Accommodating the demands of consumer privacy and data-driven ad targeting can feel like the Heisenberg uncertainty principle of the marketing world.
The more attuned you are to one, the less attuned you can be to the other.
Comscore wants to turn that notion on its head with the launch on Wednesday of its new Data Partner Network, an AI-based initiative that allows data providers to convert ID-based data sets into ID-free audience segments, which might then be used to retarget desired customers.
If it looks like a duck …
More than a dozen data providers, including AnalyticsIQ, Polk Automotive Solutions and TransUnion, integrate their data with Comscore and the new Data Partner Network. From there, Comscore’s programmatic subsidiary, which is called Proximic, turns that ID-based data into a predictive model for ID-less audiences, according to Rachel Gantz, managing director of Proximic by Comscore.
Polk, for instance, might have an audience of Toyota Corolla owners that Proximic’s AI-powered contextual engine would analyze to determine which topics that audience has “the highest propensity” to be interested in, like cruise vacations or gardening, Gantz told AdExchanger.
The process mirrors audience targeting, she said, but without identifiers. Rather, the vendor matches audience profiles from the data sets of other companies in its partner network to Comscore’s “opt-in panels,” which are anonymized records the company uses to track behavioral patterns (like, say, a demonstrated interest in cruises or cooking.
“We’re using AI to ensure that [each ID-free data set] acts like an audience segment,” she said, just without any sort of user identifiers.
Partnering with Comscore has been a “compelling narrative for how buyers can unlock the full potential” of emerging ID-free audiences,” Margo Hock, VP of digital partnerships at AnalyticsIQ, told AdExchanger.
Since launching earlier this year, Hock said more than 250 of AnalyticsIQ’s marketer clients have implemented Comscore’s new predictive audiences. A mix of new and returning advertisers have elected to use these audience segments, she added, which AnalyticsIQ sees as “a strong signal of positive performance.”
Scaling up
Working with Comscore has made it much easier for programmatic media partner MiQ to reach more people, said Ciaran Slattery, SVP of trading at MiQ.
Without sufficient scalable privacy solutions, Slattery said, the ability to optimize and retarget audiences across a wide array of inventory is “severely limited.” A lot of the data MiQ worked with in the past was based on IDs or cookies, but contextual targeting posed more of a challenge.
Targeting a domain list, Slattery said, “needs a lot of human update to consistently repeat and refine,” he said, whereas AI can refresh domain lists and find new, relevant segments in a fraction of the time.
Contextual advertising, he added, has long-term effects. Rather than “just [garnering] a few extra points of ROAS or a slightly lower CPA,” he said, it drives conversions and confirms that its ads are reaching the right people.
Recently, MiQ worked with a local tourism board to target travelers looking for higher-end experiences, like expensive hotels and restaurants. But “traditional segments” for travelers or high earners aren’t up to date, Slattery said – especially since many people (particularly those with higher incomes) may travel on a whim. By using Comscore’s segments and targeting correlated interests, like luxury vehicles, he said MiQ could extend that reach to audiences it would otherwise have missed or struggled to reach.
MiQ found Comscore’s audiences particularly useful in the context of CTV, where the naming conventions for various segments and content topics are “not universal,” Slattery said.
The Proximic AI-based solution can find new ways and places to reach relevant audiences, he said, by ingesting mountains of bidstream signals and normalizing data from across different data sets that otherwise might not sync up.
For an actual person do this, he said, would be “almost impossible.”
