The impending privacy changes coming to the iOS 14 through the deprecation of the IDFA disproportionately harm companies that have built deep user profiles for the purposes of ad targeting. Facebook, having built extremely sophisticated, precise targeting infrastructure, will perhaps suffer more than any other mobile advertising platform as a result of IDFA deprecation: all of its optimization mechanisms are built atop user-centric profiles that rely on in-app events being tagged with device identifiers. Without being able to attribute in-app events to a particular device, and thus a Facebook account, Facebook’s advanced, hyper-precise ad serving machinery is weakened.
The Sword of Damocles is meant to represent the unpredictable, potentially existential risks faced by those in power: Damocles hung a sword above his throne with a mere strand of horse’s hair. Facebook’s hyper-precise ad targeting is its power — it is the advantage it has over all other advertising platforms (apart from Google UAC). And IDFA deprecation is Facebook’s Sword of Damocles: a risk it always faced for becoming too skilled at targeting ads at users via aggregated behavioral profiles.
Facebook has grown its advertising revenue considerably since 2017, when it introduced the event-optimized App Event Optimization (AEO) and Value Optimization (VO) campaign strategies: all of that revenue growth is attributable to Facebook’s ability to calculate the probability of a specific user clicking on an ad, installing an app, and ultimately, making one or more purchases. With iOS 14, Facebook’s ability to do that is diminished: it can know which users click ads, but it can’t tie in-app events to individual user profiles, at least not in real time and in a reliable way. The changes introduced in iOS 14 were always a possibility, and the more money that Facebook made with its ability to target ads to users with stunning precision, the more likely these changes were to be enforced.
In understanding just how important these event-optimized campaign strategies have been for Facebook, it’s helpful to go back to 2017, when they had only just been rolled out. In Facebook’s App Event Optimization tool showcases the power of its data in Q1 earnings, published in May 2017, a little less than a year after AEO had been introduced, I wrote:
Facebook’s AEO product is another example of the advertising giant utilizing its massive data set on users as a competitive advantage that is almost structurally unbeatable. Other networks can optimize around in-app events, and some do, but they don’t have the first-party engagement data or the deterministic profile data that Facebook does to cluster players together based on propensity to spend. Absent those data sets, AEO is hard to do: since very few people spend money in apps, a lot of data is needed in order to build robust proxy signals for spending. Facebook has that data and almost no one else does.
ARPU in Facebook’s most important market (North America) increased 9.75% and 9.81% quarter over quarter in Q2, when AEO was introduced, and Q3 of 2016, versus 4.85% and 6.43% worldwide. Yet the quarter-over-quarter increase for North America was roughly equivalent to that of worldwide from Q1 2016 to Q2 2016, before AEO was launched.
It’s important to keep in mind that the introduction of the AEO campaign strategy was a means of improving per-user monetization in high-ARPU geographies for Facebook: AEO is designed to focus a campaign’s reach on users that are likely to complete some event, meaning costs of installs increase, but the value of each acquired user increases enough to create viable unit economics. In other words: AEO campaigns are designed to be, relative to campaigns that simply optimize for installs, low volume and high ARPU.
It bears pointing out that there are really only three ways for Facebook to increase its revenue: to scale its user base, to increase ad load, or to increase the value of its ad inventory. In North America, Facebook only increased DAU by 35% between Q4 2016 and Q4 2019 while its (global) revenue more than doubled from 2016 to 2019. And while ad load has certainly increased in that period, Facebook’s CFO noted in 2016 that ad load would reach a natural limit across its core portfolio of apps at some point in the near-term future. It seems likely that the source of Facebook’s revenue growth from 2016 until today was its ability to increase the value of its inventory through event-optimized campaigns like AEO — which, as the diagram above points out, are heavily dependent on user profiling.
The primary beneficiaries of AEO and VO campaigns, and generally Facebook’s increased capacity for precision ad targeting, have been smaller advertisers and SMBs: they were given access to the types of sophisticated targeting machinery that would have been wholly inaccessible to them otherwise. The point I made when Google subsumed all of its mobile app advertising inventory into the UAC product was that, while the lack of transparency of UAC might have irritated larger advertisers, the product gave small companies the ability to advertise by leveraging Google’s incredible advertising tools. These event-optimized, algorithmic targeting tools not only drove ARPU increases but they created opportunities for efficient advertising for smaller companies who otherwise wouldn’t be able to spend money on performance marketing at all.
Some caveats. It should be noted that IDFA deprecation only impacts app advertising campaigns, not web advertising. Although Facebook revealed that 94% of its 2019 advertising revenue was generated on mobile, the company doesn’t break that out between mobile web advertising (eg. a user clicks on an ad and is taken to a website) and mobile app advertising (eg. a user clicks on an ad and is taken to an app store). It’s also worth mentioning that, as of now, only the iOS device identifier has been deprecated (and even then, only effectively deprecated: some percentage of users will elect to opt into ad tracking, and their IDFAs will be exposed to Facebook), and the GAID on Android devices is still accessible to all apps for ad tracking.
But my best guess is that the majority of Facebook’s mobile advertising revenue is driven by app advertising, and I believe that Google will deprecate the GAID within six months. So the changes introduced in iOS 14 represent real, immediate problems to Facebook’s revenue growth. With IDFA deprecation, the strand of hair dangling the Sword of Damocles above the throne has snapped. Facebook might expect to see its revenue decrease by some meaningful percentage: just a few days before WWDC 2020, Facebook published a white paper in which it revealed that ad personalization accounts for 50% of CPM prices on Facebook’s Audience Network (the aggregated inventory across 3rd-party apps for which it manages ad targeting and serving). The conclusions in the article are withering:
It seems nearly impossible that advertisers won’t face deteriorating economics on Facebook in the short term as IDFA deprecation materially impairs Facebook’s ability to precisely target users. Over the long term, I believe that Facebook will find a path to its current level of ad serving efficiency without needing advertising identifiers. But the content of its own white paper underscores very clearly how important personalization is for ad targeting, and IDFA deprecation damages Facebook’s ability to deliver that kind of personalization.
Photo by Ricardo Cruz on Unsplash