IDFA: What It Is, How It Works, and Requesting Permission in iOS

What Is IDFA?

IDFA (Identifier for Advertisers) is a unique alphanumeric identifier assigned to each Apple device for advertising purposes. Unlike IDFV (Identifier for Vendor), IDFA is consistent across all apps on the device and can be reset by the user at any time via Settings — Privacy — Advertising — Reset Identifier.

The primary purpose of IDFA is install attribution and tracking the effectiveness of advertising campaigns on Apple devices. Ad networks use IDFA to link an ad click to an app install, build user profiles, and deliver personalized ads. Since the implementation of ATT, the percentage of users granting access to IDFA averages 20–35% according to various analytics platforms.

According to Flurry Analytics (2025), after iOS 18 the ATT adoption rate reached 92% of active devices, but the share of users agreeing to tracking dropped to 22%. This has led to a fundamental shift in the mobile advertising industry, forcing ad networks to seek alternative attribution methods.

A key feature of IDFA is that it does not contain personal user data and is not linked to an Apple ID. However, Apple classifies it as an identifier subject to App Tracking Transparency regulation.

How IDFA and App Tracking Transparency Work

App Tracking Transparency (ATT) is an Apple framework introduced in iOS 14.5 that requires explicit user consent before reading IDFA. Without consent, the asIdentifierManager method returns a zeroed identifier (00000000-0000-0000-0000-000000000000), and the app cannot track the user.

App Tracking Transparency Framework

ATT is implemented through the ATTrackingManager class with a single public method requestTrackingAuthorization(completionHandler:). When this method is called, the system displays a native dialog with text set by the developer via the NSUserTrackingUsageDescription key in Info.plist. The user can choose one of three options: "Allow Tracking", "Ask App Not to Track", or "Deny".

Authorization status is tracked via ATTrackingManager.trackingAuthorizationStatus, which can take four values: notDetermined, restricted, denied, or authorized. Before requesting ATT, it is recommended to check the current status and not call the dialog if the status is already determined — starting with iOS 17, calling requestTrackingAuthorization again with a denied status does not show the dialog and returns an error.

When to Request Permission

The timing of the ATT request critically affects consent rates. Research from Singular (2025) shows that requesting ATT on the onboarding screen (at first launch) yields a 35–45% conversion rate, while requesting it after completing an action (e.g., after registration) increases conversion to 55–65%. It is recommended to show a pre-permission screen with custom design explaining the value of tracking to the user before the system ATT dialog.

Important: the app can call requestTrackingAuthorization only once before receiving a response. Once the user chooses "Deny" or "Ask App Not to Track", the dialog will no longer appear on the device. The developer must implement fallback mechanisms for working without IDFA to ensure the app continues to function correctly under limited tracking conditions.

iOS 18+ Tracking Requirements

With the release of iOS 18, Apple has tightened requirements for apps using the advertising identifier. All apps that request IDFA or transmit device data for advertising purposes must comply with the new App Store rules. Non-compliance leads to build rejection during App Store Connect review.

Privacy Manifest and Tracking

Starting with iOS 18, every app must include a Privacy Manifest (privacy.xcprivacy file) declaring all APIs used for tracking. For IDFA, the NSPrivacyTracking and NSPrivacyTrackingDomains keys are used to specify the ad network domains to which the identifier is transmitted. Without a correct manifest, the app is rejected during App Store Connect moderation.

Apple has also introduced a requirement to specify the reason for using the advertising identifier in the app description. If the app uses IDFA for install attribution, it must use SKAdNetwork as the primary method, with IDFA only as an additional signal to improve attribution accuracy.

IDFA Alternatives

With increasing restrictions on IDFA, the industry has developed several alternative approaches. SKAdNetwork (StoreKit Ad Network) is Apple's official attribution mechanism that works without IDFA and does not require user consent. SKAdNetwork uses signed postback requests between the ad network and the App Store, ensuring user privacy.

Other alternatives include: probabilistic attribution based on IP address, OS version, and other device signals; private aggregated APIs such as Private Click Measurement (PCM) from WebKit; and proprietary ad platform solutions using machine learning to predict conversions. For the Chinese market, CAID (Cellular Advertising ID) based on mobile operator network data is also used.

Code Examples for Requesting IDFA

Implementing an IDFA request requires integrating App Tracking Transparency and properly handling all possible authorization statuses. Below is a complete example for Swift with a pre-permission screen.

ATT Request in Swift

The example demonstrates the full IDFA workflow: checking status, displaying a pre-permission screen, calling the ATT system dialog, and obtaining the identifier after authorization.

import AppTrackingTransparency
import AdSupport

func requestIDFAAccess() {
    let status = ATTrackingManager.trackingAuthorizationStatus

    switch status {
    case .notDetermined:
        // Showing pre-permission screen with explanation
        showPrePermissionScreen { accepted in
            guard accepted else { return }
            ATTrackingManager.requestTrackingAuthorization { newStatus in
                handleATTStatus(newStatus)
            }
        }
    case .authorized:
        // User already granted consent — getting IDFA
        let idfa = ASIdentifierManager.shared().advertisingIdentifier
        print("IDFA: \(idfa.uuidString)")
    case .denied, .restricted:
        // Redirecting user to settings
        redirectToSettings()
    @unknown default:
        break
    }
}

private func handleATTStatus(status: ATTrackingManager.AuthorizationStatus) {
    if status == .authorized {
        let idfa = ASIdentifierManager.shared().advertisingIdentifier
        // Passing IDFA to ad network
        sendIDFAToAdNetwork(idfa.uuidString)
    }
}

Strategies for Working Without IDFA

Since the majority of users decline the ATT request, developers need a set of strategies for effective work without the identifier. The main approach is a combination of SKAdNetwork and probabilistic attribution methods.

SKAdNetwork and Attribution

SKAdNetwork (SKAN) is an Apple framework for install attribution that does not require IDFA and does not display a consent dialog. The ad network registers its SKAdNetwork ID in the app's Info.plist, and after installation the App Store sends a signed postback with campaign information. SKAN 4.0, introduced in iOS 17, supports multiple conversions, source identifiers, and coarse-grained conversion values.

SKAN limitations: postback delay of 24 to 48 hours, no user identifier, and limited conversion values (up to 64). Despite this, SKAN has become the de facto standard for iOS attribution after ATT implementation, and all major ad networks must support it to work on Apple devices.

Predictive Identifiers

Ad networks such as Meta and Google have developed their own machine learning-based solutions that predict conversions without a user identifier. Meta Aggregated Events Measurement (AEM) uses aggregated data for attribution, while Google Attribution models conversions based on historical data.

For developers, it is important to implement multiple attribution methods simultaneously: SKAdNetwork as the primary method, server-side attribution through ad networks, and proprietary analytics tools with conversion modeling. Combining methods allows recovering up to 70–80% of attribution accuracy compared to the IDFA era. It is recommended to test different combinations of methods during beta testing to select the optimal set for the specific app and target audience.