What is SKAdNetwork (SKAN) in iOS app advertising?

Short answer. SKAdNetwork (SKAN) is Apple's framework for attributing iOS app installs to the ads that drove them while keeping users anonymous. Ad networks sign their ads, and after an install the device sends the network a delayed postback that carries no user- or device-specific data.

What Apple says it is

Apple's developer documentation describes SKAdNetwork as "A class that validates advertisement-driven app installations" and says the API "helps advertisers measure the success of ad campaigns while maintaining user privacy." It involves three participants: ad networks, source apps that display the ads (or websites in Safari), and the advertised app. Apple states that the postback it signs "doesn't include user- or device-specific data", and that apps "don't need to use AppTrackingTransparency before calling SKAdNetwork APIs".

How it works

1. The ad network signs the ad and the source app shows it.
2. The user taps, installs and opens the app.
3. As the user engages, the advertised app updates a conversion value that encodes what they did.
4. The device sends a postback to the ad network, and to the developer if they opt in.

In version 4, Apple says, "The first conversion window spans days 0 to 2; the second window spans days 3 to 7; and the third window spans days 8 to 35." That can produce up to three postbacks. The fine-grained conversion value is "available only in the first postback"; later postbacks, and lower data tiers, carry only a coarse value. Apple assigns each download a postback data tier "to ensure crowd anonymity", so small campaigns receive less detail.

Apple now also points advertisers to a newer framework: "Use AdAttributionKit for app ad campaigns on the App Store and alternative marketplaces." It publishes guidance on how the two interoperate.

A worked example (illustration)

These numbers are invented to show the logic. A game maps its conversion value to three events: finished the tutorial, reached level 5, made a purchase. A campaign gets 1,000 attributed installs. The first postbacks show 400 tutorial completions, 150 level 5s and 20 purchases. The second and third postbacks carry only a coarse value, so they show broadly whether later engagement held up, not exactly what happened. You learn which campaign produced paying players, days later and in aggregate, never which person.

What a good or bad reading tells you

A healthy SKAN setup tells you which campaigns and networks produce valuable users within the 35 days Apple's conversion windows cover. A thin one (most installs with no value, or missing fields) usually means the campaign is too small to clear Apple's privacy thresholds, or the conversion value schema was never designed. Growth consultant Marcus Burke told AppsFlyer: "SKAN really threw a wrench into the whole system."

In games and in consumer brands

SKAN attributes iOS app installs, so it is central for app and game advertisers. A consumer brand that sells on its website measures sales by other means, though it may still run app campaigns if it has an app.

Common mistakes

• Designing the conversion value after launch, so the first weeks of data are empty.
• Treating SKAN numbers as user-level data and trying to join them to individual users.
• Reading a creative as a loser before its postbacks have arrived, since Apple sends them "after a random delay".
• Ignoring that low-volume campaigns get coarser data.

Flixr's view

SKAN pushes creative decisions toward signals you can see earlier, so we judge ads on what networks keep funding and on downstream results read through a mobile measurement partner, not on clicks. Our note on which part of an ad drives installs or purchases covers how we read that.