Retail media or search: where the next pound should go

8 min read

Retail media or search: where the next pound should go

8 min read

Every e-commerce marketing director is now pitched from two sides. Retailers and marketplaces want a bigger share of the budget for retail media, while search, paid and organic, still brings shoppers to your own site. The next pound has to go somewhere, and the finance director will want to know why it went there.

Retail media is advertising sold by a retailer or marketplace on its own website, app and stores, targeted with what it knows about its shoppers. For brands that sell mostly through retailers it can be the strongest line on the plan. For brands whose margin sits on their own site, it can mean paying a retailer to rent customers you would rather own.

This article explains how retail media differs from Google search and Shopping, how we decide where the next pound goes, how to test it, and what to report upwards.

Why the split matters to the business

The money is moving. IAB UK's Digital Adspend study, produced with Oliver Wyman and published in March 2026, put UK retail media spend at £3.8bn in 2025, up 18% on the year, against £17.9bn for search. Those are market totals. They tell you where other advertisers spend, not where your next sale will come from.

Get the split wrong one way and you pay retailers for sales that would have happened anyway, often on your own brand name, while the retailer keeps the customer data. Get it wrong the other way and you give up the digital shelf where your category is actually bought, to a competitor that bid where you did not.

Get it right and the board sees every channel judged on the same measure: the extra contribution it adds after its cost, and how many new customers it brings in.

What retail media is, and how it differs from search

Retail media networks sell ads on properties the retailer owns, and increasingly on other sites using the retailer's data. Three examples UK brands will recognise:

Amazon Ads. Sponsored Products are cost-per-click ads that promote individual product listings on Amazon and select apps and websites. Amazon also offers Amazon Marketing Cloud, a clean room in which advertisers can analyse pseudonymised Amazon Ads signals alongside their own data.
Tesco Media, which describes itself as a partnership between Tesco and dunnhumby built on Clubcard insight, reaching shoppers online, in store and through offsite partners.
Nectar360, which offers digital advertising and Sainsbury's shopper insight built around the Nectar loyalty programme.

Google search and Shopping work differently. Shopping ads draw on the product data you send through Google Merchant Center, and the click lands on your own site. For a board, the differences that matter are commercial rather than technical:

Who owns the customer. A sale through a retailer sits in the retailer's customer record. A sale on your site gives you the email address, the repeat order and the lifetime value.
Who sees the data. Retail media is closed loop: the retailer sees both the ad and the purchase, so it can report sales against spend. On your own site you measure through your analytics, with the consent and attribution gaps that come with it.
Margin. A retail media sale carries the retailer's or marketplace's terms. A direct sale carries your full margin, less fulfilment and returns.
Where the shopper is. Someone searching Amazon is already shopping. Someone searching Google may be researching, comparing or reading an AI answer first, so search reaches earlier in the decision as well as at the point of purchase.

How we decide where the next pound goes

We work through the same seven steps whether the starting budget is mostly retail media, mostly search or split between them.

Start with where revenue and margin sit today. We split the last 12 months of sales by route to market (your own site, Amazon and each retail partner) with contribution margin per order after fees, fulfilment and returns. A channel that drives a large share of revenue but a small share of profit should not get the same share of the next pound.
Work out what a customer is worth in each channel. On your own site you can measure repeat purchase and lifetime value from your order data. Through a retailer you usually cannot, so we treat a retailer sale as a single order unless the retailer's data shows otherwise.
Check your own search position first. We look at coverage for non-brand category terms in organic and paid search, the quality of your Shopping feed, and whether your product and category pages are cited in AI answers. If the site is weak there, fixing it is often cheaper than buying the same demand elsewhere, and our guide to Merchant Center feed optimisation covers the feed side.
Separate defence from growth in retail media. Bidding on your own brand inside Amazon protects the listing, while bidding on category and competitor terms wins new shoppers. They are different decisions with different returns, so we report them apart. Amazon's new-to-brand audience covers shoppers who have not bought from your brand in the past 12 months, which gives a first read on whether spend is finding new customers.
Put every channel on the same measure. ROAS on a retailer and ROAS on your site are not comparable, because the margin behind each pound of revenue differs. We compare contribution per pound spent and cost per new customer instead. In Google Ads, customer lifecycle goals let Search, Shopping, Performance Max and Demand Gen campaigns bid higher for new customers, or only for new customers, which makes the comparison fairer.
Test for incrementality before scaling. Retailer-reported sales show what was attributed to the ads, not what the ads caused. We design a hold-out by region, product line or time period and measure total sales across every route to market, using incrementality testing rather than platform dashboards; our piece on marketing mix modelling and incrementality explains when a model is worth building on top.
Move money in tranches against a rule agreed in advance. Before the test starts, agree with finance what result would justify the next tranche and what would stop it. That turns a budget argument into a decision the board has already signed off.

