Contextual app targeting in the post-IDFA era

In the short time that passed since Apple’s announcement, much has been written about the new challenges that our industry will face. While it is clear that this seismic shift in the app ecosystem will have far reaching implications, times like these are ripe for industry innovation.


Much of the focus so far has been given to the impact on attribution and measurement with Apple’s rollout of SKAdNetwork, but without the IDFA as a persistent identifier, a complementary and equally important effort should be made on the campaign targeting and optimization front. While the exact nature of Apple’s updated approach towards the IDFA was not foreseeable, the writing has long been on the wall about the coming of a significant change. Fyber has been preparing for this new reality for quite some time and we are now taking our first step in this journey, by unveiling in-app contextual targeting built for the post-IDFA era.


Context that drives performance


For the vast majority of Fyber’s demand partners, which include top performance and brand DSPs, the IDFA is a pivotal input to the machine learning algorithms that drive their bidding decisions. With the IDFA expected to become a scarce resource, DSPs will have to explore alternative, privacy-aware data points and evaluate their value in driving campaign performance and ROAS.


For many, contextual targeting may seem too big of a leap from the precise user-level targeting that drives user acquisition campaigns today to provide a worthwhile alternative. However, in the absence of a persistent identifier, Fyber believes that an app-focused version of contextual targeting can deliver meaningful value to performance marketers.


In just the last few years we’ve seen multiple examples of how certain ad tech concepts from the web environment are executed very differently when imported to the app environment (e.g. header bidding vs. in-app bidding).


App-focused contextual targeting is yet another example of this trend. Unlike its web ‘sibling’ it cannot rely on written content (which is not applicable to the vast majority of apps) to deduce the context. Instead, it should orient around targeting parameters that include things such as:


  • Behavioral signals from the current app session – How many impressions did the user see in this session? For how long has the app been in use?
  • Device status signals – Does the user have a strong network connection and enough battery to download a new app? Is audio on or off? Are headphones currently being used?
  • Enriched app-level data – What framework does the game run on? Does the target audience prefer games running on Unity/Unreal/Buildbox etc.?

Click here for the full list of contextual targeting parameters


Here are two of the main reasons why we feel contextual app targeting is an exciting opportunity:


  • It provides context that is generated from a mix of data points coming from the device, the underlying technology, and the content, blended with session-level behavioral information.
  • It leverages privacy-aware parameters that cannot be used to identify an individual user or track users across apps.


Fyber will make these parameters available to DSPs as part of the bid request sent by Fyber in each and every auction, and DSPs would be able to utilize these parameters as inputs to their machine learning algorithms, and evaluate how each parameter (or combination of parameters) impacts campaign performance.


This solution will be rolled out for iOS and Android in parallel. It is very logical to assume that the GAID will have a similar fate to the IDFA, but for the time being, this release will also enable DSPs to train their machine learning models in using these parameters while a persistent identifier is still available on Android.


Many of the data parameters that Fyber will be making available programmatically have long been in use by ad networks via their SDKs. By sharing these parameters with DSPs, Fyber serves its mission to be the SDK for the SDK-less, putting DSPs at a level playing field with ad networks. We believe this will also benefit our publishers by further enhancing the positive competition between demand partners that drives superior monetization results, without asking publishers to integrate more SDKs.


Thinking about the big picture In the last two years, we have seen several trends emerge in the app ecosystem:


  • In-app bidding is gaining adoption, delivering not only better monetization results, but also more operational efficiency and much-needed transparency into ad monetization.
  • DSPs command a growing share of UA marketer’s media mix, often delivering superior results to those delivered by the ‘traditional’ UA channels.
  • Brand advertisers are warming up to the app environment in general, and games specifically, realizing that user eyeballs have shifted to apps across all audience segments.


The entire app ecosystem has a vested interest in the success of bidding, as a means to diversify demand and increase transparency and control for advertisers and publishers alike, all in a privacy-aware manner. Fyber has long been committed to the cause of promoting programmatic adoption by apps. With this release, we aim to establish standardized OpenRTB custom extensions that can be seamlessly adopted and used for targeting by the programmatic app industry.



Read these next

Contact Us

    By sharing your information you are agreeing to receive communications in regards to any questions or requests submitted on this form. Fyber will keep your information solely for internal tracking purposes and will not use this information for any other purpose. You may request to delete the information provided at any time.

    If you send us a message by clicking the "Send" button, we use a recaptcha service provided by Google LLC to check whether the message was sent by a natural person or a computer program ("bot") in order to ensure that only valid user requests are forwarded to us. Google LLC processes personal information from your browser, such as your browser settings and your click behavior on this screen. Please refer to the Privacy Policy for further information on data processing procedures of our third-party services.