Inside Airbridge
A comprehensive guide to SKAN install double-counting prevention
August 17, 2023
By
Jeongwon Kim

If you’re a marketer exploring iOS app marketing, you’ve likely encountered the difficulty of accurate performance measurement. This challenge arises from the changes introduced by Apple in April 2021 with the iOS 14.5 update, particularly the implementation of the App Tracking Transparency (ATT) policy. Since this pivotal shift, collecting Advertising Identifiers (IDFA) has become a more intricate task, introducing new complexities along the way. 

Also, numbers show the challenges marketers face. According to Airbridge’s analysis, as of 2023, only an average of 30% of iOS users globally have granted permission for IDFA collection. This makes it challenging to accurately measure ad performance for 7 out of 10 individuals. 

Airbridge offers the Attribution Overlap feature to identify any overlap in SKAN attributions. Check out this article if you are a marketer interested in seeking more accurate ad performance insights in SKAN.

💡You can also use Attribution Overlap in SKAN 3.0.
The Attribution Overlap feature is available when CSM grants access to you to change the SKAN conversion value settings page to a version that supports up to 4.0. You can see more accurate ad performance with the ability to determine double attributions on channels that only support SKAN 3.0.

Why you need Attribution Overlap

If you utilize the Attribution Overlap feature, you can easily determine whether an event has been attributed via both SKAN and the Airbridge attribution model. The SKAN and Airbridge attribution models vary in how they measure ad performance. The two attribution models differ in terms of who attributed and whether IDFA is used.

Therefore, if you measure ad performance using both attribution models concurrently, you may encounter events that meet the attribution criteria for both. The event is counted as an overlap in the ad performance measured by each attribution model. In fact, the event only happened once, but the report counts as if it happened twice.

Eliminating double attributions allows for a more accurate aggregation of ad performance measured by both attribution models.

Events attributed with SKAN are treated as attribution overlaps when the user permits  IDFA collection. The double-attributed event signifies ad performance measurable within the Airbridge attribution model independently of SKAN. Conversely, solely attributed events indicate ad performances exclusively measurable through SKAN. The count of such events is calculated by subtracting the number of attributed events in the Airbridge attribution model from the total number of events attributed to SKANs.

Here’s how to use ‘Attribution Overlap’

CPI(Cost Per Install) or CPA(Cost Per Action) is a key metric for marketers. Attribution Overlap helps you calibrate your CPI and CPA by incorporating non-overlapping SKAN events into the existing app marketing performance, providing a holistic view of attribution results. Let's delve into an example illustrating how to assess actual ad performance by categorizing events based on the attribution model and the presence of overlapping attributions.

 

First, let’s examine a sample CPI calculated for app installs that are attributed using SKAN and Airbridge’s attribution models without accounting for double attributions. The ad spend is $1000.

All 500 app installs attributed to SKAN came from iOS users, including users who allowed IDFA collection and those who did not. CPI of $2 is the ad performance calculated based on the number of app installs by iOS users. App installs by Android users are excluded.

The 800 app installs which are attributed by the Airbridge attribution model are the sum of app installs from iOS users who allowed IDFA collection and Android users. The CPI of $1.25 is the ad performance calculated based on the number of app installs from both user groups. App installs from iOS users who did not allow IDFA collection are excluded.

When using the two attribution models independently, it is challenging to identify app installs performed by specific users based on factors such as OS and IDFA collection. Here you can facilitate Airbridge’s Attribution Overlap feature. Attribution Overlap allows you to see all app installs measured by both attribution models without excluding or overlapping events. 

Consider another CPI example below, assuming that there are 400 non-overlapped attribution events for app installs attributed to SKAN. The ad spend is the same at $1000, and the app installs measured by both attribution models are combined.

The 400 duplicate unattributed app installs are from iOS users who did not allow IDFA collection. The 800 app installs, attributed to the Airbridge attribution model, are the sum of app installs from iOS users who allowed IDFA collection and Android users.

The CPI of 0.83, calculated by adding the two numbers together, reflects the ad performance based on all app installs generated through app marketing. No app installs were excluded based on the operating system, IDFA collection, or any other criteria. By leveraging Attribution Overlap, the final CPI was at least 30% and up to 70% lower than if SKAN and Airbridge attribution models were used separately.

