SKAdNetwork, often shortened to SKAN, is Apples privacy focused framework for measuring mobile app advertising on iOS. It lets advertisers understand which campaigns drive installs and early in app actions, while keeping user identities hidden and protecting device level data.
Instead of sending granular logs of impressions and clicks for every single user, SKAdNetwork sends delayed, aggregated signals. These signals are just detailed enough to help marketers optimize spend, but never precise enough to pinpoint an individual person or device.
SKAdNetwork is part of the StoreKit framework and serves as Apples official attribution system for app install advertising on iOS. When users do not consent to tracking through App Tracking Transparency, SKAN becomes the main way to measure performance.
Key ideas behind SKAdNetwork
With SKAdNetwork, you still see which campaigns and sources generate installs and revenue signals, but you cannot see which exact user did what.
Four main actors participate in a SKAN flow
These roles stay the same across SKAN versions, even as features evolve.
At a high level, a SKAN flow looks like this
Postbacks include information such as
No user identifiers are included. Timing is also randomized within a window so that events cannot be matched to a specific user session.
Since SKAN does not send event logs, conversion values become the central way to encode post install performance.
In SKAN versions before four
In SKAN four, Apple introduced an additional concept
Advertisers design their own conversion schemas within these constraints. Good schemas capture the most predictive signals early in the user journey, so that delayed reporting still reflects long term value.
SKAdNetwork four significantly expanded what advertisers can see, while keeping the privacy model intact. Some of the most important updates
Multiple measurement windows and postbacks
Instead of a single postback, SKAN four can send up to three, each tied to a different period after install
This lets advertisers see how users behave beyond the first day without exposing individual timelines.
Crowd anonymity
Apple replaced the old privacy threshold concept with crowd anonymity.
In short, the larger the crowd, the more detailed the aggregated data, because it becomes impossible to reverse engineer a single user.
Source identifier expansion
The previous campaign identifier had up to one hundred options. SKAN four expands this to ten thousand possibilities via the source identifier field.
Advertisers are encouraged to treat this as three levels inside one number, for example
This hierarchical approach fits neatly with crowd anonymity and helps preserve meaningful breakdowns even when privacy constraints limit detail.
LockWindow lets an app developer end the measurement period early if enough data has already been collected. This can speed up the arrival of postbacks and give advertisers insights sooner.
At recent developer events, Apple introduced AdAttributionKit as the next generation framework that builds on SKAN concepts.
Key points
For now, SKAdNetwork remains the foundation of privacy centric attribution on iOS, while AdAttributionKit is positioned as its successor. Both frameworks share the same underlying philosophy: strong privacy guarantees coupled with limited but useful measurement.
Stronger privacy
No user level identifiers are sent. Postbacks are delayed and aggregated so a single user cannot be singled out.
Fraud resistance
Because Apple signs each postback and controls attribution logic on device, it is harder for bad actors to spoof installs at scale.
Compliance ready
SKAN is designed to align with modern privacy expectations and regulation, which helps advertisers continue to run performance marketing on iOS.
Limited visibility into lifetime value
SKAN windows focus on early behavior. Later conversions may not be fully captured, especially for products with long payback periods.
Reduced granularity
You do not see user journeys, detailed source level paths, or flexible event streams. Everything must fit inside the conversion value logic.
Delayed optimization
Randomized postback delays mean that you cannot instantly react to campaign performance. You must work with aggregated, lagged data.
Complex implementation
Designing a good conversion scheme, integrating SKAN correctly, and aggregating data from multiple networks requires careful planning and often the help of a measurement partner.
To get real value from SKAN, teams usually focus on four areas
Many advertisers rely on mobile measurement partners or internal data teams to manage these tasks and align SKAN results with broader analytics.
What is SKAdNetwork in simple terms
It is Apples system that tells you which ad campaign drove an app install, but without revealing who the user is.
Why did Apple create SKAdNetwork
To balance performance measurement with privacy. It lets advertisers keep investing in iOS advertising while avoiding user level tracking.
What is the difference between SKAN and device based attribution
In device based attribution, networks and partners use identifiers to track individuals across apps. In SKAN, Apple performs attribution on device and only shares anonymous, delayed results.
What is a conversion value in SKAdNetwork
It is a small number chosen by the app developer that represents what a user did after install. For example, reaching a level, making a purchase, or achieving an engagement score.
How many campaigns can I run under SKAdNetwork four
You can use up to ten thousand source identifier values, which is far more flexible than the older one hundred campaign limit.
Can SKAdNetwork measure re engagement
Traditional SKAN does not handle re engagement directly. AdAttributionKit is designed to support that use case more fully.
Is SKAdNetwork going away
Apple has signaled that AdAttributionKit will gradually take over as the primary framework, but SKAN concepts, and likely SKAN itself, will remain important for some time.