Modernising Carrefour’s data architecture

About the customer 

Carrefour Belgium employs more than 11,000 people. With about 40 Carrefour Hypermarkets, more than 440 Carrefour Markets, 300 Carrefour Express convenience stores and 200 Carrefour Drive collection points, Carrefour Belgium is the first and only retail chain that helps customers at all times in their lives and at every stage of consumption.

Thanks to its local roots, in almost every municipality, and its unique range of 11,000 local products and more than 4,000 organic & well-being products, Carrefour Belgium is the ideal partner in daily life. That is why 700,000 customers, or about 80% of Belgian families per year, do their shopping at Carrefour every day.

Carrefour Belgium is part of the Groupe Carrefour, a large food retailer with more than 360,000 employees. The group is active in more than 30 countries with more than 12,300 stores and achieved a turnover of 76 billion euros in 2020 with its formats.

Gregory Pierquin joined Carrefour 7 years ago, and started his journey as “self scan product manager”.  This was an important application as there was direct contact with the customers during their shopping journey in the stores (Make sure to listen to our podcast featuring Gregory Pierquin, check it out here) 

Today, Gregory’s team manages all data platforms @ Carrefour.  Gregory considers it important to have a coherent and consolidated view on all data available on one hand, but secondly – and more importantly – he wants to be close to the master data to have a good view on the customer. And also to know and understand the customer better, to manage the relationship with the customer in a correct way and to make personalized and relevant offers.

The problem / challenge 

Carrefour is a company with a lot of data, and they have been using their data for more than 20 years.  In the beginning the company used Excel files. Today they want to use all customer data and all other available data to make the right and quick decisions in real-time.  

The Carrefour – before called “GB” – Loyalty Program has been in place for more than 50 years. The program feeds the value and importance to Carrefour of understanding their customer and to provide personalized offers. Carrefour wants to stay far away from “spamming” their clients.  

To tackle this challenge, Carrefour took it as their mission to review their data architecture and to combine all available data. An extra challenge was to avoid data duplication, because there were multiple data platforms in the past.

In striving to achieve this mission, the data team at Carrefour came across some hurdles, caused by an old-fashioned big data technology on premise that was in place.  

Carrefour Belgium started to discuss with Carrefour Groupe,  Google Cloud and Google Cloud Managed Services Partner Devoteam G Cloud as their trusted partners to achieve their important data mission.

The goal

The objective was to put in place a data hub in an accelerated manner, assuring that all Carrefour entities use the same technology. Another goal was to get a unified data model in place, as well as assuring the solution and approach are cloud native.  

From a business perspective, the objective was to have ‘easy’ access to the data, to get real-time, valuable insights to make quick decisions.

The solution 

Devoteam G Cloud proposed to use a data & analytics accelerator framework to accelerate the implementation of the Carrefour data hub.  

Carrefour chose this data framework for a number of reasons. First of all, Carrefour was looking to build a framework and welcomed the fact that partner Devoteam G Cloud had already developed a robust framework. This saved them time.  

Furthermore, the Devoteam data framework allows for solving the data replication problem and to simplify the ETL pipeline.  

Last but not least, the Devoteam data framework uses a full DevOps ecosystem (Gitlab), which was also aligned with Carrefour’s strategy to operate more in a DevOps way. 

The methodology

First, Devoteam G Cloud had some workshops with Carrefour to define how the organisation wanted to organise their data warehouse.  Once this was defined, the team deployed the data framework.  

In this case, deploying the data framework meant that a GKE cluster and all Google Cloud resources were created and configured on the Carrefour account, to be ready for the data engineers to start working on the data. The Devoteam G Cloud team analysed what Carrefour is producing as raw data every day, and then transformed this data to another format which is more useful for reporting and day to day insights and operations.  

Initially, the framework was used for some first test cases.  In the beginning, only 1-2 data engineers worked along with Devoteam G Cloud to learn and to test some use cases. Just after a few months, some of the pipelines were pushed into production.

Nowadays, Carrefour have grown their data engineering team and the data engineers are operating autonomously on a day to day basis.  Devoteam G Cloud remains on point to help them remediate bugs and respond to any feature requests they have. 

The feedback from Carrefour has been very important to the Devoteam G Cloud team working on the data framework.  The development of this framework is a continuous process, assuring that all users benefit from the latest enhancements.

The result

The next goal is to assure that on the business side, ‘getting data’ is made easy, without the need to do coding or to get help from data engineers.  The business will be able to access the data directly to make quick, real-time decisions.   

Achieving this goal will help Carrefour to become the global leader in the food transition for all, by offering their 13 million daily customers healthy quality food at the right price – every day and everywhere.