Case Study: Cdiscount uses Crownpeak Product Recommendations for algorithm orchestration
Key results
The challenge
Cdiscount, one of France's largest e-retailers with more than 9 million customers and 23 million monthly unique visitors, wanted to better segment and target its offers to personalise the customer journey. Its in-house data team was developing its own algorithms, but the volume brought by its marketplace exceeded internal resources. The retailer needed a partner to accelerate R&D and let its merchandising team control merchandising rules without relying on data scientists.
The solution
Cdiscount selected Crownpeak's Product Recommendations solution, integrated in server-side mode with no front-end development required. The platform lets Cdiscount's data teams use their own and third-party algorithms in one tool, processing large data volumes in real time, and its algorithm-orchestration capability finds the best combination of algorithms and fallback strategies on a continuous basis.
“Algorithm orchestration is something we didn't have before. Thanks to Crownpeak we now have the ability to stack strategies and algorithms using fallback strategies that allow us to find the optimum combination of different algorithms.”
SBSimon Berthet-BondetHead of Merchandising, Cdiscount
The results, in context
Complementary strategies powered by Crownpeak algorithms produced a 2.5% increase in conversion rate; integrating journey models across 100% of search-list strategies added a 3% increase; and new merchandising rules added a further 1% increase. The business team also took control of merchandising rules directly, cutting setup work that used to take several days or weeks to less than a day.