.jpg)
It has been nothing short of a disruptive year for advertisers, with many drastic changes to the mobile app advertising landscape. From iOS 14 implications on mobile attribution, to technological advances that followed, read on as we explore it all, and what it means for deterministic attribution as we know it.
What is deterministic attribution?
Deterministic attribution is the technology used to identify the channels that generated an action. Deterministic Attribution uses device IDs to identify the same user across different touchpoints, such as clicks on the ad, installs, in-app purchases, etc. Because the device ID is used, the match is 100% accurate for attribution. Deterministic attribution can work with the following 3 main attribution models:
Last-touch attribution: the conversions are attributed to the last click before the conversion action
Multi-touch attribution: this model assigns varying weights to different traffic sources for an advertising interaction. This leads to multiple channels sharing the responsibility for a conversion.
View-through attribution: this model attributes a conversion generated by an impression and not a click.
Deterministic mobile attribution solution:
Deterministic attribution technology is the most accurate way to link marketing actions and sales/conversions, as it uses a unique identifier (IDFA for Apple and GAID for Google) for every action (i.e. impression, click) and every device that is involved in the process.
This makes the attribution very precise, as opposed to probabilistic attribution models that don’t rely on such unique identifiers.
Probabilistic vs Deterministic attribution
The majority of users travel a winding path through a multitude of touchpoints before downloading your app. There are two primary identity resolution models used to bridge this identity gap: probabilistic and deterministic attribution. It's critical to comprehend how they're employed and the data they provide because each one has a particular function.
While deterministic attribution doesn’t allow for “randomness”, as it is aware of the entire user journey, probabilistic attribution relies on clues and follows a set of probabilities that the install or any other action was driven by one or more campaigns. The fundamental difference between the two is that deterministic attribution uses ID matching to tell with 100% certainty exactly what the user’s path was and what app campaigns successfully contributed to the user downloading the app, while probabilistic attribution relies on probabilities.
However, post iOS 14, the rules of the game have changed, and thanks to Apple’s App Tracking Transparency limiting tracking abilities, it is now much more difficult to manage by deterministic attribution. Comparing deterministic attribution to that of probabilistic is a vast topic, especially post IDFA. Read our blog ‘What you need to know about mobile attribution post iOS’ to gain a deeper perspective.
Deterministic Attribution and SKAdNetwork
Apple’s deterministic mobile attribution solution is SkadNetwork. Post iOS14 and all the privacy changes that came with it, SKAdNetwork is the only deterministic model allowed when users don't explicitly give their consent to be tracked across mobile apps.
SKADNetwork is Apple’s privacy-friendly way to attribute impressions and clicks to app installs on iOS apps. It shares conversion data with advertisers without revealing any user-level or device-level data. The limitation of this is that it takes from 24 to 48 hours for the data to be evaluated and then sent to the advertisers, as well as making it difficult to carry on optimization towards post-install events.
However, there is a solution for attribution in the mobile business; it is called “predictiveness” and it involves machines.
In the near future, machine learning will become more important, as it is a means to predict the success of a marketing effort.
Algorithms are created to do things a human being couldn’t do even with an infinite amount of available time. Artificial intelligence helps us understand trends, user behaviors and journeys in a way that is respectful to the user.
In the iOS 14 reality, with limited available data, machine learning can give analysts and marketers a complete view of how their efforts are performing. Thanks to our machine learning algorithms we are able to be SKAdNetwork compliant and run app install campaigns with it without lowering our performance.
