At Smadex, technology is at the core of our day-to-day. As Chief Data Officer, I make sure our team is constantly learning and improving. It’s the only way to face the challenges of this always-changing world of advertising. This brings us to the changes coming along with the upcoming SKAdNetwork based attribution and its changes to programmatic.
Given the upcoming release of the App Transparency Tracking (ATT) framework, our tech team went back to the basics. Only with one goal in mind: redesigning the way our algorithms work on the new data flow.
In preparation for the depreciation of IDFA, we’ve been reformulating how our machine learning makes sense of the data
SKAdNetwork based attribution: How will optimization work post-ATT?
With this in mind, the Smadex Data Science Team has been training new algorithms to outperform on SKAdNetwork based attribution. Our goal is to make the most of the data available: not looking only at SKAd enabled postbacks but the campaign as a whole.
Making the most of our data sources
- Frozen LAT campaigns. While running LAT campaigns before ATT full deployment, we get all the contextual information available. Once ATT is deployed, we’ll keep all the insights we have gained during this pre-ATT phase. Those learnings will still be available for future campaigns.
- Living device ID data. The need for the opt-in rate will likely lead to a very low number of available device IDs. After ATT deployment, learnings can be extrapolated to the users that have chosen to opt-out.
- Living Android campaigns. On the whole, most advertisers have an Android version of their iOS app, and we’ll keep receiving the same data as of now. In effect, our algorithms can extrapolate those Android learnings into the iOS campaigns with excellent results.
- Living SKAd postbacks. For SKAdNetwork based attribution, we’ll need to work within the restrictive Apple’s privacy threshold. At Smadex, we are developing new functionalities within our platform, such as the new Flexible Mapping.
Are you up to the SKAd challenge?
The most amazing thing is that we can combine different sources depending on the amount of data available for each one; that’s where the magic happens. As a result, we can leverage the information provided by each of our data subsets to fine-tune the algorithms that provide a single output for each ad request we process in real-time.
All in all, it’s a high priority for us to work with our partners ahead of the full implementation of App Tracking Transparency and SKAd based attribution, minimizing the transition to the new post-iOS14 era.
Shoot us a message to get more information on how to operate and succeed in the post-IDFA world.algorithmsATTattribution modelsIDFAios14optimizationSKadnetwork