full set of attribution features, integration with all ad networks, and Protect360 fraud protection.

At the Enterprise tier, additional features include: a dedicated platform instance, custom SLAs, a personal account manager, team training, and a security audit. Enterprise pricing is calculated individually and typically starts at $20,000 per year for apps with medium traffic up to 500,000 installs per month.

Additional costs include: integration with premium ad networks, enhanced Raw Data Report with hourly S3 export, and multi-platform accounts for agencies and large advertisers. AppsFlyer also offers a free 30-day trial period to explore the platform without entering payment information.