In this episode of Unlocking Retail Media, Kevel CEO James Avery sits down with Melissa Burdick, co-founder and president of Pacvue, to discuss the evolution of commerce media, the shift toward incrementality-focused spending, and why API-first infrastructure is critical for retail media success.
Brands are becoming more cautious about where they spend their retail media dollars. After years of rapid growth and experimentation across hundreds of retail media networks, the conversation has shifted from "how fast can we scale?" to "is this actually incremental?"
In this episode of Unlocking Retail Media, Avery speaks with Burdick, who spent a decade at Amazon helping launch the consumables business and pioneering early advertising capabilities before co-founding Pacvue in 2017. Now operating across more than 100 retail platforms, Pacvue provides brands and agencies with an AI-powered commerce system to manage, optimize, and measure advertising across the fragmented retail media ecosystem. Here are the five biggest takeaways from their conversation.
Takeaway #1: Brands Are Demanding Proof of Incrementality, Not Just ROAS
"Two or three years ago, it was kind of like a non-negotiable. You'll be on retail media. Everyone's rushing into the retail media space. Then there is this backing up of like, whoa, is this incremental?" Melissa Burdick, Pacvue
The initial excitement around retail media has given way to scrutiny. Brands invested heavily across multiple networks, often shifting budgets from one retailer to another without generating genuine sales growth. While ROAS metrics within individual retailers may look strong, those numbers don't reveal whether advertising dollars are actually driving new purchases or simply redistributing existing demand.
Burdick explained that brands now use what she calls "IROAS" metrics to understand true incremental return on ad spend. "When you're in a silo and you're looking at one retailer and another retailer, their ROAS metrics are their own," she noted. "Your ROAS may look amazing, but actually it's not really incremental."
This shift reflects a broader maturation in how brands evaluate retail media performance. The focus has moved from demonstrating measurability to proving business impact. Retailers that can't clearly demonstrate incrementality risk losing share of wallet to networks that can.
Takeaway #2: The Three Pillars of Commerce Media Require Different Playbooks
Burdick outlined a framework for understanding how commerce media has expanded beyond traditional retail media networks. She identified three distinct pillars, each requiring its own strategy:
Discovery Commerce: Platforms like TikTok and Instagram where consumers discover products through influencer marketing and upper-funnel content. Brands need to understand that research suggests approximately 20 touchpoints occur before a purchase, and many of those begin on social platforms where people aren't yet ready to buy.
Retail Media: The hundreds of retailers offering sponsored products and advertising across their owned properties. This pillar has matured significantly, with brands now demanding better measurement, incrementality proof, and full-funnel capabilities that extend beyond basic sponsored product ads.
Agentic Commerce: The newest pillar, where consumers use ChatGPT and other large language models to research and potentially purchase products. "That's the trend I think that we know the least about," Burdick acknowledged. "The model continues to change, but that's where people are going to ChatGPT and LLMs to ask, what should I be buying?"
Each pillar requires different content strategies, measurement approaches, and technology integrations. Brands can't apply the same playbook across all three, and retailers need to understand where they fit within this broader ecosystem.
Takeaway #3: If You're Not Experimenting with AI Shopping Tools Now, You're Already Behind
Burdick pointed to Walmart's approach as a model for how retailers should think about agentic commerce. Two years ago, Walmart's CEO asked his executive team how they'd prepared for a meeting using AI tools. The team responded that AI was banned at Walmart. That policy was quickly reversed, and now every associate has access to AI tools.
"They created internally, like Marty, Sparky, they're experimenting," Burdick said. "They were one of the first ones to announce a relationship with ChatGPT. So they're out there really trying to figure it out and experiment."
Burdick personally uses Amazon's Rufus frequently, sharing an example of researching an adapter for her electric car that would work with Tesla chargers. "I asked Rufus, give me the Tesla adapter that works with my BMW car, and it gave me what I needed. That would've been, I would've had to read a product detail page and try to figure out, does this work with my car?"
