There are two contexts in which SKAdNetwork (SKAN) can be used together with Google.
- Google as an ad network – the use case for advertisers.
- Google as an ad monetization platform- the use case for publishers.
Google Ads for advertisers
SKAdNetwork integration enables advertisers to measure the performance of their Google App Campaigns for Install (ACI).
Google ads Privacy Threshold (post updated in Q2 2023): >100 installs per campaign in 24 hours.
Privacy Threshold or Crowd Anonymity after SKAN 4 is a crucial component of SKAdNetwork.
When there are fewer installs in a campaign than in the setup of the privacy threshold for ad network, advertisers should expect the following:
- NULL in Conversion Values instead of actual values in the reporting.
- Ad networks won’t have ConversionValues for automatic ad optimization.
- Missing source app ID
- The source identifier is going to have two digits instead of three or four
The end results will depend on the number of installs in the campaign. There are 4 combinations of the received data that depend on the volume of installs in the campaign.
Google limits the number of campaigns per app that can be run simultaneously to 8.
Google Ads SKAdNetwork integration differs from the integrations with other ad networks – Google does not send copies of postbacks back to MMPs / Analytics. As an alternative, aggregated SKAdNetwork data is available through Google Reporting API.
This creates a number of limitations compared to traditional integrations:
- Data is updated retroactively with a lookback of 5 days. For example, data pulled via API 4 days ago might change when refreshed again.
- There is no real time data. It may take up to 72 hours to update the data.
- Conversions attributed via click and impression are mixed together. There is no way to see the type of attribution.
- API does not differentiate between ‘Null’ and 0 in Conversion Values. Both values are marked as 0.
Advertisers can receive postbacks for iOS 15+ users acquired by Google ads directly from Apple. In order to do this advertiser need to set up copy of postbacks from Apple. This allows to to gain more extensive insights.
Pro tip
Moreover, Google states that both conversions reported in the API and displayed on the dashboard are modeled. That means that Google uses its own logic to attribute conversions. Apple SKAdNetwork is only one signal for them. Therefore, there will always be discrepancies when you compare Google conversions and Apple conversions. Modeled conversions also require extra time to appear in the reporting. It may take up to 5 days to display modeled conversions.
How to use SKAdNetwork with Google Ads
Firebase
Latest version of Firebase SDK enables developers the following:
- Automatically register an app for SKAdnetwork attribution
- Track “first open” conversions
- Compatible with MMP SKAdNetwork solutions
Supported MMPs
There is no huge difference in the Google Ads SKAdNetwork setup with MMP. They are used to map conversions values to specific events in your app and to display SKAdNetwork data in the reports.
Advertisers that are using SKAdNetwork conversion value schemas’ provided by MMPs need to make sure their Google Analytics 4 in-app events match the in-app events of the MMP.
This is required because events may have different naming conventions. For instance, if in MMP’s dashboard the event is specified as “in_app_purchase”, an advertiser needs to ensure that it corresponds to the “purchase_one” event in Google Analytics 4. Failing to map these events to each other would prevent Google from determining which SKAdNetwork conversion values are linked to your partner events and Google Analytics 4 events, and this could make it difficult to calibrate the SKAdNetwork models effectively.
Google Admob for publishers
Make sure to integrate the latest version of Google Mobile Ads SDK. Older versions require additional setup.
If publishers use Google Admob monetization, they need to define SKAdNetworkIdentifier for all used monetization providers.
This form is currently undergoing maintenance. Please try again later.