Conforama’s small ROI steps on artificial intelligence

Pricing, product returns, recommendation, fraud… Sarenza has developed multiple uses of data and algorithms as part of a time-to-market oriented approach. Sephora has set up an AI Factory to industrialize its models. Carrefour is also accelerating strongly on data recovery.

However, the level of maturity is very heterogeneous in the retail sector. For some companies, the AI ​​and data strategy has yet to be designed and executed. Conforama has recently initiated its own, as its marketing director was able to present during the Hub Retail & E-commerce.

Conforama adopts a Data roadmap

For this first step, the brand has opted for the field of customer relations, where the adoption of AI in France is still low, indicates Léonard Cahon, senior consulting manager for Artefact. This use concerns 7% of French companies, compared to 50% worldwide. This assessment, he explains by the complexity of the projects.

For Conforama, the use of AI to personalize recommendations in its newsletters is however considered a quick win. The company indeed claims a project carried out in eight weeks, with gains estimated at several million euros. Each week, the distributor contacts a base of three million customers by e-mail.

“The personalization issues are key (…) We have therefore chosen to work on the definition of a Data roadmap intended precisely to advance this personalization”, declares Mélodie Charles, the marketing director of the furniture brand.

With its service provider Artefact, Conforama has thus defined a roadmap comprising short, medium and long-term objectives (up to two years), and structured around use cases. The purpose: “to deliver extremely concrete and rapid results in order to be able to develop more and more use cases”.

Personalization of newsletters to fund other use cases

Following an audit, the distributor drew up a list of priorities and rolled out the foundations necessary for their implementation. And for its first application of AI, Conforama opted for a “relatively simple use case. Personalization issues were already partially addressed at Conforama,” explains Mélodie Charles.

This consists in personalizing the recommendations sent every Tuesday by e-mail. The purpose of this newsletter is “to support the brand’s commercial highlights”. Now, recommendations are based on customer appetite. Previously, the e-mailing offered eight generic recommendations to the three million customers.

At present, 24 million personalized recommendations are sent to this database every week. But above all, the investments made in AI have quickly paid off.

“The use case has been largely profitable for Conforama, since we have increased the click rate on our e-mails by 15% and doubled the rate of return of customers in stores and on”, welcomes the marketing manager.

Direct financial and productivity gains

The distributor also claims productivity gains among its teams. The selection of products to recommend, a “quite time-consuming” task, was entrusted to a recommendation engine.

Employees previously responsible for these actions “have been repositioned on value-added and more strategic tasks to be able to develop new use cases.”

For this application of artificial intelligence, Conforama collected customer, product and transaction data. This represents 15 input data sources, plus 25 staging tables.

Their processing by a model, which is based in particular on 35 automated tests, is used to determine, for a consumer, the most probable purchase. This processing takes filters or business rules into account.

Key points: data availability and method

The operational process makes it possible to guarantee, for example, that a recommended product is in stock, or to push current promotions. Before arriving at a high-performance AI algorithm, Conforama and Artefact carried out around fifty tests.

Among the key success factors, the brand and its partner cite in particular the use of technologies that are easily controlled because they can be integrated into the IS. From day one of the project, Conforama’s IT teams were involved. The mode of work is another lever of action.

“Assume that you are going to run into problems. It’s normal, it’s innovation. Imagine an organization [soit des méthodes agiles et une équipe pluridisciplinaire, NDLR]allowing you to identify and solve these problems quickly”, advises Léonard Cahon.

Source link -97