In late June, Apple made a major announcement at its annual Worldwide Developers Conference (WWDC 20).
In line with their focus to protecting user privacy, Apple announced they would be changing the IDFA (ID for Advertisers) data-capture process from an opt-out to an opt-in process. Many industry observers are calling this the death of the IDFA and the subsequent ability for targeting and attribution within app-based advertising.
To understand the significance of this announcement, it is worth reviewing what the IDFA is and the benefit it provides developers, ad networks and app owners/advertisers. The IDFA is an identifier that is appended to an individual device using iOS 6 and newer. It is the cookie equivalent of the Apple operating system - a piece of code to track an individual user, append data to the individual, and be able to track the actions they take. Namely, it is the ability to track which apps people download and how much money they spend within those apps.
Changes to the IDFA
Changing the IDFA to an opt-in process is predicted to reduce visibility and tracking of what an individual does within the App Store and particular apps by up to 95% (Adexchanger, Soomla). The fear creeping through the industry is that reduced tracking removes the ability for targeting and personalization (essential for performance efficiency) and tracking (essential for attribution and optimization).
Within FRWD, one of our four key products is based on marketing efficiency. While efficiency can be achieved via a number of different levers, targeting and personalization through the application of data is one of the most prominent means to deliver efficiency. The IDFA tracking change directly impedes advertiser efficiency within apps and the App Store.
With smartphones accounting for 67% of digital traffic and apps contributing ~70% of total minutes spent on smartphones, the change to tracking has widespread effects (Comscore, June 2020).
Apple currently holds a 42% share of the app market with 1.85m available apps (Statista, Q1 2020). Following Google’s announcement in January to deprecate third-party cookies by 2022, it is widely expected Google will follow suit with their Android operating system and require an opt-in process for their equivalent Google Advertising ID (GAID) (AdAge, Jan 2020). The implication is Apple’s decision will indirectly impact nearly the entire smartphone and app market.
What are the key implications and solutions?
- Personalization: Unsurprisingly, 89% of digital businesses invest in personalization according to Forrester. Gartner projects smarter personalization can help increase profits by up to 15%. Personalization does not just mean which message will be served, it also includes retargeting, frequency capping, etc.
With a reduction in the data being captured, there will be a direct impact on the ability to target different audiences and personalize messaging – meaning the performance gains can no longer be achieved at the same scale.
- Attribution: The fundamental impact is losing the ability to track the performance of campaigns at an individual level, i.e. it will no longer be possible to tell if a particular ad drove an app install and whether the install delivered any revenue. This makes metrics such as cost-per-install and return-on-investment essentially redundant as they can no longer be calculated. Furthermore, it removes the ability to understand which audiences were most likely to install apps and spend money within the app. This makes predictive lifetime value virtually impossible to calculate and therefore removes the ability to optimize campaigns towards higher-performing audiences.
One major solution Apple announced is the soon-to-be launched SKAdNetwork: tracking which aggregates data to allow ad networks to understand the total number of apps downloaded from advertising campaigns. While the aggregate data is key to the program, it is also a barrier. To maximize efficiency, it is critical for tracking and optimization to be performed at an individual level. Based on industry forums and conversations held with industry experts, the general consensus is the SKAdNetwork in its current form is an inadequate substitute and in need of enhanced tracking capabilities.
With realistic opt-in rates expected to be around 10-30% (Adexchanger, Soomla), this provides a small sample size of conversion data to attribute success. It also requires a large extrapolation off a small sample to forecast revenue generation. Put simply, it will be difficult for advertisers to understand how their app campaigns have performed from an ROI perspective and challenging to extract audience insights for profiling, segmenting and targeting.
Why would Apple (and inevitably Google) make such as a drastic change – especially creating a barrier to the ad networks who provide a key revenue source?
It comes down to user privacy. Customer data is an extremely valuable and risky commodity. In the wake of the Cambridge Analytica scandal, Facebook lost $35bn in market value (Fortune, March 2018). That’s the equivalent of ~47m iPhone 11 sales (Apple.com, August 2020). By aggregating data and explicitly requesting users to opt in, Apple is attempting to avoid a serious data breach.
