Retail media for an England buyer: networks, formats and an entry decision

Retail media is advertising enabled by inventory or data a retailer or marketplace controls. It can appear on a retailer's website or app, away from those properties, or inside a physical store.

The useful boundary is ownership and permission: an advert beside a shop is not automatically retail media. Nor is ordinary merchandising advertising merely because a brand benefits.

An England business considering this market should define one commercial exchange before sizing anything. Name the retailer or marketplace, the advertiser, the inventory, the audience or location, the data used, the payable event and the evidence promised after delivery. No official dataset currently measures that complete population for England.

What to take away

Use a definition that survives procurement

IAB UK's retail media definition covers digital advertising space, retail data assets and in-store opportunities owned by a retailer or marketplace and made available to brands. It divides opportunities into on-site, off-site and in-store environments and includes targeting, optimisation and measurement.

That is a trade definition, not a statutory category. It also reaches further than the definition used in IAB UK's Digital Adspend 2025 study. The study labels advertising within retailer or marketplace owned digital properties as retail media. Its reported total therefore cannot be silently expanded to all off-site and in-store activity.

Write a one-page scope card, a checklist for one deal, with these fields:

If any field is unknown, mark it unknown. A familiar label does not repair a missing contract or data flow.

Keep the parties separate

The retailer owns or lawfully controls the relevant shop, site, app, customer relationship or media inventory. A media owner or network packages access to that inventory. The advertiser commissions the communication. An agency may plan, buy or create it. A platform can serve the advert, while a data supplier or measurement provider may handle audience or transaction records.

One group can perform several functions, but the evidence still needs named roles. A retailer's product record is not the advertiser's claim substantiation. A platform impression is not proof that a person noticed the advert. A payment record is not automatically incremental sales. Each organisation should approve only the decision it owns.

Who sells retail media in the UK

Retail media is sold UK-wide, not by English region, so an England buyer normally buys from a UK network. Examples of retailers selling advertising on their own channels include Tesco (through Tesco Media & Insight Platform), Sainsbury's (through Nectar360), Ocado Retail, Boots (through Boots Media Group), Asda and Amazon Ads. Trade coverage also lists Morrisons, Co-op, Superdrug, John Lewis, Waitrose, Currys, M&S and B&Q.

Offers differ by retailer and none is ranked here. Compare them with the UK retail media network comparison, and see how grocery and health retailers sell sponsored placements.

Ad formats and how they are bought

No rate card is quoted here, because prices vary by retailer, category and season and are normally quoted for each deal. Ask for the billing unit, any minimum spend, the reporting you will receive and the VAT position. Compare self-serve and managed buying, and see what retail media agencies charge.

What the current market evidence can say

IAB UK and Oliver Wyman reported £3.8 billion for UK retail media in 2025. The published methodology and category note say the study uses member submissions, public figures, statistical modelling and qualitative interviews. It also says the method changed from the half-year 2025 work and that 2024 was restated.

This is a current UK industry estimate for the study's online definition. It is not an England total, an official statistic, a retailer count, advertiser demand, gross merchandise value or proof of a particular network's revenue. The method does not provide the England allocation needed for this guide.

The ONS Retail Sales Index internet sales dataset, released 21 August 2026, records Great Britain internet retail sales by store type and time. It measures retail activity, not advertising.

The ONS UK business activity, size and location dataset counts enterprises and local units from an Inter-Departmental Business Register snapshot, with geography and SIC classifications. Neither isolates retail-media owners, advertisers, inventory or spend.

Do not multiply those business counts by an assumed media fee. Do not apply a population share to the IAB UK estimate. Both moves would create precision without an observed relationship.

Build an England evidence register instead

Start with identifiable media propositions (the selection checklist shows how to record them) offered by retailers or marketplaces that have a verifiable England operating connection. The geography rule must fit the question. A registered office in England may establish company identity but not where an audience was reached, an advert ran or revenue arose.

For each admitted proposition, record the legal contracting entity, inventory class, date available, advertiser eligibility, charge basis, reporting definition and source. De-duplicate trading brands that share the same offer. Exclude general digital advertising, ordinary product listings, retailer turnover, supplier marketing pages without an inspectable service, and any off-site use whose retailer-data role cannot be shown.

Two different denominators may be needed. A supply study could count eligible media propositions. A demand study could count qualified advertiser briefs or completed contracts. The latter must not use sales enquiries alone. Show briefs received, briefs meeting the scope, proposals issued and campaigns contracted as separate states.

