Has commerce media’s meteoric rise already outpaced our ability to measure it?
That’s the unsettling question lurking beneath the surface of the industry’s breakneck growth. Caroline Giegerich, vice president of artificial intelligence at the IAB, points out a stark reality: this sector, poised to hit a staggering $105 billion by 2025, is expanding so rapidly that our established measurement tools—the humble click, the ubiquitous impression, the ever-present ROAS—are increasingly obsolete. They’re like ancient mariners trying to chart a course through hyperspace.
The AI Middle Manager Emerges
Giegerich paints a compelling picture of artificial intelligence not as a futuristic takeover, but as a sophisticated, even “mysterious,” orchestration layer. Think of it as a hyper-competent middle manager, capable of juggling inventory, pricing, customer behavior, and media buying with an efficiency that would make human armies of analysts weep into their spreadsheets. This isn’t just about automating tasks; it’s about asking fundamentally smarter questions. Are campaigns actually driving new sales, or just nudging loyal customers towards their usual sock purchase? Are margins protected? These are the business outcomes AI can finally illuminate, empowering the humans still guiding the ship.
“The challenge is that the measurement has not kept step with the growth in the business,” she said in an interview with Beet.TV editorial director Lisa Granatstein at the IAB Connected Commerce Summit.
The Shifting Landscape: Beyond Owned Properties
And it’s not just about what we measure, but where. For years, brands obsessed over their carefully curated digital gardens—websites, apps, social feeds. Now, Giegerich highlights a significant gravity shift. The real influence, she argues, lies in places brands can’t directly control: creator content, earned media, even—and this is a wild one—a marketer recently discussing a Wikipedia page as a key touchpoint. Why? Because AI systems may accord more weight to these seemingly external endorsements than to a brand’s own polished self-promotion. It’s a humbling realization that our meticulously crafted narratives might be losing ground to machine-verified authenticity.
This seismic shift necessitates a profound architectural change in how we construct information.
“We used to build things for humans to consume. Now we need to build things for machines to understand,” Giegerich emphasizes. This isn’t a minor tweak; it’s a complete re-engineering of data. If product data isn’t structured for machine comprehension, those AI agents hunting for Nikes will simply walk on by.
The Rise of the Unsupervised AI Shopper
Looking ahead, Giegerich foresees a future where AI doesn’t just assist shopping, but executes it autonomously. Imagine instructing your AI agent to find size eight red Nikes for under $200, and then stepping away while it navigates the complex digital marketplace on your behalf. This autonomous future, however, demands extreme caution.