To emphasize its commitment to user privacy, Apple used its presence at the Consumer Electronics Show (CES), an event traditionally used to showcase new products and upgrades, to talk solely about user privacy. In their first CES participation in 28 years, the focus on privacy was then supported by a large above-the-line marketing campaign.
Similar discussions of user privacy have occurred in the last three years. In May 2018, the industry went through a similar change when the General Data Protection Regulation (GDPR) came into effect. While industry experts predicted the swift decline of user opt-ins and the “crumble of the cookie”, the industry appears to have self-regulated itself. Quantcast reported 90% opt-in rates as users traded off exchange of their data for access to web content (The Drum, July 2018).
Does this offer a glimmer of hope for app advertisers? Potentially.
However, a key difference is Apple’s opt-in requirements are two-fold: the opt-in is required both within the app where the advertising appears as well as any app downloaded directly from the advertising. Higher opt-in rates are likely going to require some level of incentive from the app owner.
Overall, there is no doubt a direct impact on audience sizes and the potential data available for targeting and attribution will occur. The impacts will be felt on both the advertiser and app owner/vendor side.
For App Advertisers:
- There will be a smaller subset of the target audience available for targeting, which will impact conversion rates and efficiency.
- Potentially higher costs for media as ad networks place a premium on the reduced targetable inventory available.
- Attribution will be less precise, particularly in the initial 6-12 months. This will impact the ability to assess campaign performance, strategy and overall ROI.
- Lower precision of predictive lifetime value.
- A potential greater reliance on existing apps using alternate tracking measures may rise, such as email and phone number login (e.g., Facebook/Instagram, YouTube, etc.). This may raise the walls on the so-called “walled gardens”.
In response, advertisers have three non-mutually exclusive options:
- Shift spend to alternative platforms (e.g. social).
- Negotiate to pay less for what will essentially be lower quality audiences (e.g. lower CPI).
- Forge stronger partnerships with publishers and tech platforms to facilitate data sharing to offset higher share of anonymous users.
For App Publishers/Owners:
- Reduction in inventory quality due to a higher share of anonymous users is possible.
- A poorer user experience as user opt-in is required and potentially requested on each app visit, interrupting the user from accessing the content and causing higher early churn rates.
That said, publishers and app owners also have options:
- Create “average audience” profiles based on non-anonymous user base, offering advertisers a view on what kind of audience they can expect to reach to offset the uncertainty of targeting fully anonymized users
- Create “data walls” or incentives where the content within the app can only be accessed via opting into IDFA tracking, similar to how publishers responded to GDPR on web browsers. Example incentives have included:
- Gaming apps: providing access to certain levels or gaining rewards to be used in-game by opting in.
- News apps: gaining access to a certain amount of articles for free by opting in.
Conclusion and next steps.
Apple’s commitment to user privacy is only going to steepen. To help navigate the changing landscape, the FRWD team can support in multiple ways:
- Help advertisers restructure and prioritize their targeting. Apple previously allowed unlimited campaign lines for targeting. However, with the introduction of the SKAdNetwork, advertisers are only able to use 100 lines for targeting (RevenueCat, June 2020). As such, the setup of campaigns is critical to provide the best opportunity to gather insights.
- Setup a test-and-learn approach comparing app activity to other channels (e.g., social) to benchmark performance and uncover the optimal marketing mix to deliver app installs and maximize ROI
- Support app owners and publishers with an opt-in framework i.e. when the opt-in should be requested, what should be offered in exchange for opting in, the frequency with which the opt-in should be requested, how people can later opt-out, etc.
- Model how we expect different incentives to impact opt-ins, returning users and revenue generation.
- Help advertisers and app owners assess the impact on their current in-app advertising MarTech, including evaluating current ad networks and attribution platforms and devising plans on how to adapt and mitigate the impact of the upcoming changes.
- Support private equity funds exploring the space or companies with existing portfolio impacted by IDFA changes by assessing business impact and working together on mitigating initiatives.