"Shoppers don't live inside these static audience labels," said Sean Black, general manager at Audience Path, during EMARKETER's Future of Digital summit, challenging the industry's reliance on traditional demographic targeting. The executive argued that understanding real-time consumer behavior, not outdated cookie data, is the key to effective audience targeting in retail media.

Most retail audience marketing "starts with false assumptions" about who consumers are and what they want, Black argued. He built that case on an analysis of 5 billion monthly video streams across 10,000 publishers.

Challenge what platforms actually know

Black urged marketers to question how advertising platforms build their audience segments and what data they use to train AI models.

"You can't just go into ChatGPT and type in an audience," said Black. "It's gonna come from Reddit. It's gonna come from open web data. It's not about real behavior."

Traditional audience labels, like "deal conscious tech operator" or age-based segments like "18 to 24," rely on cookies that could be six months or a year old. These static identifiers don't capture what consumers are actually doing today or what their current intent is, according to Black.

Marketers should make partners prove their audience quality by answering two questions: What signals are you capturing, and how did you identify the audience's actual intent?

Focus on behavioral patterns

Tracking consumption patterns, and how their frequency changes over time, is how Black proposes to understand shopper intent. He used his own family as the example: He, his wife, and his mother all had the same intent when the couple was expecting twins, but their behavior diverged.

"The frequency in which we were looking at the content is what helped define who we were," he said. His wife consumed baby-related content constantly, while he searched for parenting guidance less frequently. Despite the same end goal, their content consumption patterns revealed different roles and needs.

The approach analyzes what people are watching, reading, and engaging with in real time through video streaming behavior. This creates what the company calls "mirrors," behavioral profiles built from actual consumption patterns rather than cookie-based assumptions.

Audience members with the same intent reach it through different paths. Understanding these behavioral differences allows for better creative messaging, improved channel selection, and more effective scale.

Behavioral signals complement existing data

Behavioral targeting should expand existing strategies rather than replace first-party or commerce data, Black emphasized.

The recommended approach stacks multiple signals: streaming behavior, search patterns, social engagement, and demographic information. This creates a more complete picture of consumer intent.

"What are you training models on?" Black asked, urging marketers to scrutinize partners who claim audience knowledge based solely on supply or first-party data.

Marketers should demand "humanized data" from partners: behavioral insights that reveal consumption patterns and intent signals rather than another AI-generated option.

Real-world application: King's Hawaiian case study

A King's Hawaiian campaign targeted consumers whose behavioral signals indicated last-minute Thanksgiving shopping needs, according to results Black shared.

The day before Thanksgiving, the campaign ran creative that let viewers select products and add them directly to Instacart or Amazon shopping carts for same-day delivery. That execution showed how understanding behavioral patterns can drive immediate purchase action beyond what demographic targeting delivers.

The campaign achieved a 57-second average engagement time, according to Black.

Behavioral targeting lets marketers time messages precisely and match creative to where consumers are in their purchase journey.

Watch the full session.