How has ATT impacted mobile advertising?

The Financial Times this weekend published an analysis of mobile app advertising data provided by analytics firm Branch. The article concludes that Apple’s advertising product, Apple Search Ads, has captured a material portion of the mobile app advertising market since Apple made its App Tracking Transparency (ATT) privacy policy mandatory in April of this year. The author of the article published the below graph on Twitter, which captures the crux of the dynamic as measured by the Branch data:

While, anecdotally, I can affirm that, based on my limited view of the market, Apple’s ASA has captured increased market share since ATT was introduced, I do believe the data used in the article is open to misinterpretation that potentially overstates the broader market developments in that time. There are a number of reasons for this:

  • Branch admittedly works primarily with non-gaming apps. Since gaming is the dominant category in terms of mobile app install ad spend, this data is biased, but my general sense — which is not quantifiable — is that the shift in ad dollars from games advertisers into ASA and away from other channels likely aligns with what is depicted in this graph;
  • This data only captures mobile app install ads spending and not mobile web ads spending. Any examination of ASA market share will of course be scoped to mobile app install ads spending since ASA only traffics in mobile app install ads, but nonetheless, the larger part of the mobile advertising market is missing in this analysis;
  • This data is presented as a percentage of total attributed installs. Given that these figures are percentages and not absolute values, the total market size of iOS installs is a relevant consideration which is omitted here, ie. if the total market shrank, Apple’s increased share might not be indicative of increased revenue or scale. Note that I don’t believe this is the case, but without this context, interpreting this data to any conclusion is difficult.

The article cites Evercore ISI’s recent research note which estimates that Apple’s advertising revenue will reach $20BN in 2024 from around $2BN in 2021. I covered this estimate in Apple robbed the mob’s bank, part 2, although the article proposes that Apple might earn $5BN in advertising revenue this fiscal year (remember that Apple’s 2022 fiscal year started on September 26, 2021).

It’s important to note that Apple slow-rolled adoption of iOS 14.5 and iOS 14.6 and didn’t aggressively push the ATT-mandatory update to iOS devices until early June 2021, just ahead of its WWDC developer conference. So the dramatic expansion of Apple’s advertising business only began in earnest four months ago when a majority of iOS devices were running an ATT-mandatory iOS version, and it has already reached a projected (or, rather, projectable) run rate of $5BN in that time.

This is stunning growth for an advertising channel that was mostly seen before ATT as an ancillary source of marginal new installs for most app developers. It’s not unexpected, of course: I have covered the ways in which Apple privileges ASA relative to other networks under the strictures of ATT, primarily through the provision of a separate and more granular reporting API than SKAdNetwork.

But in the broader context of the mobile advertising market, does $5BN matter? $20BN? As I point out in Apple robbed the mob’s bank, part 2, $20BN would represent 17% of Apple’s services revenue, which is an important, narrative-affirming growth vector for the company. But if Apple gains $20BN in advertising revenue from some combination of other channels, against the backdrop of a buoyant digital advertising ecosystem that will likely see significant growth over the next few years, what difference does this re-distribution of budget make to advertisers?

My answer to that question doesn’t differ meaningfully today than it did as a prediction in IDFA deprecation: winners and losers, which I wrote more than a year ago in October 2020.

Direct response mobile web advertisers, particularly DTC and e-commerce advertisers, are panicked: they have no ASA equivalent to channel money through, and the ad spend that was erstwhile deployed on Facebook has simply evaporated — and, with it, the attendant revenue. Facebook’s Content Fortress perhaps provides some opportunity to reclaim spend through shifting the consumer transaction experience into the Facebook or Instagram app, but even for a reasonably evolved product like FB Shops, full adaptation to the ATT environment will take some time. In the short term, DTC and e-commerce companies are likely to feel immense pain as a result of ATT.

App advertisers are likely to see a degradation in performance even with available advertising alternatives such as ASA and broker networks simply because Facebook, through its sophisticated tools like AEO and VO, was so efficient at reaching relevant users before ATT. Broker ad networks are able to fingerprint for attributing, which makes measurability on those channels seemingly comparable to the pre-ATT environment — although, as Alex Bauer notes in episode three of my ATT: One Month In series, fingerprinting is mostly good at finding a click to attribute to an install and not necessarily the right click — but nonetheless, ad spend follows performance. In “Mobile advertising spend can’t just disappear”, I describe how, absent improbably large operating margins for advertising campaigns, ad spend will shrink in the face of degraded channel performance. This reality simply can’t be escaped, although I identify some mitigation tactics in this piece.

What will perhaps change even more dramatically, however, is the discourse animated by this topic. Firstly, ATT is becoming old news: I published The coming war between Apple and Facebook in January 2017, Apocalypse Soon: What happens when the iOS advertising ID is deprecated? in February 2020 (ahead of Apple’s announcement that the IDFA would, in fact, be deprecated), and Facebook may take 7% revenue hit from Apple privacy changes in January 2021. I have been talking about a potential paradigm shift in mobile measurement for four years, and talking about it almost daily since WWDC 2020.

Regardless of my belief in the purity (or coarseness) of Apple’s intentions with ATT, that privacy policy was rolled out, and it’s a reality now. Debating its merits seems unhelpful and unproductive. As I make clear in The IDFA is the hydrocarbon of the mobile advertising ecosystem, I believe the IDFA facilitated an invasion of user privacy that was not justifiable and that it needed to be retired. The fact that I don’t think ATT was the perfect or most righteous means of doing that is irrelevant at this point.

Second, this conversation space has become febrile and toxic. The role that social media plays in society and the guardrails that should be installed to constrain its negative influence should be discussed. But the primary concern of this blog and the broader Mobile Dev Memo community is mobile advertising and freemium economy design. And with respect to Facebook specifically, I simply have no skin in this particular game. I’m happy to engage with the obstreperous church of digital advertising denial, but debate further afield than that becomes unproductive quickly. I believe that personalized advertising is a public good, and I intend to dedicate the words on this blog to further that purpose.