From targeting to prediction: digital advertising’s paradigm shift

Personally, I’m in two minds. There’s a difference between how I feel about my privacy, and what I do to preserve it. Perhaps it’s one of those things that I care about very deeply, but I’d like someone else to take care of it, please. And maybe that’s why brands stand to gain so much by showing that they understand customers’ privacy and that they’re taking care of it. Acting fast and loose with an intimately detailed record of someone’s preferences and behaviors is no way to build trust and loyal relationships.

Differentiation and Apple’s lucky halo effect

Google and Ipsos collaborated last year, to demonstrate that when a brand offers a good privacy experience, it can expect to see an uptick in brand preference, trust, and advertising effectiveness. That research was a long time coming, with cleverly measured findings that tell us the bleedingly obvious. The advice is: to offer users meaningful choices about their data, show that you remember those choices, and let them review them. Doing that helps users feel good about you. No kidding.

We care about privacy, but often we don’t do anything about it. We never think before we tap ‘accept’. You might wonder how no one doing anything caused such a massive change in the way the advertising industry treats data – the industry is undergoing an enormous course correction that few predicted. It seems like only yesterday that we were pitching digital attribution as the pinnacle of measurement.

Now, the pitch goes something like, “Get over it. We can’t track people reliably anymore.” Most of the work being done in digital analytics today is about either making the most of a subset of data from the users we can still track accurately (especially identifiable customers), or measuring advertising and making it more effective without any personal data at all.

It hasn’t happened because consumers took action. It’s happened, I suggest, because ten years ago, in 2013, when Edward Snowden’s 41 PowerPoint slides revealed the Prism program, Apple barely appeared anywhere on it. What a gift: a differentiated brand position, sitting on a plate. By 2019, the brand’s messaging was ‘Privacy. That’s iPhone,’ and today we have an industry scrabbling to find value in our new favorite type of data, first-party. The only thing that consumers have done to contribute is to carry on buying iPhones.

Digital advertising’s two fundamental shifts

Before Apple’s Intelligent Tracking Prevention (ITP), we followed cookies as they appeared across websites and built audiences defined by big lists of them so that we could show ads in browsers that were probably being used by the people who were probably at least slightly inclined toward clients’ brands. Now, two big things are different.

First, we largely don’t target audiences anymore. Instead, we rely on machine-learning-powered technology in ad platforms to bid to advertise to increasingly relevant people. These machine learning models are trained to drive after a goal; success is only as good as the data that represents that goal. Instead of investing in third-party marketing audiences, we’re analyzing the customer records that represent lasting success for the advertiser, and using that to predict the future value of new acquisitions. That accurate prediction of lasting business value, mined from first-party data, is how we’re driving better results.

Using a smaller volume of personal data to predict a better measure of success is proving more valuable than trying to track everyone and guess who to advertise to. Hindsight is a wonderful thing.

The second big thing, and no less a paradigm shift, is that we have become increasingly focused on personally identifiable information (PII): real names, email addresses, and phone numbers. This sounds less private than pseudonymous cookie values, but the key distinction is that we can never know anyone’s PII unless they’ve actively given it to us. It’s another shift of focus, toward data belonging to customers we know, and away from data belonging to users, we have to guess about.

Ad platforms and publisher networks know the same people, and here’s the trick: both the advertiser and the platform have matching records of the same encrypted email addresses. By matching them, we can know that the customer we acquired is the same person who saw our ad on the publisher network’s website. Not a cookie in sight. Apple’s ITP has become so defensive that we’re on a route to giving up on cookies altogether – third-party, first-party: the lot.

I wonder whether we’ll stop there. Digital advertising and its capabilities seem to blend, slowly but surely, on- and offline, into a single big business we might as well just call marketing. As the power of AI to predict continues to develop, every advertising strategy becomes full-funnel (paid search included).

As the reality of user tracking has focused us on the face of the customer, not the faceless user, we should be getting better at thinking about who these people really are, and how to engage them.

To read more from The Drum’s latest Deep Dive, where we’ll be demystifying data & privacy for marketers in 2023, head over to our dedicated hub.