‍The retail media opportunity is real. So is the risk of squandering it. Networks that launched fast are now facing advertisers asking harder questions about incrementality, proof, and whether the numbers actually hold up. The 2026 Media Outlook series opens exactly there: not with predictions, but with a conversation about what separates networks that are building something durable from those that are busy looking busy.

The discussion brought together three practitioners with distinctly different vantage points. Leora Kelman, Managing Director and Partner at Boston Consulting Group, who leads BCG's retail media practice, offered the macro-strategic view. Danilo Tauro, Co-Founder of CartographAI, brought the practitioner and investor lens. At the center was Nagarajan Chakravarthy, Chief Digital Officer at iOPEX Technologies, whose work spans digital transformation across retail, media, telco, and high-tech, and who has guided organizations through every phase of the journey, from crawl to walk to run.

Three perspectives. One question underneath all of them: what does it actually take to build a retail media network that lasts?

Stop Predicting. Start Scenario Planning.

Retail media is growing. WPP estimates it will overtake TV ad revenues in 2025, reaching roughly $180 billion globally. But Leora was quick to reframe what that number actually means for the people running networks today: growth doesn't eliminate uncertainty, it amplifies the cost of being wrong about it.

Her prescription was direct: move away from point predictions entirely. 

"We are really bad at predicting consumer behavior in normal times," she said. "Especially in a moment where there is so much technological change, so much consumer behavior change." 

What BCG is pushing clients toward instead is scenario planning — mapping out multiple futures, identifying no-regret moves that hold across all of them, and tracking the signals that tell you which scenario is actually unfolding. Over 60% of consumers have already used AI for shopping in some capacity, she noted. 

Danilo reinforced the point from a market structure perspective. Incremental retail media dollars are flowing disproportionately to large, technically sophisticated players. For everyone else, the old playbook of activating sponsored listings and watching revenue print is no longer sufficient. The space is bifurcating, and the gap is widening.

3 Signals You Can No Longer Ignore

Nagarajan traced the arc of retail evolution before landing on what he sees as the defining shift of this moment: hybrid commerce, where human shoppers and AI agents coexist in the same environment, and networks have to be ready for both.

From that, he identified three signals worth watching closely. 

First, infrastructure readiness: whether a human or an AI agent is doing the shopping, the experience must be seamless and simplified. Networks that build for one and ignore the other are architecting their own irrelevance.

Second, incrementality. ROAS alone is no longer sufficient. "What is very, very important is the margin," Nagarajan said. Networks need to ask which specific categories and SKUs are driving real incremental value, not just spending budget, particularly as agentic commerce changes what gets surfaced and why.

Third, proof-of-concept velocity. As advertisers grow more sophisticated, they're behaving more like investors, demanding evidence before scaling commitment. Leora added her own signal to this: she tracks website traffic across the top 10 retail sites, watching for the moment traffic starts ticking down as an early indicator that agentic shopping has reached genuine scale. "When that traffic starts ticking down," she said, "we know a real thing."

Big Rocks, Pebbles, and Sand

On building a differentiated demand strategy, Leora introduced a framework borrowed from product development that cuts through the complexity: big rocks, pebbles, and sand.

Big rocks are strategic suppliers — the brands you have a sales relationship with, regardless. What they want from non-top-tier networks is simplicity: the ability to activate campaigns in their own seats, through their agencies, via API. 

Pebbles are smaller, less sophisticated advertisers for whom the network is already their primary sales channel. They don't have agency support. The tooling needs to live inside the merchant portal, almost invisible in its ease, a prompt that says your SKU is underperforming, here's one button to fix it. 

Sand is programmatic aggregation: partnerships with ad tech platforms that unlock broader demand. "Not a super-specific ad server that doesn't talk to anyone else," Leora said pointedly, "but really making sure those API connections are open."

Nagarajan anchored the framework in execution: every retailer has a strength, a loyalty base, a captive audience. The winning formula starts by compounding that strength, not replicating what Amazon built, but building toward what only you can offer.

Fix the Foundation Before You Chase the Future

Nagarajan shared an account that will resonate with anyone who has watched a promising retail media launch hit its first wall. A network had gone live quickly, with third-party ad servers, agency support, and strong early momentum. Five or six months in, brands started asking harder questions about incrementality. The numbers didn't hold up.

The recovery required going back to basics: triage, stabilization, and rebuilding from the ground up. Advertiser briefs. Media planning. Mid and post-campaign analysis. Converting one-time budgets into evergreen commitments. Only once the foundation was solid did the team focus on what was working for top suppliers and use those results to educate the broader base.

The AI innovation came last, and deliberately so. The team vectorized campaign data and built a conversational agentic interface that sales teams could query directly before supplier meetings. The lesson from that sequence was the most quotable of the session: 

"You can't just jump with Agentic." 

AI accelerates what is already working. It cannot substitute for the work that hasn't been done yet.

Closing the Loop: Where Retail Media Wins the Bigger Budget

Danilo brought the conversation back to what he called the marketer's holy grail: closing the loop between media viewership and purchase data. Upper funnel and lower funnel have historically operated as separate worlds — different teams, different KPIs, different budgets, different technology. Retail media, he argued, is uniquely positioned to bridge that gap.

Amazon did it by connecting retail signals to its own CTV properties — Prime Video, Fire TV, Thursday Night Football. Walmart acquired Vizio. Allegro in Central Europe is partnering with regional publishers. Mercado Libre is working with Netflix. 

The pattern is consistent: the networks winning incremental brand dollars are the ones proving full-funnel influence, not just lower-funnel conversion.

Leora added a dimension that sharpens the opportunity further: creators. "If you could mix creators with the first-party data of retail to both better target and close the loop, now I have a much more powerful narrative channel, powered and proven by incredibly powerful first-party data." That combination, she argued, is the kind of premium inventory that moves brand budgets at scale.

Why This Matters Now

Retail media carries a structural advantage most channels simply cannot replicate: credible data, real buying behavior, and measurable influence at the point of commerce. What this conversation made clear is that the advantage is not automatic. As Danilo put it in closing: the winners are the ones who plan for multiple futures, pick the right priorities, and treat retail media as a business transformation — not a tech project.

Watch Episode 1 of the 2026 Media Outlook for the full conversation, the frameworks, and the strategic signals shaping what comes next.