​​SKAdNetwork (SKAN) is Apple's attribution framework for iOS 14 that aims to provide a privacy-friendly way for app developers to measure the effectiveness of their ad campaigns. SKAN 4.0 is the latest version of the framework 一 released on October 24, 2022, for versions equal to or later than iOS 16.1 and iPadOS 16.1, and it introduces several new features and improvements.

With each update, SKAN brings not only a lot of options but also great complexity to mobile attribution. In this post, we will share the business-side or marketing-side implications of the SKAN 4.0 changes you need to know to stay ahead of the game.

Four marketing implications of SKAdNetwork 4.0:
- Small-scale testing and operations get easier
- Creative-level or country-level attribution can partly come back
- Cohort analysis for ARPU and ROAS can partly come back
- Web placements now trackable with SKAN

Small-scale testing and operations get easier

One of the major improvements in SKAN 4.0 is that it makes small-scale testing and operations easier for app developers. Under SKAN 3.0, the privacy threshold made it difficult to run small-scale ad campaigns. This was because the framework required a minimum number of installs per campaign before it would provide any SKAN postback for in-app events, making it difficult to test and optimize campaigns with a limited budget.

In order to protect user privacy and ensure that conversions are recorded only for a statistically significant number of users, SKAN has a privacy threshold that limits the number of conversions that can be recorded for a given app. Determined by the number of installs per campaign ID that have occurred within the past 24 hours, the threshold is typically known to be within a range of 10-20 installs. If a campaign falls below this, the conversion value field in the postback associated with that campaign will be assigned a null value.

SKAN 4.0 introduces the concept of "crowd anonymity," which is achieved through a four-tiered system ranging from tier 0 to tier 3. Depending on the tier that a source is assigned to, advertisers may receive either fine-grained or coarse-grained values to gather information about users' post-install activities. The first is a 6-bit value allowing for up to 64 combinations while the latter is an approximate value that can be represented by low, medium, or high.

If a campaign/source is assigned to tier 0, no information will be provided, which is similar to the situation when the privacy threshold was not met in SKAdNetwork 3.0. However, if a campaign/source is assigned to tier 1, advertisers may still receive coarse-grained values, which is less specific than fine-grained values but still provides some insights into user behavior. Apple states that advertisers can "receive less specific attribution information when the privacy thresholds for conversion values are not met," which suggests that getting into tier 1 may be less difficult than meeting the privacy threshold under SKAN 3.0.

Creative-level or country-level attribution can partly come back

‍

‍

In earlier versions of SKAN, only campaign-level analysis was possible. SKAN 4.0 partially restores this functionality with the introduction of a new parameter called “source-identifier,” which replaces the previous “campaign-id.” The source-identifier can be up to four digits in length (0~9,999), up from the previous two digits (0~99), and can be used to represent additional dimensions such as creative, country, and format.

Cohort analysis for ARPU and ROAS can partly come back

‍

‍

SKAN 4.0 partially restores the ability to perform cohort analysis such as day-X average revenue per user (ARPU) and return on ad spend (ROAS) through the use of three postbacks, an increase from the previous single postback mechanism. The conversion window has also been extended to a maximum of 35 + 6 (24 ~ 144 of random delay) days, providing deeper insights into post-install user activities. These postbacks are sent consecutively within a designated time window from the date of app installation and can be used to reconstruct the user's journey after installation.

Mobile measurement partners (MMPs) can use these postbacks, attribute them to a specific date of installation, summarize all user journeys and present them in a cohort analysis format. However, it's worth noting that SKAN 4.0 applies a random delay to each postback, making it difficult to accurately attribute it back to the specific date of installation. 

Web placements now trackable with SKAdNetwork

One of the new features in SKAN 4.0 is the ability to track web ads in Safari browser, in addition to mobile placements. This allows ad networks, MMPs, and advertisers to see how app installs and in-app events are attributed back to the ad placements on the web. This can provide valuable insights into the effectiveness of cross-channel ad campaigns. Please note that this feature will be only supported for the Safari browser.

Looking ahead

Overall, SKAN 4.0 introduces several new features and improvements that make it easier for app developers to measure the effectiveness of their ad campaigns in a privacy-friendly way.

Airbridge is currently working closely with ad networks to support SKAN 4.0, and we are applying changes to our existing conversion value settings page to reflect the SKAN 4.0 specifications more accurately. We want to ensure that our clients are able to take full advantage of the benefits that SKAN 4.0 has to offer. Tune in for more updates!