Set up Attribution Overlap in Airbridge right now

How do I set up the Attribution Overlap in Airbridge to obtain the correct CPI and CPA? You can set up Attribution Overlap in Airbridge in just 3 steps below.

  1. Update to the version that supports up to SKAN 4.0 conversion settings page
  2. Select ‘Attribution Overlap’ from Measurement types
  3. Add Attribution Overlap related metrics and GroupBy in Actuals Report

To measure event data with SKAN, you need to set up conversion values. When setting up SKAN conversion values in Airbridge, you can map multiple app events and event attributes to each conversion value. Additionally, you can also map double attributions to a conversion value.

To enable mapping, update the Airbridge iOS SDK to version 1.33.1 or later and switch the SKAN conversion value setting page in Airbridge to the version that supports up to SKAN 4.0. You can ask your Airbridge CSM to grant access to the SKAN conversion settings page.

💡One thing to note before changing the version
Once you switch to the SKAN conversion settings page that supports up to version 4.0, it's not possible to revert to version 3.0. However, the conversion values set in version 3.0 can still be utilized in the updated version. This enables you to continue measuring event data from ad channels that only support SKAN 3.0 by using the version that supports up to SKAN 4.0 on the conversion settings page.

On the SKAN conversion value settings page, select ‘Attribution Overlap’ under the Measurement type. The toggle activates when this option is selected.When Attribution Overlap is enabled, any in-app events mapped to a SKAN conversion value can be identified as having overlap attribution through the Airbridge attribution model. App installs are automatically included in the determination when you map a conversion value in Attribution Overlap.If you select Attribution Overlap, you can also identify overlap attribution for conversion values sent from ad channels that do not support SKAN 4.0.

👉 You can find more information about SKAN conversion settings in the Airbridge user guide.

After collecting sufficient SKAN data, you can review overlap attribution determinations by choosing a metric and group in the Actuals Report. You'll find SKAN metrics that explicitly differentiate between overlap and deduplicated installs.

  • SKAN Overlap Installs (App): Shows the count of app installs attributed to both SKAN and the Airbridge attribution model.
  • SKAN Dedup Installs (App):  Shows the total count of app installs solely attributed to SKAN, excluding those attributed with the Airbridge attribution model

Additionally, there is a GroupBy option. Enabling the ‘SKAN Is Also Attributed By Airbridge(IDFA)’ GroupBy allows you to examine the overlap attribution of SKAN metrics, providing three possible values: true, false, and blank.

  • true:  Shows the number of SKAN metrics attributed to both SKAN and the Airbridge attribution model.
  • false: Shows the number of SKAN metrics attributed to SKAN and not attributed to the Airbridge attribution model.
  • blank: Shows the number of SKAN metrics attributed to SKAN but the “Attribution Overlap” component is not included in the conversion value mapping, or the install is attributed to Meta Business.

👉 You can find more information on how to retrieve SKAN data in the Airbridge user guide.

Airbridge is always delivering accurate ad performance

While the overlap attribution of SKAN events may lead to unmeasured iOS app marketing performance, the ability to more accurately assess ad performance in the era of privacy marks another milestone for marketers.

In addition to attribution, Airbridge has completed the development of the version that supports up to SKAN 4.0 functionality for marketers. Marketers can now preset SKAN conversion values within Airbridge. These preconfigured SKAN conversion values will become increasingly significant as more ad channels adopt SKAN. We are dedicated to advancing these features while closely monitoring SKAN-related trends.

Amplify the success and precision of your iOS app marketing performance with SKAN through Airbridge.

Đăng ký nhận newsletter
Nhận thông tin về các xu hướng marketing, insights.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
Check out our all-in-one package that fits every stage of your growth.
Jeongwon Kim
Product Writer
Subscribe to the newsletter for marketing trends, insights, and strategies.
Get a mail whenever a new article is uploaded.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.

Liên hệ ngay

Tìm hiểu cách bứt phá tăng trưởng
cho ứng dụng.
Đăng ký ngay để nhận insight hàng tuần từ 20.000+ chuyên gia marketer hàng đầu về ứng dụng
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.