With hardware updates not making headlines like they used to, Apple has spent the last year attempting to dominate tech news, instead, with announcements of a range of new privacy protection features and policies. The most significant of these is without a doubt App Tracking Transparency (or ATT for short), a part of the iOS 14.5 update.
Receiving an updated dashboard in iOS 15, ATT appears to the user as a simple pop-up message whenever a mobile app is opened for the first time. In the message, users are asked if they would like to allow the company behind the app in question to track their activity and access their data. At no additional cost to them, the user can say 'no.' This represents a groundbreaking change in Apple policy and - unsurprisingly - users are opting out of tracking in droves.
Globally, as many as 90% of iPhone users have already chosen to disallow apps from collecting and storing their data.
While this has been hailed as great news by privacy activists, companies in many data-driven industries have begun to express concern. Perhaps the loudest outcry has been heard from digital marketers and the platforms on which they rely.
Each year, social media platforms like Facebook, Twitter, and TikTok generate more than 150 billion dollars of revenue from advertising. Most of this success owes itself to the data these companies collect from users.
With easy access to data over the last few years, a business model has arisen in the digital space that involves the application of AI and machine learning to the personal information of users. Social media platforms package these technologies as easy-to-use advertising tools, usually sold using a cost-per-impression or cost-per-click model. Businesses and digital marketers use these tools to create highly-tailored advertisements which target specific segments of the platforms' user bases.
Given that iOS users make up a huge chunk of social media's user base, one might wonder: Will large digital platforms still be able to sell their advertising products going into the future?
As the saying goes, “necessity is the mother of invention,” and several major platforms have already announced plans to get around the new restrictions placed on them by the ATT feature. The most notable of these workarounds are being implemented by market leader Facebook, which generated about 84.2 billion dollars in advertising revenue in 2020.
Facebook's approach is to 'depersonalize' all data that they collect using encryption. Basically, after being processed by the company's so-called "Privacy-Enhancing Technologies," the data remains analyzable by AI but becomes untraceable to any particular user. This, Facebook hopes, is in compliance with both GDPR and Apple policy.
Although this solution has only recently been implemented, it does seem to have a good shot at keeping AI/ML technologies at play in the digital marketing space. As of the time of writing, Facebook's solution still has yet to be shut down by Apple. The same can't be said about TikTok's somewhat nefarious attempts to track users by 'fingerprinting' their devices...
So far, a few key trends have emerged, providing a decent glimpse into how things could look for advertising and AI in the coming years.
First, there can already be observed a clear shift among marketers away from quantitative metrics and towards qualitative ones. Ads are no longer being created just to target specific demographics, but instead to have greater mass appeal. As such, companies are not focusing on the micro-level, but on the macro — on things like overall customer satisfaction and gross sales. Of course, AI will continue to be applied to this data to help make informed decisions about campaigns.
Second, it is now clear that with fewer ways to target customers on a personal level, digital marketers will need to experiment with other variables. One example of a non-personal variable that can be analyzed is the number of conversions per ad against the time that the ad had been published. Digital marketers may want to experiment with automating the publication of ads, so that they can be put out at the ideal moment. Neural networks that can predict the best possible publication times are set to become a go-to tool. Other 'depersonalized' variables could be things like ROI, ad spend, and the number of likes on a post.
Finally, digital marketers are already applying the lessons they learned over the last few years of AI-powered predictive advertising to tried and tested marketing strategies like gated offers, loyalty programs, and email marketing. A very simple example of such a campaign would be to put out an ad offering a coupon code in exchange for filling out a survey and leaving an email address. This, in effect, represents an 'opt-in' on the part of the user. Any data that is willingly shared can be used without qualms.
If one thing seems certain today, it's that privacy stands to be one of the chief driving forces behind the next era of digital marketing. Yes, at times it may seem a pesky restriction, but ultimately it is only through overcoming obstacles that innovation can be achieved. We know this from experience.