Improving data accessibility and operational efficiency

GEODIS, headquartered in Levallois-Perret, France, is a global leader in logistics and transportation. The company generates around 12 billion euros in revenue, employs 50,000 people across 70 countries, and offers a wide range of services, including intercontinental and road transport, last-mile delivery, and contract warehouse management.

To maintain its competitive edge, GEODIS recognized the need to optimize its data management strategy by improving operational efficiency and driving business growth.

Historically, GEODIS operated using decentralized data teams across its various business units, leading to significant data silos. This fragmentation resulted in several challenges, including:

• Complexity and performance issues: The existing mix of big data and cloud-native technologies created a complex environment, slowing down projects and hindering performance.
• Lack of visibility: Incomplete visibility into the data landscape made it difficult to leverage data effectively across the organization.
• Scalability and accessibility: Difficulty in accessing and scaling data hindered the launch of new projects and impacted overall efficiency.

The company needed a unified data platform to de-silo its data, enhance data accessibility, and streamline operations.

GEODIS modernized its Data Platform and built a single data access service with Starburst Enterprise, which established shared working principles across business units. Key aspects of the solution included:

• Unified data access: Starburst provided a single point of access to GEODIS’ diverse data sources, regardless of the underlying storage system or the type of data, including ADLS, Hive, Impala, Kudu, Kafka, Apache Iceberg, MongoDB, and ElasticSearch.
• Performance and scalability: Starburst’s powerful query engine replaced Hive, Impala or MongoDB BI Connector, delivering 6X faster performance and allowing queries to run in an average of 3.5 seconds. This transition from complex heterogeneous components to Starburst’s modern engine was crucial for handling high-concurrency, real-time analytics workloads, which includes 139 000 business requests per day. The ease of using SQL is a plus for their data processing platform.
• Seamless integration: The platform integrated seamlessly with GEODIS’ existing Cloudera architecture for security management and data cataloging, extending Ranger security rules and replicating the Hive metastore.

The implementation of Starburst Enterprise allowed GEODIS to simplify data access, optimize query performance, and facilitate the integration of new consumers and BI tools.

The deployment of Starburst Enterprise led to the following improvements for GEODIS:

• Real-time insights: Business analysts gained the ability to perform analyses faster and use near-real-time data, improving decision-making processes. “Business analysts perform analyses faster and have near-real-time data usage due to the rapid querying capabilities of live data,” noted Amato.
• Operational efficiency: Starburst became the go-to platform for data access, supporting nearly all use cases, including financial reporting, supply chain optimization, and real-time warehouse management. “Starburst has helped to streamline operations and exceed business lines expectations. Starburst is now the main tool for business units to build reports by federating data from the data lake and data warehouse with huge positive impacts on simplicity and performance,” said Hicham Guelai, GEODIS’ Data Architect.
• Increased productivity: “Starburst significantly simplified our work, as data analysts no longer needed to rely on data engineers to prepare the data and can use near-raw data,” said Amato.

Future outlook:
GEODIS aims to enhance data product management by building data products. The company is also organizing around AI topics, implementing tools for AI/GenAI use cases, and ensuring smooth integration between Starburst and Dataiku for future AI projects. The ongoing collaboration with Starburst will continue to enhance GEODIS’ data landscape, driving further innovation and efficiency. “Our partnership with Starburst is crucial as we advance our data initiatives and integrate AI into our operations,” concluded Amato.