While the model for agentic commerce continues to change, retailers who wait to experiment will find themselves years behind competitors who started early. "It's kind of like in that early stage of the cycle where it's not really in use yet, but they're figuring it out," Burdick explained. "If you're not starting to experiment and understanding it today, then you're gonna be definitely left behind."
Takeaway #4: API-First Infrastructure Is Non-Negotiable for Scale
"My big phrase that I say to every retailer is API first," Burdick emphasized. "It’s really what unlocks the ability to be fast, flexible, and be able to scale."
Pacvue operates as the single interface through which brands manage campaigns across more than 100 retail platforms. When retailers launch new features only in their user interface without exposing them through APIs, those capabilities effectively don't exist for the vast majority of brand spend flowing through platforms like Pacvue.
"What's challenging about that is, especially for a platform like Pacvue, is that we're working with all the brands and agencies and so they don't get access to the cool new stuff first if it's not pushed into the API," Burdick explained.
Measurement capabilities represent one of the biggest gaps. While Amazon has established a gold standard with Amazon Marketing Cloud, making data available in truly computable ways, most retailers lag far behind. "We have about a hundred variables that we always ask retailers for when we launch and, depending on who the retailer is, maybe we'll get 10%, 20%, 50% of the variables that we ask for," Burdick noted.
Retailers need to understand that API exposure directly correlates with advertiser spend. Brands using platforms like Pacvue simply can't activate features that aren't available programmatically, regardless of how sophisticated the UI experience may be.
Takeaway #5: Most Retailers Underestimate What It Takes to Drive Demand
"One of the biggest things is if you build it, they will come," Burdick said when asked what retailers underestimate about launching retail media networks. "People think, oh my, okay, it's here now. Where's all my demand? Where's all my sales?"
Building the infrastructure represents only the first step. Brands already have their budgets allocated, often locked into annual plans with established retailers. New networks face the challenge of proving their value, demonstrating incrementality, and competing for finite marketing dollars.
Burdick stressed the importance of partnership from day one: "Retailers should reach out to the experts in the space and interview them before they start a retail media network. They should talk to you [Kevel]. They should talk to Pacvue."
Success requires understanding the ecosystem, identifying the right technology partners, and having realistic expectations about timelines. "It actually takes a lot of work to bring that demand to the table," Burdick noted. Retailers that approach network building with humility and a willingness to learn from partners who've scaled other networks will move faster and avoid costly mistakes.
Integration across e-commerce and advertising also proves critical. "Retail media networks kind of launch, and they launch in a silo and they don't take into account any of the sales inventory, like the e-commerce components of the business," Burdick observed. The most successful networks tie promotions, coupons, and inventory data directly into advertising capabilities, creating a unified system rather than treating media as a separate revenue stream.
The Path Forward for Retail Media
Commerce media has matured from a land-grab phase into an era of accountability. Brands demand proof of incrementality, not just impressive ROAS numbers within siloed retailer dashboards. The expansion into discovery and agentic commerce means the touchpoints between initial product awareness and final purchase now span multiple platforms and technologies.
For retailers, the priorities are clear: build API-first infrastructure that enables programmatic access to all capabilities, experiment now with AI-powered shopping tools, and approach network building with realistic expectations about what it takes to attract sustainable demand. The retailers that protect their technology investments and evolve how they reach customers across all three pillars of commerce media will be positioned to capture share as the space continues to mature.
Partner with experts who've scaled networks before. Start experimenting with agentic commerce even though models are still changing. Make your data and capabilities accessible through APIs. Most importantly, recognize that simply launching a retail media network doesn't automatically generate demand. It takes strategic planning, strong partnerships, and a willingness to prove incremental value to brands that have more options than ever before.
Listen to the Full Conversation on Unlocking Retail Media
For more insights like these, tune in to the full episode of Unlocking Retail Media, the podcast where Kevel CEO James Avery sits down with industry leaders and innovators shaping the future of retail advertising.
Listen to this episode with Melissa Burdick here.