When retail media wins

Most of your sales already happen through retailers, which is common in grocery and household goods. The retailer's shelf is your shelf, and the retailer's data is the closest you will get to your customer.
Amazon is where your category is bought, and competitors already bid on your brand and category terms there. If you run a marketplace yourself, the same questions apply from the other side, and our visibility playbook for marketplaces and aggregators sets out how we approach them.
You are launching into a retailer's range and need trial at the point of purchase.
You need audience data, such as loyalty card insight, that no other channel can give you.

When your own site and search win

Your margin, repeat purchase and customer data sit on your own site, as they do for a direct-to-consumer brand.
Your search coverage has gaps: category terms where competitors rank and you do not, a thin Shopping feed, or pages that AI answers never cite. The e-commerce visibility playbook sets out how we close them.
Your range changes quickly, as in fashion, where your own site can show the new season the day it lands; our visibility playbook for fashion e-commerce covers that model.
Most of the proposed retail media budget would go on your own brand name, defending sales you would probably win anyway.

How to measure it and what to report upwards

Report the channels together, on measures finance already uses:

Incremental contribution by channel: the extra margin each channel added in tests, after its cost.
Cost per new customer by channel, with lifetime value on your own site shown alongside.
Share of sales by route to market, quarter on quarter, so the board can see whether the business is becoming more dependent on retailers.
In retail media, the split between brand and non-brand spend, and the share of attributed sales that were new to brand.
Total revenue against total marketing spend, as a check that channel results add up.

Be careful with retailer-reported numbers. A retail media network marks its own homework with data you cannot audit. Use its reporting to optimise inside the platform, and use your own tests and total sales to make budget decisions.

Mistakes we see most

Comparing retailer ROAS with site ROAS as if a pound of revenue were worth the same in both places.
Funding retail media from the search budget without first testing what search was delivering.
Spending most of the retail media budget on brand terms and reporting it as growth.
Treating attributed sales as incremental sales.
Forgetting that Google's results now include Shopping listings and AI answers, so weak search visibility also loses sales that end up at retailers.

Questions to ask your team or agency

What share of our retail media spend is on our own brand terms, and what happens to total sales if we pause it in a test region?
What is our contribution per pound spent and cost per new customer in each channel, on the same definitions?
Which non-brand category searches do we lose on Google today, and what would it cost to win them?
Which test would settle the question, and what result would make us move money?

How we approach it

We plan search as one budget across organic, paid and AI visibility, and we will say when retail media is the better home for the money. If you are weighing up the split, our integrated search team can design the test with you. Get in touch to start that conversation.

Frequently asked questions

What is retail media?

Retail media is advertising sold by a retailer or marketplace on its own properties, such as its website, app, stores and loyalty programme, and increasingly on other sites using its shopper data. UK examples include Amazon Ads, Tesco Media and Nectar360. Because the retailer sees both the ad and the purchase, it can report sales against spend, although attributed sales are not the same as extra sales.

What is the difference between retail media and Google Shopping ads?

Google Shopping ads use product data from your Merchant Center feed and send the shopper to your own site, where you keep the margin, the customer record and any repeat purchase. Retail media sends the shopper to a product page on the retailer's site, and the retailer keeps the customer. Retail media reporting is closed loop, while search is measured through your own analytics.

How should a brand split budget between retail media and search?

Start from where revenue and margin sit today, then compare channels on contribution per pound spent and cost per new customer, not platform-reported ROAS. Brands selling mostly through retailers usually need retail media, while brands with strong direct margins usually get more from search. Move money in tranches, each backed by an incrementality test whose success threshold is agreed with finance in advance.

How do you measure retail media incrementality?

Run a hold-out test: pause or reduce spend in some regions, product lines or time periods and compare total sales across every route to market against a control group. Retailer dashboards report attributed sales, which include purchases that would have happened anyway, particularly on brand terms. Agree the measure and the success threshold with finance before the test starts.

Is Amazon advertising worth it for brands that also sell on their own site?

It can be, if Amazon is where your category is bought and competitors are taking those shoppers. Separate brand defence from category and competitor campaigns, check what share of sales came from shoppers new to your brand (Amazon counts those who have not bought from you in the past 12 months), and compare the margin with a direct sale before scaling spend.