Mobile App Advertising in 2022
Whilst all eyes are glued on the web world and digital advertising’s migration to the “cookieless” future, the mobile app ecosystem has had significant changes over the past few years which will also come to a head in 2023 if we are to believe Google’s latest timelines for Android.
It has never been more complex to navigate the mobile app side of digital advertising, with the dual juggernauts of Google & Apple responsible for how all advertisers attempt to use App in their media strategy. This blog will cover some of the key areas to pay attention to
TECHNOLOGY
Fundamental to all types of media is having a strong foundation in technology, in order to aid & deliver an effective mobile app strategy. The key questions revolve around how to collect your first party data through your mobile app, which often sits in the form of a SDK (software development kit) that would be installed within the app to help do this.
However with the movement away advertising ID’s such as Apple’s IDFA & Google’s AAID, your technology stack needs to be strong enough to bring data sources together with a foundation based on your own first party data. Add to this the legal requirements of consent, as seen primarily with Apple’s ATT framework on the App Store since iOS14.5, suddenly data can fall away quite quickly. Whilst Google will be making bigger changes going into 2023, some quite large ones are already starting to take place in a similar theme to ATT where consent plays a role in zeroing out the AAID as a default.
For Mobile App in particular, the role of a Mobile Marketing Platform (MMP) is often a big question as to whether there is value in investing in one. It is very fair to say that all of these have had their technology handicapped by the removal of advertising ID’s, but the smarter ones are evolving their approach in a means to keep value. It also removes the need to implement multiple SDKs on an app, which can sometimes be quite long winded processes. There is also a nice interlink with some of the newer technologies on the block, namely Customer Data Platforms (CDPs) and Data Cleanrooms (DCRs).
The other question for technology also comes down to measuring success, which we will dive into more later. But let’s make it perfectly clear that an app install based approach vs in-app events based approach to mobile app are largely different with more complication more than ever. GA4 has been touted as a solution here but it is not going to be able to provide a complete view of all channels in an app based world, in the same way Firebase could never do it.
ACTIVATION
As with the web world, identifiers have been fundamental to how paid media campaigns are targeted on mobile app. The collection of advertising ID’s has lent itself to ad networks, DSPs & vendors collecting data in order to curate / build targeting functionality from. You could argue that unlike the 3rd party cookie, that this targeting is way more accurate / stronger than the web world ever was.
But in a world where these identifiers may ultimately not exist anymore or be dependent on consent being given, this will put certain vendors out of business as well as make it more strategic in how to go about planning a media strategy that doesn’t depend on advertising ID’s.
As usual, certain Walled Gardens retain an advantage in their offering of app based users, providing you stay in their worlds. This is not just specific to the social platforms, but with Google (especially on Android) and how we see the fact that Google Ads app based campaigns continue to be seen as the strongest platform to run app install ads on.
On the iOS side of things, if you are not running Apple Search Ads then you are at a material disadvantage over your competitors, purely to the fact that ASA continues to be the only media platform exempt from all of Apple’s changes since iOS14.5 (more on this later). The expectation is that Apple will continue to evolve its ads offering over time and is needed to keep an eye on.
More traditional targeting options do exist but need a good review before looking to expand with. If you are working with ad networks, understand exactly what value are they giving you from a targeting perspective but also a measurement perspective. Contextual is seen as a safety net especially in the web world but is much more complex in app so do not be fooled by certain vendors who say their technology may work the same here. Data is still power in ad targeting but again reviewing this is going to be key, though some interesting options may be telco powered data by the likes of Skyrise Intelligence or even app based data from the likes of PushSpring.
Finally it would not be Mobile App without talking with some specific nuances within this ecosystem. The most common could be argued to be deeplinking, to which there are 2 different types (deferred deeplinks vs standard). What is clear however is that forms of deeplinking as a concept is challenged where it relies on an advertising ID such as a IDFA. But this should not put off advertisers, considering the power of where it does work in order to drive users to specific sections of an app from a click & how different vendors can extract information around deeplinks to better personalise ad campaigns.
