Retail Media’s Next Test: Turning Siloed Data Into Shared Wins
Floor decals and shelf talkers don’t get much respect in the era of programmatic advertising. But they were retail media before anyone called it that – and understanding that lineage matters for grasping where the industry is headed next.
Retail media has traveled a long road from in-store signage to what are now, in some cases, barely distinguishable from full-scale media platforms. The networks built by Walmart and Target have grown sophisticated enough to compete for national brand budgets – but the sector still faces stubborn structural problems around measurement, identity, and data fragmentation that no amount of AI enthusiasm will paper over, according to one industry veteran.
“You’re seeing these retailers create truly transformative media networks with really, really sophisticated capabilities,” said Claire Wyatt, GM, BD and retail sector lead at The Trade Desk, in this video interview with Beet.TV. But she was equally candid about the gaps that remain.
From floor decals to full-stack platforms
Wyatt, who spent time at Albertsons earlier in her career, traced retail media’s origins back to physical stores – the branded end-caps, shelf talkers, and floor graphics that manufacturers paid for long before a pixel was ever served. The digital pivot came when e-commerce gave retailers a new inventory to monetize, and someone realized ad space on a product page was worth something.
She credited the team at Walmart’s Triad unit – the in-house group that managed supplier marketing programs – as “probably the pioneers of maybe what we would know of retail media today.” From that starting point, the category has ballooned. Global retail media ad spending is projected to reach $100 billion by 2027, up from an estimated $75 billion in 2025, according to eMarketer forecasts, as retailers increasingly package their first-party data alongside ad inventory.
Wyatt pointed to Walmart Connect and Target’s Roundel as benchmarks for what the category can become – networks that have built genuine media infrastructure, not just bolted an ad server onto a shopping site.
The measurement problem
For all the sector’s growth, Wyatt described measurement as a persistent frustration – one she has carried through most of her career. The core value proposition of retail media, she argued, is the retailer’s first-party data: both for audience targeting and for closing the loop between an ad impression and an actual purchase. The trouble is that the industry hasn’t agreed on what “closing the loop” actually means in practice.
“Despite the fact that retailers have access to all of this great data, there’s still such intense discussion in the industry about what do we do with that data and how does it actually drive outcomes,” she said. “Can we agree, and then can we agree across retailers if those outcomes are the same?”
The debate over incrementality versus return on ad spend cuts to the heart of it. Wyatt argued that basic ROAS figures, while useful as a rough business metric, don’t answer the more important question of whether advertising is actually moving the needle. “When you’re really thinking about is my media adding value, it needs to be an incremental ROAS number,” she said. Progress is being made, she acknowledged, but “not as fast as everybody would like.”
Garbage in, garbage out – even with AI
The question of what separates winning retail media networks from also-rans drew a pointed answer. Scale alone, Wyatt suggested, only gets the largest players so far – and for everyone else, a generic pitch about retailer brand equity isn’t enough.
“It’s not enough to say you should spend with me because I’m XYZ retailer,” she said. The networks pulling ahead are those that translate the actual differentiators of their retail business into their media product, creating coherent value on both sides.
On AI – which she noted was unavoidable at industry events, suggesting a “shot game” at conferences would leave attendees unable to walk out – Wyatt is pragmatic. “It’s very much garbage in, garbage out,” she said. “If you haven’t done the work to create a really solid data foundation, implementing any kind of AI efficiencies is going to be really, really hard.”
Floor decals and shelf talkers don’t get much respect in the era of programmatic advertising. But they were retail media before anyone called it that – and understanding that lineage matters for grasping where the industry is headed next.
Retail media has traveled a long road from in-store signage to what are now, in some cases, barely distinguishable from full-scale media platforms. The networks built by Walmart and Target have grown sophisticated enough to compete for national brand budgets – but the sector still faces stubborn structural problems around measurement, identity, and data fragmentation that no amount of AI enthusiasm will paper over, according to one industry veteran.
“You’re seeing these retailers create truly transformative media networks with really, really sophisticated capabilities,” said Claire Wyatt, GM, BD and retail sector lead at The Trade Desk, in this video interview with Beet.TV. But she was equally candid about the gaps that remain.
From floor decals to full-stack platforms
Wyatt, who spent time at Albertsons earlier in her career, traced retail media’s origins back to physical stores – the branded end-caps, shelf talkers, and floor graphics that manufacturers paid for long before a pixel was ever served. The digital pivot came when e-commerce gave retailers a new inventory to monetize, and someone realized ad space on a product page was worth something.
She credited the team at Walmart’s Triad unit – the in-house group that managed supplier marketing programs – as “probably the pioneers of maybe what we would know of retail media today.” From that starting point, the category has ballooned. Global retail media ad spending is projected to reach $100 billion by 2027, up from an estimated $75 billion in 2025, according to eMarketer forecasts, as retailers increasingly package their first-party data alongside ad inventory.
Wyatt pointed to Walmart Connect and Target’s Roundel as benchmarks for what the category can become – networks that have built genuine media infrastructure, not just bolted an ad server onto a shopping site.
The measurement problem
For all the sector’s growth, Wyatt described measurement as a persistent frustration – one she has carried through most of her career. The core value proposition of retail media, she argued, is the retailer’s first-party data: both for audience targeting and for closing the loop between an ad impression and an actual purchase. The trouble is that the industry hasn’t agreed on what “closing the loop” actually means in practice.
“Despite the fact that retailers have access to all of this great data, there’s still such intense discussion in the industry about what do we do with that data and how does it actually drive outcomes,” she said. “Can we agree, and then can we agree across retailers if those outcomes are the same?”
The debate over incrementality versus return on ad spend cuts to the heart of it. Wyatt argued that basic ROAS figures, while useful as a rough business metric, don’t answer the more important question of whether advertising is actually moving the needle. “When you’re really thinking about is my media adding value, it needs to be an incremental ROAS number,” she said. Progress is being made, she acknowledged, but “not as fast as everybody would like.”
Garbage in, garbage out – even with AI
The question of what separates winning retail media networks from also-rans drew a pointed answer. Scale alone, Wyatt suggested, only gets the largest players so far – and for everyone else, a generic pitch about retailer brand equity isn’t enough.
“It’s not enough to say you should spend with me because I’m XYZ retailer,” she said. The networks pulling ahead are those that translate the actual differentiators of their retail business into their media product, creating coherent value on both sides.
On AI – which she noted was unavoidable at industry events, suggesting a “shot game” at conferences would leave attendees unable to walk out – Wyatt is pragmatic. “It’s very much garbage in, garbage out,” she said. “If you haven’t done the work to create a really solid data foundation, implementing any kind of AI efficiencies is going to be really, really hard.”
Series Videos
