The unreliability of third-party ad targeting signals spells opportunity for publishers offering their own first-party targeting tech.
Hearst has dramatically improved ad targeting across its publisher network since it launched AURA, an AI-based targeting solution, last June, Mike Nuzzo, SVP and head of data at Hearst, told AdExchanger. The company has seen improvements in addressability “between 30% and 200%,” depending on the audience segment being targeted, he said.
AURA is able to provide better addressability, Nuzzo said, by combining behavioral signals on readers’ browsing habits with contextual signals about the content they’re reading, as well as AI-powered audience modeling based on both of those signals.
Prior to launching AURA, Hearst was “confined to either contextual on its own or behavioral on its own,” he said. And behavioral signals derived from cookies or deterministic IDs on the open web represent “probably a 30% addressable market,” he said, hence the need to combine inputs and expand the audience graph using modeling.
Curating stability
The improvements in ad targeting are helping Hearst demonstrate to advertisers just how deep its audiences and content insights go, said Jennifer Dorre, SVP of ad product and data at Hearst. And they’re helping Hearst prove the old adage that publishers understand their content and audiences better than anyone else, she said.
Plus, AURA’s audience matching capabilities are making it easier for Hearst to court and retain ad budgets, Dorre said. Which helps keep Hearst’s ad business on a stable trajectory despite the traffic losses from generative AI search that every publisher is dealing with, she added. (She declined to offer specifics on how Hearst’s search traffic or ad revenue has been impacted by AI adoption.)
AURA has been particularly effective proving Hearst’s understanding of purchase intent signals, Nuzzo said. For example, it can analyze data derived from Hearst’s affiliate marketing content and guess which other Hearst readers across brands like Elle, Men’s Health and Country Living might purchase similar products. This modeling capability predicts audiences matches “almost twice as well as either the behavioral counterpart or the contextual counterpart on its own,” he said.
Because AURA is able to translate all that data in a way that actually impacts campaign performance, it’s more convincing than a publisher pitching the strength of its data without results to back it up, Nuzzo said. “People don’t respond to data – we are not ones and zeros,” he said. “They respond to relevance and resonance.”
What Amazon adds
Once AURA models a target audience, Hearst packages it into a curated deal. Until recently, AURA audiences could only be purchased directly from Hearst’s sales team. But Hearst’s partnership with Amazon Ads, announced in June, now lets advertisers buy AURA segments via Amazon DSP using deal IDs.
“We’re seeing more programmatic demand, and we want to meet those buyers where they’re already transacting,” Dorre said.
The partnership with Amazon Ads should also boost AURA’s ability to model audiences with high purchase intent, Nuzzo said, because it gives Hearst access to shopping data from Amazon’s ecommerce platform. These shopping signals update every 24 hours, so campaign targeting can be optimized daily, he said. And the audience matching is privacy protected by an encryption system provided by Amazon Publisher Cloud, he added.
The connection between the two platforms also helps Hearst surface less obvious overlaps between audiences. For example, he said, a pet food advertiser might want to target AURA’s “pet parents” audience. But there could be latent purchase intent signals around the “family travelers” audience, too, that could be surfaced by ecommerce data coming from Amazon.
Being able to demonstrate that Hearst’s purchase intent modeling matches well against Amazon’s sales data establishes “validation between the two parties, which we think lends credibility,” Nuzzo said.
Hearst isn’t giving Amazon DSP access to everything AURA has to offer, though.
Advertisers targeting AURA audiences via Amazon DSP will only be able to purchase ad inventory in standard IAB formats, Dorre said. However, Hearst also offers some proprietary, nonstandard ad units, such as its Hero Collection full-frame video takeover ads. This proprietary inventory is only available to advertisers that purchase AURA audiences directly from Hearst, she said.
This limitation in ad formats aside, Hearst is looking forward to deepening its relationship with Amazon as it explores new ways to use AI to optimize its ad offering, Nuzzo said. And Hearst is bullish about the technology enabling new ways for publishers to capitalize on their data.
“This ability to understand how AI helps find audiences that were otherwise invisible,” he said, “is going to bring increased enthusiasm to legacy publishers and how they think about their first-party data assets.”
