The Storefront is Changing. Product Data is Taking Centre Stage
The brands getting ahead are treating their product data as a strategic asset instead of a back-office function and building the infrastructure now to compete in an agent-driven world, says Lori Schafer, CEO of Digital Wave Technology.
Topics
Retailers competed by perfecting the digital storefront for years. Better product pages, stronger merchandising, richer imagery and sharper pricing strategies became the foundation of ecommerce success.
That model is beginning to change.
As AI agents increasingly influence how products are discovered, compared and purchased, brands are no longer optimising solely for human shoppers. They are preparing for software that can evaluate products, compare attributes, verify claims and make buying decisions without ever visiting a traditional product page.
The shift goes far beyond another marketplace strategy. It changes what determines visibility, how merchandising works and where competitive advantage is created. In this environment, structured, governed product data becomes just as important as branding, pricing or customer experience.
For Lori Schafer, CEO of Digital Wave Technology and a veteran of more than three decades in retail analytics, AI and digital commerce, this marks one of the biggest structural shifts the industry has faced since the rise of mobile commerce.
In this interview, Schafer discusses why agentic commerce is reshaping retail, why product data has become a strategic asset, how digital merchandising is evolving beyond human shoppers, and why regional marketplace challenges increasingly begin with the same underlying data problem.
Excerpts from the interview;
You predicted the impact of mobile, analytics, and social commerce in your book, Branded! Are marketplaces and agentic commerce driving a similar shift, and what are brands still underestimating?
Yes, and arguably a bigger one. Mobile, analytics, and social shifted where and how consumers shopped. Marketplaces and agentic commerce shift who is shopping. When an AI agent is the buyer, the brand loses its traditional storefront, its merchandising voice, and even its pricing leverage.
What most brands still underestimate is that this isn’t another channel to manage; it’s a disintermediation of the brand-consumer relationship itself. The consumer may never see your product page again. An agent will see it first and filter on your behalf.
Brands are staffing for this like it’s ecommerce 2.0, adding headcount to manage marketplace listings, when the real imperative is rethinking the operating model: the data structures, the governance workflows, the way content is created and maintained.
The brands getting ahead are treating their product data as a strategic asset instead of a back-office function and building the infrastructure now to compete in an agent-driven world.
You shouldn’t be asking “How do we rank better on Amazon?” The right question is, “How does our product data perform when a machine is the buyer?”
After years of advising global retailers, where do you see the biggest operational challenge today: marketplaces or the data infrastructure behind them?
The data infrastructure underneath, without question, and I’ve seen this pattern repeat across every major technology transition for thirty years. The marketplace looks broken. Products rank poorly, content is inconsistent across channels, pricing is out of sync, and inventory signals are wrong.
But when you dig into why, you almost always end up in the same place: the data layer can’t keep pace with the speed, granularity, and syndication requirements of modern commerce.
Most brands are running marketplace operations on top of PIM, ERP, and merchandising systems designed for a single channel, a slower cadence, and a human audience.
The internal alignment problem is just as real as the technology problem. In most organisations, the teams responsible for product data — merchandising, IT, digital, marketing — don’t share ownership of it.
Merchandisers maintain data in one system, digital teams in another, and nobody has a complete, governed single source of truth. When the marketplace demands an attribute that doesn’t exist anywhere in your infrastructure, the process to create it, validate it, and syndicate it can take weeks, if it happens at all.
The marketplace isn’t your data problem. It’s your data problem’s report card. Every syndication failure, every ranking drop, every compliance flag is the marketplace surfacing a structural weakness that existed long before your first product listing.
You’ve spent years in retail analytics and AI. How is the rise of agentic AI changing the definition of “good product data” compared to the search-and-click ecommerce era?
In the search-and-click era, “good” meant discoverable and persuasive to a human, such as keywords, hero imagery, and lifestyle copy.
For agents, “good” means machine-interpretable, verifiable, and contextually complete. Agents don’t browse, they reason. They need structured attributes, consistent taxonomies, provenance, claims that can be validated, and relationships between products.
The definition has been completely inverted, and most organisations haven’t caught up to what that means operationally.
Marketing-driven copy that’s vague or aspirational actively hurts you because an agent will deprioritise or misclassify a product whose attributes are incomplete or ambiguous. It has no patience for “premium quality” if it can’t verify what that means.
What this looks like in practice:
- Does every size carry the right weight attribute?
- Are ingredients listed in a format a regulatory agent can parse?
- Is your taxonomy consistent enough that a recommendation engine can reliably group and compare products across your catalogue?
Those are the questions that determine whether your products are visible or invisible in an agent-mediated commerce environment. Good data is now an algorithmic asset, not a creative one.
Retailers once optimised product content for human shoppers. As AI agents increasingly influence discovery, how does that change digital merchandising?
It flips the discipline from persuasion to precision. Human merchandising rewarded storytelling, emotion, and visual hierarchy. Agent-driven merchandising rewards completeness, accuracy, attribute richness, and semantic clarity. The “shelf” becomes an API surface.
Merchandisers will spend less time curating banners and more time governing taxonomies, training models, managing attribute coverage, and ensuring their product data competes well inside ranking and recommendation systems they don’t control. The new merchandiser is part data steward, part AI strategist.
The process changes are significant, too. Content workflows built around human review cycles — someone writes copy, someone approves imagery, someone publishes to the channel — aren’t fast enough for agent-grade data requirements.
You need automated validation, real-time completeness scoring, and feedback loops that surface gaps before they affect ranking or compliance.
We’re already seeing retailers create new roles that sit at the intersection of merchandising, data governance, and AI operations. That’s the right instinct. It’s important not to treat this as a technology problem alone, or it will struggle. Instead, treat it as an organisational capability to have a lasting advantage.
Are global brands facing the same marketplace data challenges everywhere, or are regional commerce ecosystems evolving differently?
The core challenge — fragmented, incomplete, inconsistent product data — is universal. But the shape of the problem varies significantly, and global brands tend to underestimate that until it costs them.
- North America is marketplace-dominant and Amazon-centric. The pressure is on syndication velocity and competitive pricing. Amazon’s standards for attribute completeness and content quality have become the industry baseline. Brands that can’t keep pace get buried.
- Europe is regulation-heavy — GPSR, Digital Product Passports, ESG disclosures. The compliance burden falls squarely on product data infrastructure. It’s not enough to have a well-described product; you need provenance, certifications, supply chain traceability, and claims that can withstand regulatory scrutiny. Brands without governed master data have real exposure.
- APAC, especially China and Southeast Asia, is livestream-and-super-app driven, requiring rich media and real-time updates at a cadence Western brands aren’t built for. The speed requirement alone breaks traditional PIM workflows.
- Latin America is a hybrid story: marketplace consolidation around players like Mercado Libre, combined with strong omnichannel and department store traditions in markets like Mexico, where retailers such as El Palacio de Hierro are balancing luxury brand storytelling with marketplace-grade data discipline.
Cross-border tax complexity, multi-currency pricing, and Spanish and Portuguese localisation add a layer that most global PIM strategies underestimate. It’s a region where getting the data foundation right is a genuine competitive differentiator.
- Emerging markets often leapfrog directly to mobile-first marketplace models, which paradoxically creates an opening for brands with clean, structured data — because the baseline is lower and the advantage of getting it right shows up faster. Same disease, different symptoms, and in some cases, a faster path to the cure.
ALSO READ: Beyond Clicks: The Shift Towards Outcome-Based Advertising
