Retail media and Shopping ads at a glance
Retail media, as the retail media guide describes it, covers four broad kinds of placement: sponsored listings in the retailer's search results, placements on product and category pages, display and homepage slots, and off-site audiences built from the retailer's shopper data. It is bought self-serve in the retailer's own interface, as a negotiated package, or through a demand-side platform where the retailer offers that route. Its reports come from the retailer's own sales, which is both its appeal and its weakness: the seller defines what counts as an attributed sale, one retailer at a time. The guide names no retailer, network or ad product, and neither does this page.
Shopping ads are built from a feed submitted to Google Merchant Center, whose product data specification documents the attributes each item needs: title, description, link, image, price, availability and the rest. The platform decides from the feed which searches each product matches, assembles the ad from the image, title and price, and sends the click to the product's link. Standard Shopping shows these listings on search results and the Shopping tab with the advertiser controlling product groups and bids; Google's help page on Performance Max describes a campaign that uses the same feed, with asset groups and audience signals, across Google's inventory towards a conversion or value goal. The shopping and catalogue ads guide and the Shopping and Performance Max hub cover how to run them, and the product feed guide covers the feed; this page does not repeat either.
Compare table
Retail media vs Shopping ads
| Row label | Retail media | Shopping ads |
|---|---|---|
| Where the ad appears | On the retailer's site and app, and on other sites using its shopper data | On Google's search results and Shopping tab, and across Google's inventory with Performance Max |
| Who sells it | The retailer, self-serve or as a negotiated package | Google, through Google Ads |
| Who can buy it | Usually brands whose products the retailer stocks | A business with a product feed and pages where the products can be bought |
| Whose data targets it | The retailer's: what its shoppers search for, view and buy | The brand's feed, matched by the platform to searches, plus any audience signals supplied |
| Where the click lands | The retailer's product page; the retailer makes the sale | The product page named in the feed, usually the brand's own store |
| Who defines a sale | The retailer, including whether other products from the brand count | The brand, through its own conversion tracking |
| Reporting window | Set by the retailer; ask for it before reading any figure | Set in the brand's own conversion settings |
| Customer data kept | Aggregated reports under the retailer's rules; the customer stays the retailer's | The order and the customer record, in the brand's own systems |
| Creative needed | The product as the retailer lists it, plus banners for display placements | The feed's image, title and price; asset groups as well for Performance Max |
| How to read its return | Against a holdout by region, store group or audience; never on the reported figure alone | Against break-even return from margin, and a holdout before scaling |
The third and fifth rows decide most cases. Retail media is open to brands the retailer stocks and leads to the retailer's checkout. Shopping ads need a feed of products the advertiser sells and lead to the advertiser's own pages. So for many product brands the choice is not between two ads for the same sale. It is between the sale that happens in someone else's shop and the sale that happens in the brand's own.
When to choose retail media
Choose retail media when most of your sales happen on retailers' sites and shoppers search there rather than on the web. The FMCG marketing guide explains why that holds for everyday products: people choose between brands on a retailer's results page, influenced by what they remember and what is visible, and the brand rarely sees who bought. Sponsored listings and product page placements sit exactly there. They suit a new product with no reviews or organic position on the retailer's site, and a brand whose rivals are taking the sponsored positions on its category searches.
Choose it knowing what you are buying. The reports come from the seller of the product being evaluated, the customer stays with the retailer, and paid listings on searches where the product already ranks may move sales from organic to paid without adding any. The retail media guide's questions for any proposal, starting with how an attributed sale is defined and whether a holdout can be run, are the price of entry. Where the brand also lists its own products on a marketplace, the marketplace channel page covers sponsored placements from the seller's side, bid against break-even after fees.
When to choose Shopping ads
Choose Shopping ads when you sell from your own store and people look for your kind of product on the web. The ad is built from data you control, the click lands on your page, the order and the customer are yours, and the margin is the whole retail margin rather than what is left after a retailer's share. Targets can be set from that margin: the break-even calculator gives the return an order must reach to pay its costs, and the ROAS calculator checks a campaign against it. Standard Shopping gives the most control over which products compete and what each bids; Performance Max takes the same feed further across Google's inventory, and the comparison of Performance Max and Standard Shopping covers that choice.
