How using AI to create predictive insights can give publishers an edge

The way advertisers can use data is changing thanks to stricter data-protection laws and coming changes to third-party data sharing. We spoke with Jürgen Galler, co-founder and CEO at data management platform 1plusX, about how the use of AI by publishers can help them traverse these challenges.

Why should AI-powered predictive insights be a key consideration for publishers as we move away from third-party cookies?

If you really want to create value out of raw data, it's not just about what a user tells you but also what you observe—the interactions, behaviors, trends, and so on. Machine learning allows publishers to create value out of their data. Value that is usable for content personalization but also for ad targeting.

The future is going to be much more complicated because you're actually going to have a huge variety of potential identifiers: classic, first-party cookies; some third-party cookies, supported by some browsers; mobile identifiers; log-in identifiers. In addition, you're also going to have contextual identifiers, such as which article somebody is reading or which video they’re currently watching.

What AI does is make this decision [who to target] an adapted one, a process that we call adaptive tracking and adaptive targeting. It goes beyond just creating an audience for users that you have an email for, or creating an audience for users that you want to actually reach contextually.

Even with a rich tapestry of third-party data, digital advertising hasn’t been great at predicting what we want. Why should things improve once all that third-party data goes away?

The decline in third-party cookies has been something that has grown over the years. Chrome is obviously the market leader, and Google is now saying it’s going to switch it off in 18 months. But the reality is that this has been going on, step by step, for the past several years.

By using an adaptive mix of contextual targeting and first-party user targeting, you increase reach and you increase quality.

The fact is, third-party data is very often bad data. It's been collected maybe months ago, has been mixed up and shared, and is just bad quality. If you use first-party data in real time and you add to it with predictions, embedding contextual signals and so on, then you're much closer to the user.

If consumers knew what was going on behind the scenes, would they be comfortable knowing their data was being used in this way?

I think so.

If I'm booking a hotel somewhere, and I haven’t finished the booking cycle, but then I go and read something online somewhere else, and then I see this hotel again, that's annoying. That's all third-party data sharing.

But if I'm a Spiegel or Telegraph reader and I give them the right to use my data, that's because I like them and I have a better experience if I go there. I get more accurate recommendations for the content because they know that I like surfing, Italian food, and traveling to Asia. That's worth something.

The next step is for media companies to give consumers more control over their data.

In the future, maybe we could have a framework where you could have a dashboard where users can set approvals and disapprovals, centrally, so that they are in control. They’ll also then realize how their data is being used and where it’s being used. That’s the right direction, and in the end, it should feel better for consumers.

Have these AI-based solutions got better traction in places like Europe, where privacy and data-sharing laws are a little ahead of the curve?

Absolutely, though it still took time.

When we saw GDPR [the General Data Protection Regulation] coming, we implemented a lot of relevant features, and we thought everybody was going to buy our platform. But they didn’t. Everybody waited.

Many businesses saw that the typical third-party data they were using had gone, so they retreated from the market. However, they quickly realized they needed new ways to create this level of targeting data so they could serve their clients. It was a commercial motivation for many.

Then, the big brands realized that they only wanted to work with media companies that were doing super correct GDPR-style data processing. So, the media companies then had to prove that they were super clean. Now, we’re getting calls from media companies with global user bases that want to be GDPR and CCPA [California Consumer Privacy Act] compliant.

Other countries, meanwhile, have their specific privacy rulings, and there are going to be more and more coming, so it’s becoming increasingly important to have a very strong and well-customized data setup to be compliant.

"Behind the Numbers" Podcast