At the Retail Technology Show this year, retail media wasn’t being framed as a future opportunity. It felt immediate, commercial and under scrutiny. Across the programme there was a clear shift in tone, away from potential and towards proof. If this is going to be a serious revenue line, it has to show impact in the same way as any other channel.
Frasers Group’s ELEVATE session sat right in the middle of that conversation. Led by Richard Lallo, Group Head of Customer Marketing, it focused less on the promise of retail media and more on what has to be built to make it work. The headline themes were familiar on the surface, scale, personalisation, omnichannel activation. What made it more interesting was how those pieces were connected, and the level of infrastructure required underneath them.
The starting point is data, but not in the abstract sense that tends to dominate these discussions. Frasers has spent the past few years pulling together a single view of its customer, using email, device and transaction data to stitch together profiles and remove duplication. Each customer then sits within a dataset that includes more than 20 attributes, covering behaviour, preferences and engagement.
That foundation underpins everything else. With around 32 million customer records, roughly half of the UK adult population, the business has both scale and range. That range matters. It spans customers buying entry level sportswear through to high spend luxury, across categories including fashion, fitness, gaming and home. What that creates is not just reach, but a view of how people move between categories and price points, which is where a lot of retail media value sits.
From there, the conversation shifts into activation. One of the stronger points made in the session was that retail media only becomes meaningful when it moves across the full funnel. Frasers is working across onsite, offsite and in store inventory, using the same audience logic to reach customers at different stages, from awareness through to conversion. That includes everything from sponsored products and digital screens to out of home placements, experiential activations and connected TV.
What changes with a unified dataset is how those channels are used. Rather than planning media around formats, campaigns are built around audiences. That means identifying a segment with a clear propensity to buy, then mapping inventory to where that audience is most likely to engage. In practice, that can mean placing out of home media in very specific locations based on where target customers are concentrated, rather than defaulting to high traffic sites.
A campaign example with Nike illustrated how that plays out. Built around a New Year training push, the campaign identified likely buyers within Liverpool and Manchester, then activated across those regions using a mix of in store activity, gym based experiences, digital screens and out of home placements. Content from physical activations was reused across channels, creating continuity rather than fragmentation. The approach was less about blanket coverage and more about concentration, focusing spend where it was most likely to convert.
Measurement is where the model either holds up or falls apart. This is where Frasers is pushing the idea of closed loop reporting, linking media exposure directly to sales. Instead of reporting on impressions or clicks in isolation, performance is tracked against incremental sales uplift, using control groups to compare regions or stores where campaigns are active versus those where they are not.
The numbers shared reflect that focus. Across campaigns, the business is reporting returns of more than £12 for every £1 spent, alongside an average uplift of 33% in new customers and an 87% repeat booking rate from brand partners. Those metrics are doing a specific job. They move retail media out of the experimental budget and into something that can be compared directly with other marketing investments.
There is also a clear push towards automation. The next phase is about making the platform easier for brands and agencies to use, reducing friction in planning and activation. The direction of travel is towards a model where a campaign brief feeds into an orchestration engine, which then identifies audiences using first party and external data, generates channel specific creative through AI, and deploys it across inventory. Performance data is then fed back into the system, improving targeting and creative decisions over time.
What sits underneath all of this is a more pragmatic view of the retail media market. It is not an easy win. Grocers are already well established, media agencies have strong relationships with brands, and there is no shortage of inventory. What differentiates one retailer from another is not the presence of media space, but the quality of the data and the ability to activate it effectively.
That is the real takeaway from the session. Retail media only works when it is built on a foundation that connects customer understanding with media execution and commercial outcomes. Without that, it risks becoming another layer of complexity in an already crowded market. With it, it starts to look like something much more durable, a channel that is measured, optimised and scaled in the same way as the rest of the business.