Diagnose demand without inventing it

A practical demand signal is a repeated buyer problem with inspectable evidence. It might be an advertiser asking for a defined retailer audience, a retailer unable to reconcile booked and delivered inventory, or a campaign report that cannot connect a claimed outcome to an eligible order population. Record the observation, source, date, owner and counter-explanation.

Regulatory friction creates work, but it does not prove commercial demand. The CAP Code requires non-broadcast marketing communications to follow its rules, including recognition and misleading-advertising provisions.

The CMA's price-transparency guidance covers mandatory fees, taxes and charges when products are promoted to consumers. Those controls matter where a retail-media placement shows a product and price. They do not forecast advertiser spend.

Privacy evidence is another boundary. The ICO's guidance on storage and access technologies page history marks 29 April 2026 as the end of its draft status. It covers cookies, pixels, scripts, tags and similar technologies under PECR and, where relevant, UK GDPR.

Consent or another permitted route for a device operation does not establish statistical validity. A sound measurement design does not supply privacy permission.

Compare five operating models on one job

Choose a bounded job, such as selling and measuring a sponsored placement on a retailer product-search page. Then compare the current merchandising route, direct retailer media sale, managed network, off-site activation and data or measurement service against the same fields.

No route is automatically cheaper, safer or more effective. Supplier charges, internal labour, taxes, makegoods and refunds need comparable units before a financial decision. Buyer weights remain blank until the organisation states its risk and purpose.

Make market entry reversible

The entry test should use synthetic or specifically authorised records. Inspect one placement through booking, creative review, serving, consumer presentation, reporting, correction and withdrawal. Do not expose a live audience merely to complete procurement evidence.

Advertising review checks whether the communication is recognisable and whether objective claims have support. Consumer review checks product information, prices and choices. Privacy review maps each purpose, device operation and recipient. Accessibility testing covers the advert, controls and destination journey with defined user tasks. Security review covers access, logs, supplier connections and incident response.

The NCSC's secure online service guidance treats identity, access, monitoring, bots and incident management as operating concerns. These controls need evidence for the actual configuration. A supplier security page cannot prove the buyer's account is correctly set up.

Set stop conditions that good results elsewhere cannot offset. Hold entry if the retailer has no inventory authority, a claim is unsubstantiated, price information is misleading, device access has no approved route, or an accessible journey fails.

Hold entry if privileged access cannot be revoked, delivery cannot be reconciled, or campaign records cannot be exported.

Treat opportunities as hypotheses

An opportunity exists only when a named buyer has a costly problem and a feasible controlled test. Possible investigations include improving inventory evidence, separating consent-aware reporting from targeting, making campaign approval records portable, or reconciling media delivery with eligible transaction states. None is an England forecast.

For each hypothesis, keep the counter-signal. A retailer may already have a reliable internal process. Advertisers may not accept the proposed measurement. Necessary data may be unavailable or unlawful to use. The service may add more manual review than it removes. Stop when the evidence cannot distinguish the proposal from the current route.

Entry decision for 2027

For a UK brand that already sells through one retailer, the default is a single controlled test on that retailer's on-site sponsored-product or search placement, before off-site or in-store activity. Fix a budget you can afford to lose, a start and end date, and the outcome you will count. Use a holdout design from the incrementality test guide so the result is not only the retailer's own attribution.

Scale only if the delivery report reconciles to the invoice, claims are substantiated, and the test shows sales you would not have made anyway. Do not enter if the retailer cannot give reports you can check or if the privacy and advertising reviews are incomplete. Recheck the IAB UK definitions and study method when the next edition appears, and test one retailer-advertiser exchange at a time.

Before you act

Common questions

Which UK retailers sell retail media?

Examples include Tesco, Sainsbury's (Nectar360), Ocado Retail, Boots, Asda and Amazon Ads, among many others. They sell UK-wide, so an England buyer buys from the same networks as the rest of the UK.

Can I use the IAB UK figure as an England market size?

No. The £3.8 billion estimate is a UK industry figure for the study's online definition. It is not an England total, a retailer count or proof of any network's revenue, and no population share should be applied to it.

What does retail media cost?

There is no published rate card. Costs depend on the retailer, format and season, and are usually quoted per click, per thousand impressions or as a fixed fee. Ask for the billing unit, minimum spend and VAT position before comparing offers.

How should a first-time buyer start?

With one controlled test on a single retailer's on-site placement, a fixed budget and a holdout, as set out in the entry decision above.