Retail media for an England buyer, from a definition that survives procurement to an entry decision

Retail media is advertising enabled by inventory or data controlled by a retailer or marketplace. 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, and ordinary merchandising is not 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.

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 scope card with these fields:

• legal retailer or marketplace entity and the property it controls;
• media location, format and eligible audience;
• advertiser, agency and any platform or data supplier;
• campaign objective and the event that triggers a charge;
• data sources, controller or processor positions and device operations;
• sales or other outcome definition, attribution window and exclusions; and
• correction, withdrawal, evidence export and retention route.

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.

What the current market evidence can say

IAB UK and Oliver Wyman reported GBP 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 on 21 August 2026, records internet retail sales in Great Britain 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 dataset 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 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 can create 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 addresses mandatory fees, taxes and charges when products are promoted to consumers. Those controls matter where a retail-media placement presents a product and price. They do not forecast advertiser spend.

Privacy evidence is another boundary. The page history for the ICO's guidance on storage and access technologies marks 29 April 2026 as the point when its draft status ended. 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, and 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.

Route	Commercial event	Control point	Evidence required	Main dependency
Existing merchandising	no advertising charge unless separately agreed	retailer product presentation	approved product and placement record	retailer process
Direct media sale	buyer-defined booking, delivery or other billable unit	retailer-owned inventory	order, creative approval and delivery record	retailer sales operation
Managed network	contract-defined service and media charge	divided by agreement	retailer authority, agency actions and platform log	supplier and contract
Off-site activation	contract-defined audience or media event	retailer-data permission plus external inventory	data lineage, delivery and reconciliation	platform and data chain
Measurement service	buyer-defined report or analysis unit	source and method ownership	population, event definitions and uncertainty	access to compatible records

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 non-compensating stop conditions. Hold entry if the retailer cannot show authority over the inventory, a claim lacks substantiation, price information is misleading, device access has no approved route, an accessible journey fails, 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.

Decision for the 2027 planning horizon

Use 2027 as a scheduled recheck, not a promise of growth. The initial decision is whether one defined retailer-advertiser exchange has enough evidence for a controlled pilot. It is not whether retail media is fashionable or whether a broad market figure is large.

Before publication, re-open the IAB definitions and study method, regulator pages and relevant contracts. Assign named advertising, consumer, privacy, accessibility, security, commercial, accounting and measurement reviewers. Keep this page on editorial hold if the definition, geography, evidence population or legal status changes.