Retail media and customer analytics is software and analysis that reads every basket recorded by a retailer's loyalty card, and uses it for two jobs. The first job is to set the retailer's prices, promotions and ranges, meaning which products each store stocks. The second job is to sell brands advertising aimed at the retailer's own shoppers. The analysis is often called customer science. It began with a British supermarket's loyalty card in the 1990s, and is now sold to retailers in many countries.
A loyalty card links each basket to a person. Over months, the records show how often each household shops, which brands it buys, and when it switches. Analysts group shoppers by these habits: families who cook at home, young people who buy ready-to-eat food, shoppers who buy only on discount. Then they ask practical questions. Does a discount on cooking oil bring in new shoppers, or does it only give money away to people who would have bought the oil anyway? Which products are bought together? Which products could a store stop stocking without losing customers?
Consider a supermarket chain with 90 stores across Tamil Nadu and 25 lakh loyalty members. The analysis shows that a store near an IT park in Chennai sells far more ready-to-cook food than the chain's average. So the chain gives that store more shelf space for it. The analysis also shows that a 20% discount on one brand of rice mostly went to regular rice buyers, and brought in few new ones. The chain drops that promotion. These recommendations go to price and promotion optimisation software, which sets the next prices.
The same data supports the second job. A shampoo brand wants to reach shoppers who have recently bought a rival shampoo. The chain finds 1.2 lakh members who did so in the last three months. It shows them the brand's advertisement in its own app and website. The brand never receives the list. Four weeks later, the chain uses the same loyalty records to report how many of those members bought the brand's shampoo. Selling advertising on a retailer's own app and website in this way is called retail media, and the system that sells it is a retail media network.
Retail media is attractive to brands because the retailer can prove sales, not only clicks. That proof comes from first-party data, which the retailer collected itself. The basket records arrive from loyalty programme software. Proving which advertisement caused a sale is a question of marketing attribution.