Since Apple’s App Tracking Transparency (ATT) framework came into effect, the company has started confronting users with daunting warnings about favorite apps tracking you across apps and websites owned by other companies, and inviting consumers to opt out of this tracking. The impact on measurement has been material. Tim Cook would call such opt-out language “merely descriptive” — and would neglect to recall that Apple invented IDFA for sound and pragmatic reasons (including privacy). However, it could be argued that the language is scary and intended to persuade people to opt-out. It is important to note that app publishers are not permitted to include their own justification at the point of decision. Only Apple’s framing is available.
Since the changes were made in April of 2021, estimates of opt-out rates have been all over the map. Our initial models (chart below) used 80% (meaning 80 people opt out of tracking and 20 people permit it). Since that time, we’ve seen estimates ranging from 95% to 60% opt-out rates. But credible players in mobile infrastructure who see data on thousands of apps across all hardware types have noted that the opt-out rates have been declining since the initial surge, and an estimate of 60 to 65% is probably more accurate today. But Apple not the only company to keep an eye on when it comes to tracking and measurement.
Two key players to watch in terms of evolution of measurement are Facebook and Snap. Both companies are mobile-forward, and both have built material businesses around delivering measurable advertising for customers on phones. Facebook has acknowledged that the changes in iOS would have “on the order of” $10 billion impact on sales in 2022. Analyzing the 2022 revenue impact from Apple’s privacy changes, Lotame estimates that the IDFA change will have a total impact on the companies of almost $16 billion (+9%). Again, with the lion’s share of that impact (81%) coming from Facebook.
Both companies have spoken extensively about the impact of the iOS changes on tools and capabilities, and both are adapting rapidly. For SNAP, the company has said it is adapting by a combination of pushing customers to use Apple’s SKAD Network (“SKAN”) and by developing and enhancing native tools such as the Snap Pixel and their Advanced Conversions (“AC”) tools. Snap also acknowledged that many of its customers would have to move their focus “up the funnel”.
The issues are that the native tools may have limitations and may require Snap to interoperate more with other ad tech players. As for Apple’s SKAN tool, it provides only aggregated data, and it delivers with a 24-hour delay. These are very material limitations for conversion-oriented marketers. Snap has said it will take time to work through all of this.
Facebook, equally frustrated, even used the term “rebuilding”: we’re rebuilding a lot of our ads infrastructure so we can continue to grow and deliver high-quality personalized ads. Facebook called its new tools “Aggregated Event Measurement.” But at the same time Facebook also threw Apple under the bus for providing data that is both delayed and aggregated. The suggestion is that Facebook — which does not play well with the rest of Ad Tech — will have difficulty providing measurement better than Apple’s dumbed-down SKAN utilities.
One major initiative at Facebook did become highly salient: its focus on ecommerce infrastructure for its marketer customers. This was described by Sheryl Sandberg as one of seven key initiatives, and it was mentioned numerous times in the 4th quarter earnings call. The idea is for Facebook to continue to deliver and enhance end-to-end tools for its ecommerce seller. From the first marketing impression, to the ecommerce store and experience, to financial fulfillment, customer retention, and remarketing, Facebook wants everything to happen on Facebook.
The issue is that any time any part of the chain moves outside of Facebook’s app, Facebook can’t deliver the measurement and attribution its customers require. This became painfully clear with the iOS changes. Facebook’s mission to fix it will lead it to the erection of higher walls around its version of the internet, and more firmly bunkering its marketer customers inside.
When you consider how other players may be future-proofing mobile marketing strategies, initiatives will be similar to those of Snap and Facebook: using the SKAD network while holding their noses, building home-grown measurement tools that often won’t work as well, and relying on the ad tech industry to provide adapted tools, such as the Lotame Panorama ID which can link identity across devices and platforms without the IDFA. App developers will also be looking to those partners that provide tools and infrastructure, such as AppLovin or MoPub to deliver new tools and techniques.
And all this enthusiastic turbidity and debate leaves all of us wondering: “What will Google do?” Keep in mind that Google doesn’t have the fire to lock down mobile phone ad IDs that Apple does. Apple uses the cover of “privacy” to make these changes. But others feel that Apple is restricting the ability of others to monetize traffic on its devices without its direct involvement (and economic participation).
Google, on the other hand, for the infrastructure it delivers, has benefited by providing tools and capabilities that are open and used by everyone. In June of 2021, shortly after Apple made its changes, Google announced it did not plan to bother users with a push demand for opt-in on each app (both Apple and Google always have allowed global device opt-out). After that, Google was quiet until it capitulated a few weeks ago. As of mid-March, Google is saying that it will “replace” GAIDs (Google’s version of the IDFA) on phones with its web-based “privacy-sandbox” tools, and then phase out GAIDs in about two years.
What does that mean for cross-application tracking today? Well, even if we lay aside second-best workarounds that are progressively emerging, there is still a lot to do with mobile ad IDs. In the U.S., iPhones have about 58% market share, but some reports are seeing 40% of users opt-in. Android phones have about 42% market share and will have GAIDs available for at least two more years — an ocean of time. Considering this collectively and scoped to the U.S., that still leaves marketers with access to a mobile advertising ID on about two-thirds of inventory. Have fun while it lasts!