MEASUREMENT
And finally we come to measurement on Mobile App. For this we need to split into 2 key areas of App Install & In-App Events, as well as bring in technology into the equation.
For App Install, this is naturally tied to the App Stores on both Apple & Google Play but also can align to how an “open” is defined within the actual advertiser’s app. For Apple in particular with the exception of Apple Search Ads (which uses Apple's Attribution API with lots more functionality), this is all now defined by the SKAdNetwork, a mechanism for which Apple notifies relevant parties on whether they played a role in an app install. Understanding the SKAdNetwork, how it operates with different media vendors & the concept of “bits” can be quite challenging but is now a fundamental part of iOS based app measurement.
For In-App Events, this is traditionally where SDKs have been able to track but the difficulty is now tying it back to advertising id’s that may no longer exist in the future or be tied back to consent. Lets not forgot that Apple’s ATT framework requires consent on both the advertiser (Brand X) & media vendor you are serving an ad on (Meta / Snap / TikTok etc) in order for it effectively work. There does also exist other identifiers such as Apple’s IDFV, which is still useful for analytics but is not really something valuable from an advertiser’s measurement perspective.
Two key points to make here on in-app events. First is that the SKAdNetwork on iOS side is capable of tracking this but with quite significant challenges / limitations. There are only a limited number of “bits” that can be used as an advertiser, so it is important for the advertiser to set this up appropriate for events that matter to them in an app. These are known as Conversion Values which is a 6-bit value between 0-63 & has privacy thresholds applied to it. This also means that things like calculating accurate ROI via the SKAdNetwork is aggregated at best. On the Google side, whilst a lot of it is currently relatively dependent on the AAID, this will move into the Privacy Sandbox on Android by late 2023. Crucially unlike 3rd party cookies, Google have not said they are getting rid of its advertising ID AAID through this Sandbox but their intention to replace all functions of it suggests that at some point they will. Somewhere in the middle is where the likes of GA4 & MMPs fit in, to which for measurement have certain approaches that do depend on advertising identifiers but also without.
This brings the idea of fingerprinting into play, a concept that is using other identifiers like ip address / OS version to triangulate users. Pretty much every MMP is doing this in their own ways to alleviate measurement concerns. But the writing is on the wall for fingerprinting not just on Apple / Google side but in general & I expect some of these approaches to get found out as Apple in particular continue to bring newer functionality to iOS e.g. Private Relay.
BRINGING THIS ALL TOGETHER
As with the web world, the app ecosystem will continue to shift. By the time Google hopefully successfully launches some of the Privacy Sandbox proposals on Android, some of which really does bring efficiency to the ecosystem (SDK Runtime), the app strategy will have fully migrated away from the legacy IDFA/AAID approach.
The key points for any advertisers both looking to launch an app but also ultimately just serving ads in an app environment should consider are:
Make sure your technology is fit enough to cover a world without advertising ids for paid media sake but also collect / manage / activate your first party data alongside website / offline / CRM needs
Plan for a mediaplan that works more tied on publisher specific USPs & signals that do not rely on identifiers (geo for example has always been king of mobile app)
Do not be afraid to trial out newer media platforms who have made the effort to integrate with iOS / Android in a more futureproof way whilst also keeping an eye on what Apple may end up doing beyond Apple Search Ads
Educate & learn how technology shifts have impacted measurement, both from the nuances of API’s (SKAdNetwork & Privacy Sandbox on Android) but also what other technologies offers (GA4 / MMPs / DCRs)
For a dual iOS / Android app strategy, understand the limitations of what may not be possible on one vs the other but also when it comes to reporting / measuring success, how stating something like a ROI may not be accurate
All of these points should fit in with a wider advertising strategy that may span offline, web & emerging channels with a more macro measurement framework in place to monitor success