Groupe Beaumanoir: Revolutionizing the
Fashion Industry With AI

Full Interview Transcript

Hello, my name is Briac Le Dur, Data and CRM Director at Groupe Beaumanoir. The Beaumanoir Group, in a few words, is a textile group from Saint-Malo, specializing in ready-to-wear. Today, we have five ready-to-wear brands: Cache-Cache, Bonobo, Bréal, Morgan, Caroll, a multi-brand commerce platform, Sarenza.com, and a logistics subsidiary, called C-Log. Today, there are 2,500 points of sale around the world and 1,600 stores in France.

Can You Shed Some Light on Your Collaboration With Dataiku?

The collaboration with Dataiku, started in 2016. At that time, the Beaumanoir Group heavily invested in its tech stack. Between 2015 and 2016, a large-scale big data project was launched. Since then, Dataiku immediately became central to our business activities. The main users of the solution, over the past eight years, have been analysts and data scientists. Today, we have around 75 models in production. Scoring models, purchase propensity models, segmentation models, and sales forecasts, for all our brands and all business areas.

What Major Data Science Challenges Have You Faced in the Retail Sector?

The retail sector is a sector in danger, a very complex sector, and a highly competitive one. It’s a sector in which we find many brands and a sector in which margins become smaller every time. A major challenge for us today is therefore to succeed in positioning ourselves in this very complex market. 

A second major challenge for us, is also the increase of costs related to visibility of our brands, especially through digital channels where the costs of customer acquisition are increasingly high. We needed to accelerate using data, accelerate our transformation to be able to precisely address two major challenges which are operational efficiency, to achieve profitability, and also to increasingly support our marketing teams in using marketing engine tools that are available to them and always striving to seek better returns on investment. 

We are also very lucky, because the Beaumanoir Group has a very rich database. In terms of customer data, we’re talking about 4% to 5% of customers ingested in the Customer Data Platform (CDP), we’re talking about 88% of contacts who can be activated. Also, in terms of client insights, we have approximately 300,000 clients who contact our customer service, providing us with very interesting information regarding our challenges or the areas for improvement that we need to address. We also survey 500,000 customers per year and we also have between 30 and 40,000 data insights on store reviews that we can use. 

This is a gold mine of information that we need to make the most of. And the last important point as well is that it’s a gold mine in terms of product attributes data that we can analyze and cross-reference with our customer data, since we’ve reached about 300,000 EANs over the last two years.

How Has the Beaumanoir Group’s Data Utilization Evolved After Adopting Dataiku?

When in 2016 we chose the Dataiku solution, all teams quickly adopted it quite simply, whether they were part of the more technical teams, like the data scientists, or whether they were more business users. The data scientists quickly adopted it not only because the implementation was very simple, it only took a few days within our Azure infrastructure, but also because of the ease of upskilling of our teams.

They were able to quickly benefit from using the solution whether in terms of preparation, data visualization, or simply the capacity of the solution to test a multitude of models and to choose the best for our various use cases. So this has been a very positive point. And what has really been a game changer for us, is the aspect of automation bricks which allowed us to scale on all our projects. This deployment was really challenging before using the Dataiku solution.

Can You Illustrate How Dataiku is Being Used?

The first project that illustrates the use of the solution Dataiku within the Beaumanoir Group, is our ability to calculate a purchase appetite score for all our customers, for all our brands, and in all departments. Today, this is something that we use very regularly in all our targeting, whether it’s CRM, whether it’s on the classic direct marketing channels or even when we create look-alikes or when we do media purchases on platforms such as Meta or Google. 

This has allowed us to significantly improve our return on investment, or even to adapt the content that we can push to our customers depending on whether they are likely to purchase an item or not according to a department. We can target much more the messages we send them. 

The second project that we can highlight is a project that is more recent. Since 2016, or even 2014-2015, we already had a CDP in place within the group, with very specific uses. And with the Dataiku solution, since 2018-2019, we set the goal or at least, we challenged ourselves to be able to integrate this solution within our ecosystem. That’s why, through WebApp solutions that we made available to our CRM teams, today, we are able, thanks to the Dataiku tool, to both target and work on personalized content for our customers using product recommendation models and to develop connectors on the various media channels that we can activate on the market.

This is a project that took two years, but also a project that today allows us to save on a market solution, since today it costs us much less than a solution off the market. More importantly this solution is more suited in terms of productivity gains since we worked on it, we shaped it, adapting it to our use cases and in particular our multi-brand use cases, in the context of Cache-Cache, Bonobo, Bréal.

What Does the Future Look Like for Dataiku Within the Beaumanoir Group?

After eight years of collaboration, we can say that we have gone quite far in terms of machine learning and deep learning with Dataiku. We accelerated our data transformation within all the brands of the Beaumanoir Group. Now, we need to go even further. And going further, means working more and more on AI and Generative AI.

We launched within the Beaumanoir Group a project called Kalliopé. Kalliopé is a muse for all that is eloquence, and epic poetry. This project aims to push and encourage each employee of the Group to suggest their own Generative AI project. This project also has a purpose of regulating, organizing and structuring, so that we don’t do anything random either with Generative AI. Dataiku should be able to precisely help us in this journey, and help us make all these Generative AI projects successful.