Saint-Maclou selects Fivetran and Snowflake to modernize its business data platform

Fivetran Streamlines Real-Time Data Ingestion for Saint-Maclou, Transitioning to an ELT Approach

Fivetran’s architecture is helping Saint-Maclou, a French flooring retailer, automate the ingestion of data from approximately 15 sources into Snowflake. By moving away from manual processes, Saint-Maclou adopts an Extract, Load, Transform (ELT) strategy to enhance data accessibility, reallocate technical resources to higher-value tasks, and prepare for advanced analytical uses, including artificial intelligence.

In many organizations, data projects often encounter a hidden challenge: data collection. When a platform’s data feeding relies on custom development and constant maintenance, information is delayed, hampering departments’ ability to leverage its full potential.

Saint-Maclou, based in Lezennes in northern France, operates 132 stores and a logistics warehouse, serving both individual and professional customers. With over 60 years of experience, the company employs 1,400 people.

Legacy Decision-Making Systems Impeding Analytical Initiatives

As part of its information system modernization, Saint-Maclou initiated a gradual overhaul of its data platform. This evolution allows the company to better utilize data from sales, installation services, finance, marketing, logistics, and inventory management. The operational goal is to access data quickly, responding to more frequent and evolving business demands without service interruptions.

Prior to this transformation, Saint-Maclou relied on a legacy decision-making system supported by manually developed and maintained data streams. Each new analytical project required an extensive pre-phase to identify relevant source databases, tables, and columns, followed by custom development for data ingestion.

This setup placed a significant burden on development, maintenance, and support. Data engineers were heavily occupied with data collection, often at the expense of transformation and analysis work. Project timelines were inevitably prolonged, directly impacting finance, supply chain, human resources, marketing, sales, and logistics, while limiting opportunities for experimentation with advanced analytical scenarios and AI-related use cases.

Snowflake: The Centralized Data Warehouse

To alleviate these challenges, Saint-Maclou opted for a modern cloud-based data architecture with Snowflake as its central database. This decision was critical in creating a data warehouse capable of accommodating growing data volumes and supporting diverse analytical uses.

“We faced significant costs related to development, maintenance, support, and licensing fees due to our previous workflow,” says Salmane Khamlichi, Head of Data at Saint-Maclou. “To accelerate data availability and generate value quickly, we shifted to a modern cloud data architecture with Snowflake as our central database.” With this foundation in place, the reliable and rapid ingestion of data became a structural priority.

Fivetran Automates Ingestion and Promotes an ELT Approach

To feed data into Snowflake, Saint-Maclou selected Fivetran and transitioned from an ETL (Extract, Transform, Load) to an ELT methodology. Fivetran was chosen for its native compatibility with Snowflake, robust performance validated during a proof of concept, and its ability to handle large SQL Server databases and critical data sources.

The ELT approach involves loading raw data directly into the data warehouse before applying transformation processes. In contrast, ETL requires prior data transformation in intermediary tools before loading. ELT takes advantage of the computing power of cloud data platforms like Snowflake to execute transformations closer to the analytical use cases, enhancing traceability and flexibility. This strategy simplifies data flows, minimizes specific maintenance needs, and accelerates the integration of new data sources. While ETL remains relevant in highly regulated environments, it often introduces delays and limits architectural flexibility when business requirements evolve rapidly.

Reduced Timelines and a Focus on Value-Driven Exploration

Currently, Fivetran automatically synchronizes data from around 15 sources, both on-premises and in the cloud, into Snowflake. These sources include Microsoft AX ERP, logistics tools, sales systems, inventory management systems, as well as multiple marketing applications, including the upcoming Salesforce CRM. This setup eliminates data collection bottlenecks, with transformation managed closer to the data warehouse, particularly through DBT Core.

Saint-Maclou reports that the most immediate gains are in speed. Integrating new sources, which once took days or even weeks, can now be accomplished in a matter of hours. This shortened ingestion cycle allows projects to stay within the same sprint, reducing costly back-and-forth.

Organizationally, data engineers are less burdened by operational tasks. Data ingestion no longer requires the equivalent of a full-time position, enabling resources to be redirected toward data transformation and analysis. The stability of synchronizations and the automatic management of schema changes enhance confidence in available data, reducing the risk of service disruption.

A Unified Foundation for Advanced Analytical Uses

Centralizing all data within Snowflake provides Saint-Maclou with a unified view, where information was previously scattered across multiple business applications. This consolidation fosters collaboration between data teams and marketing, sales, and logistics teams.

It also paves the way for more strategic projects, such as sales forecasting, advanced analytics, and AI initiatives. Meanwhile, Fivetran’s capabilities for tracking data consumption help manage costs associated with data projects. “Today, data collection is no longer a concern. Thanks to Fivetran, our teams can focus on transformation, allowing our work to reflect its true added value,” concludes Salmane Khamlichi.

Scroll to Top