Facebook has reacted to Apple's effective removal of ad tracking in iOS 14 by announcing it will be reworking its entire mobile advertising platform.
Previously, Facebook has generated billions of dollars in advertising and driven billions of app downloads annually by using Apple's device-level anonymous IDFA ad tracker to directly target (or personalise) mobile ads to iOS users.
However, in iOS 14, users will be more easily able to opt-out of ad tracking.
Most commentators expect the proportion who do so to increase from around 30 percent at present to 80 or 90 percent, which will render advertising methods that rely on IDFA tracking irrelevant.
At least that's Facebook's view.
A new approach
It's announced that it will no longer use IDFA for its own apps running on iOS 14, also releasing a new SDK for its partners to use.
This will rely on Apple's own SKAdNetwork ad network, which effectively becomes the only method of running ad campaigns on iOS 14.
Unfortunately, however, SKAdNetwork offers only very limited functionality, which means it will be more difficult to target users and measure the effectiveness of ads. This will reduce the effectiveness of user acquisition, as well as the advertising revenue many developers depend upon.
"App developers and publishers should expect lower CPMs," Facebook comments.
It’s not clear whether Apple views this lack of functionality as a feature to impact the advertising businesses of competitors such as Facebook and Google, or whether the company doesn't have the experience to build an effective advertising platform.
Facebook says the changes decrease the ability of its and its partners' ability to monetise effectively
Either way, Facebook says the changes decrease the ability of its and its partners' ability to monetise effectively.
Indeed, it adds Apple may end up making this new SDK so ineffective that Facebook may have to shut it down.
In the long term, it's expected that Facebook will launch a new advertising platform to enable it to target users and track advertising campaigns without targeting individual users.
This would deal with the specific issue of iOS 14 and the more general issues around consumer privacy that underpin in.
Such so-called 'probabilistic' models rely on analysing high volume of open data and then using machine learning techniques to try to work out how different types of users will react to ads; also constantly revising the result to optimise user acquisition.
UA expert Eric Seufert calls this move "substantial"; heralding a "new era in mobile measurement".