Retail media for small UK brands: plan and judge a test

What to take away

Retail media is advertising sold on a retailer's own digital properties. On-site formats appear in search results, product pages and category listings on the retailer's website or app. Off-site formats extend the retailer's first-party shopper data to external platforms such as Meta and Google. For a small brand, the first retail media test has three parts: pick a route you can afford, set the budget and the control before spending, and judge the result on incremental sales.

How entry costs are charged

The headline cost is the media itself, charged per click or per thousand impressions. Self-serve options such as Amazon Ads charge no monthly platform fee. That means an independent shop can start with a few hundred pounds and no contract. Sellers on eBay have a comparable self-serve route in Promoted Listings, which places their listings higher within eBay's own marketplace.

Managed networks work differently. They sell campaigns through a sales team, and most set a minimum spend per quarter before they will take a booking. Ask the sales team for the minimum in writing, and for what it includes, before you plan anything else.

The UK market is dominated by retailer-owned networks. Tesco Media, Sainsbury's Nectar360 and Boots Media Group are the best-known examples, and each sells access to its own shoppers rather than to a general audience.

Add the costs that sit outside the media invoice: creative, feed management, and the consent work needed to use customer data. Some of the retail media costs that surface later never appear in the first quotation, so ask for them by name: account or platform fees, creative refreshes, product feed fixes, reporting or data add-ons, and VAT on agency fees.

Using your own customer list for an audience means handling personal data lawfully. The ICO's direct marketing guidance sets out what that involves.

Two laws sit behind that guidance: the UK GDPR and the Privacy and Electronic Communications Regulations (PECR), which between them set the rules on consent and opt-outs for marketing.

A fee quoted as a percentage of media spend looks small next to a large budget and large next to yours. Ask for the cash figure at your own spend level before you compare anything.

Why managed networks set minimum spends

Minimum spends exist because a sales team has to be paid for. A network that requires a quarterly commitment is pricing human time, not advertising space.

A small brand can respond in one of two ways. Test on a self-serve platform first to learn how the auction behaves. Or take a single placement for a fixed period and treat the cost as research.

Read the minimum as a floor on your commitment, and check whether unused spend carries over into the next period.

None of these routes removes the risk. They only tell you what the risk costs before you sign.

A worked example on a £3,000 test

The figures below are illustrative, chosen to show the method rather than to forecast a result. The test splits a £3,000 budget into £2,400 of media and £600 of setup, runs for 60 days against a baseline period of the same length, and is judged on incremental revenue.

If the £2,400 of media buys £5,000 of incremental revenue at a 40 per cent gross margin, that is £2,000 of gross profit before the £600 of setup. The test has not paid for itself yet, and now you know by how much.

Timing matters, because UK retail sales move sharply with season. The ONS monthly release Retail sales, Great Britain is the public record to check before you choose a test window, so you can compare the same weeks a year earlier.

Measuring whether it paid

Reported return on ad spend counts sales that would have happened anyway. A small brand cannot afford to pay for those twice.

The cleaner approach is to hold back a control group or a control period. In a holdout test, a matched set of products, regions or weeks runs without the ads, and the difference in sales between the advertised and held-back sets is the incremental effect. For a single-product brand, a before-and-after comparison against the same weeks a year earlier, adjusted with the ONS retail sales series, is the fallback. If you cannot run a proper test, treat the result as directional rather than proven.

If you are deciding who runs the account, compare self-serve and managed routes before you agree to a percentage fee.

One more check: a sponsored placement must be recognisable as an advert. The ASA advertising codes are the standard to test against: the CAP Code requires marketing communications to be obviously identifiable as such.

For a small brand the verdict is rarely a simple yes. It is a budget line, a minimum spend and a measured result, checked in that order.

Common questions

Do I need a large budget to start with retail media?

No. Self-serve platforms charge per click with no monthly fee, so a few hundred pounds is enough to learn how the auction behaves. The constraint is your time, not the entry price.

Which route is cheapest for a small brand to test?

Self-serve on Amazon Ads, using your own product data and existing creative. It keeps the fixed cost low while you find out whether the audience responds. Shopify Audiences works differently: it helps target ads on external platforms rather than selling space on a retailer's own site, so treat it as an off-site option.

How long before a small brand knows whether it worked?

Give a test at least one full seasonal cycle, ideally 60 days, with a baseline period of the same length. Anything shorter mixes the campaign effect with ordinary week-to-week variation.