What is Retail Media

Retail media is advertising sold by a retailer or marketplace on its own website or app, and sometimes beyond it, using what it knows about its shoppers. A brand pays to show its product higher in search results, on product pages or in banners, right where people are about to buy.

• Seller of the ad: The store itself, such as Amazon, Flipkart, Myntra, Nykaa or Blinkit, not a search engine or social network.
• Where it runs: Inside the shopping app, next to the Add to Cart button, and sometimes on other sites using the retailer's data.
• Data behind it: What shoppers searched for, viewed and bought on that platform.
• Measurement: Sales are linked to ads on the same platform, which is called closed-loop measurement.
• Main formats: Sponsored products, sponsored brand banners, display and video.

This lesson follows one example: a family-run brand from Kochi that makes cold-pressed coconut oil and sells it on Amazon, Flipkart and Blinkit. A shopper who types "cold pressed coconut oil" sees dozens of bottles, and retail media decides whether the Kochi bottle is among the first few.

Key Characteristics of Retail Media

• Close to the sale: The shopper is already inside a store and searching for the product type, so very little persuasion is needed.
• Built on first-party data: The retailer uses its own shopper data instead of third-party cookies, which is why first-party data has become so valuable to marketplaces.
• Pay per click or per view: Most sponsored product ads charge when a shopper clicks, while banner and video ads often charge per thousand impressions.
• Measured in sales: Reports show ad spend, ad sales and ROAS for each product, not just clicks.
• Many networks: Each large retailer runs its own retail media network, with its own tools and rules.

How Retail Media Works

1. The brand lists its product: The coconut oil must be live on the marketplace, in stock and backed by a complete listing, because every ad sends shoppers to that listing and a weak one wastes money, as explained in product feed optimization.
2. The brand creates a campaign: It chooses the products to promote, a daily budget, and the searches or competing products to target, such as "virgin coconut oil".
3. An auction decides the slots: When a customer in Chennai searches "cold pressed coconut oil 1 litre", the marketplace compares every advertiser's bid with the relevance of each product, and the winners receive the sponsored positions.
4. The ad appears with a label: The Kochi bottle shows near the top of the results, marked as sponsored.
5. The sale is attributed: If the customer clicks and buys, the marketplace credits the order to that advertisement and displays it in the brand's campaign report.

Types of Retail Media

The platform tutorials cover these formats in detail: Amazon Ads, Flipkart Ads and Blinkit Ads.

Example: Retail Media for a Kochi Coconut Oil Brand

• Goal: Sell more of the 1 litre bottle on Amazon without losing money on individual orders.
• Margin first: The owner calculates what remains from each sale after product cost, marketplace fees and delivery, which gives the minimum ROAS the advertising can accept, as shown in break-even ROAS per SKU.
• Search and competitor targeting: One sponsored products campaign targets "cold pressed coconut oil" and "coconut oil for cooking", and another places the bottle on product pages of larger, more expensive brands.
• Quick commerce test: On Blinkit, where customers order for tonight's cooking, the brand tests a small search campaign in Kochi and Bengaluru, following the approach in quick commerce marketing.
• Review: After two weeks, the owner studies the search term report, pauses searches that spend without selling, and compares every product's ROAS with its break-even number.

Benefits of Retail Media

• Buyers, not browsers: Ads reach people who are already shopping for the product type.
• Clear results: Spend and sales sit in one report, so the brand can judge each product.
• Small budgets work: A seller can begin with a modest daily budget and increase spending on the products that perform.
• Possible organic lift: Additional sales may help a product's organic position on some marketplaces, although no platform guarantees this.

Limitations of Retail Media

• Walled gardens: Each retailer reports only on its own sales, so it is difficult to see the complete picture across Amazon, Flipkart and Blinkit.
• Credit can be too generous: A shopper who searched for the brand by name might have bought anyway, yet the ad still takes credit; the ROAS vs ACoS vs TACoS lesson shows how to spot this.
• Costs rise during sale events: Competition for the same searches increases during festive sales, so bids climb and margins shrink.
• More work: Every network has its own tools, reports, policies and approval rules.

How AI Changes Retail Media

What AI Automates Now

Marketplaces now use AI to pick which searches an ad shows for, to set bids automatically, and to write ad headlines or make product images. Some offer AI tools that suggest keywords or budgets for a new campaign.

What Still Needs a Human

Only the brand knows its real margin, its stock position and which products it wants to push. A person must set the ROAS floor, decide how much to spend on the brand's own name, and judge whether an ad sale was genuinely incremental, because automated bidding chases whatever goal it is given.

Risk to Watch

Automated campaigns can spend quickly on searches that bring sales at a loss, and they can keep buying the brand's own name, which looks impressive in reports but adds few new customers. Set a daily budget cap and read the search term report every week.

Do It with AI

Use this prompt to turn a marketplace search term report into a practical action list; it works in ChatGPT, Claude or Gemini.

1. Download the search term report from the marketplace advertising console and delete any column that could identify a customer.
2. Run the prompt, then recalculate two or three ROAS figures by hand to confirm the columns were read correctly.
3. Apply the changes in the advertising console, and review the same report again after two weeks.

Check Before You Use It

• Facts: Recalculate a few ROAS figures yourself, since AI tools can misread columns.
• Brand fit: Keep search terms that matter for the brand's image even when they are new and have few sales.
• Compliance: Follow each marketplace's ad policies, and never make health claims about the oil that the brand cannot prove.