Shopping ads are a weak substitute for retail media when the buyer will not leave the retailer. Someone doing a weekly grocery shop on a grocer's app is not searching the web for a sauce. For a brand in that position, more Shopping ads will reach people looking for the product online, which is a different audience from the one choosing at the grocer's shelf.
Decision table
| Situation | Choose | Why |
|---|---|---|
| Sold mainly through retailers, shoppers search the retailer's site | Retail media, with a holdout | The ads sit where the choice is made |
| Sold mainly from your own online store | Shopping ads | Your data, your margin, your customer |
| Sold both ways, with an own store that is growing | Shopping ads first, retail media tested | Own-store sales keep the customer; retail media must prove it adds sales |
| A launch at one retailer with no reviews or organic rank | Retail media for a period | Sponsored placement gives visibility until the product earns its rank |
| A retailer's package arrives with a projected return | A pilot with a holdout before signing | The projection is the seller's figure |
| A brand with no product pages of its own | Retail media and the retailers' listings | Shopping ads need pages where the product can be bought |
| Both running and reporting returns above target | Holdouts on each, then compare incremental return with margin | Reported returns follow different rules |
Running both: one product range, two buyers' journeys
A brand sold through retailers and from its own store has two kinds of buyer. One searches the web, sees a Shopping ad and buys direct. The other opens a retailer's app, searches the category and adds the product to a basket with everything else. Retail media and Shopping ads can run side by side because they meet those two journeys at different points; trouble starts when they are judged as if they met the same one.
Three rules keep the pair honest. First, keep the budgets and targets apart: retail media's return is the retailer's figure on the retailer's sales, Shopping ads' return is the brand's figure on its own orders, and a single blended return mixes two definitions. Second, read each against what it caused, using a Holdout GroupA randomly chosen part of an audience kept from exposure to measure what exposure caused. rather than the platform's credit. The comparison of iROAS and ROAS explains why only the return measured against a holdout can justify a budget line, and the incrementality testing guide sets out the designs, regional for retail media where the retailer allows it, and a geography or campaign holdout for Shopping ads. Third, set each target from the margin that channel actually earns. A sale through a retailer earns the brand its wholesale margin; a sale in its own store earns the retail margin less fulfilment. The comparison of marketplaces and your own store sets out the wider trade between selling through someone else's shop and your own, which is where the budget split between the two ad types finally comes from.
A sauce brand weighs a grocer's package against its own Shopping ads
Glomvaris Sauces, an invented sauce brand, sells through two grocers and its own online shop. The larger grocer offers a retail media package of £40,000 over three months, projecting £4.50 of attributed sales for every £1. The brand's own Shopping ads spend £8,000 a month at a reported return of 4.5. Before deciding, it asks the grocer for a one-month pilot of £12,000 with placements shown in six regions and withheld in three similar ones. The grocer reports £54,000 of attributed sales, a return of 4.5. Against the holdout regions, sales grew 3%, which implies £21,600 of extra sales, an incremental return of 1.8. Glomvaris earns 30% on each pound of grocer sales, so the extra sales bring £6,480 of margin against £12,000 of spend; at that margin the incremental return would need to reach 3.33 to break even. In the same month it adds £4,000 to Shopping ads and holds out two regions; the extra spend brings £11,200 of additional own-store orders, an incremental return of 2.8, against a break-even return of 1.82 on its 55% margin, so the extra spend earns £6,160 of margin. Glomvaris moves the extra money into Shopping ads and offers the grocer a smaller test limited to sponsored listings for a new flavour's launch.
To check a retailer pilot at the wholesale margin rather than the own-store one, the ROAS calculator tutorial for retail media enters the lift and reads it against that higher break-even.
Common mistakes
- Comparing the two reported returns. Each follows its own definition of a sale and its own window.
- Judging retail media on the retailer's figure. The seller marks its own work; only a holdout shows what the ads caused.
- Expecting Shopping ads to reach the retailer's shoppers. People buying inside a retailer's app are not searching the web for the product.
- Setting one return target for both. A retailer sale and an own-store sale earn the brand different margins.
- Running Shopping ads on a weak feed. The feed is the campaign; fix titles, prices and availability before adding budget.
- Paying for searches on your own name without testing. Sponsored listings on terms where the product already ranks often buy sales the organic listing would have made; ask the same question of Shopping ads shown to people searching for the brand.
- Signing a package without a pilot. A short test with a holdout costs less than a quarter of spend that cannot be judged; the comparison library lists other pairs that are easy to confuse.
