Kocélla Mechouek’s interview
How the analysis of multi-channel Customer feedback has allowed Decathlon to innovate and become a pioneer of excellent Customer Experience.
I’m Product Owner in the Data Science / AI team at Decathlon. More particularly, I manage all the projects and use cases related to the offer, the design, and the creation of products. We are actually working on several subjects at the moment such as the Semantic Analysis of our product reviews, store reviews, interactions with the Customer Relationship Center… We are either in the POC (Proof of Concept) phase or in production with other AI technologies related to image processing, price elasticity, graph theory…
At the beginning, the Customer Relationship’s Center needed a solution that allowed them to quantify in a macro way the themes that generate dissatisfaction. Every collaborator was able to cite the main irritants felt by our customers, but no one could quantify them or follow their evolution over time. Then, we rapidly identified a need for our Product Leaders. In fact, they were reading and answering every customer product review, but did not have a tool to analyze them quickly and easily. It’s such a rich raw material to help design future product versions based on more qualitative information and to facilitate innovation.
When we started this Semantic Analysis project, we surveyed three companies including Proxem. One of the advantages of Proxem was its capacity to natively analyze a large number of languages. Then, we submitted a dataset to Proxem in order to measure the added value of opting for an external company rather than developing a solution internally. On this dataset, Proxem managed to go further in the analysis, in a few days (vs. several months internally). Futhermore, we don’t have the required profiles internally (infolinguistes for example).
For two years, the collaboration with Proxem has been very good and effective. Bugs are fixed within a very reasonable time. And all my demands of product evolutions (mostly about the Dashboards) are always taken into count with a great deal of care. I meet and talk regularly with Proxem’s teams, which helps to reinforce our relationship.
Some managers use the solution to identify action plans that improve user satisfaction. The semi-automatic moderation project of the product reviews allows Decathlon to save money because the increases in the number of reviews would have forced us to hire extra people to manually moderate the reviews. This semi-automatic process allowed us to absorb this strong increase without recruiting.
We plan to move towards Speech To Text and to analyze, in addition to the written reviews of our customers, their phone conversations. We also plan to analyze more product types in the context of the analyze of the customer product reviews. For several months, we have discussed internally and with Proxem, in order to frame a first use case of a chatbot. A last idea to think about: real-time analysis of questionnaire reviews.