Système U: Modernizing its data architecture using Google Cloud

About Système U

Système U, comprises almost 1,600 retail outlets in France and around 73,000 employees. Its origins go all the way back to 1894—over a century ago—when the very first retail groups began to emerge. From convenience stores to superstores, the development of each retailer in the group relies on the services provided by three central entities: U Enseigne (retail, marketing, finance, human resources management), U Log (supply chain management), and U Iris (IT).

Industries: Food & Beverage, Retail & Consumer Goods
Location: France
Products: Google Cloud, BigQuery

Tell us your challenge. We're here to help.

Contact us

Improving the performance of Système U's data information system.

  • Streamlined its process over time, by merging its data lake and data warehouse into one single platform, thanks to BigQuery
  • Plans to ingest data in real time (streaming), to open up new opportunities for the cooperative
  • Merged its data lake and data warehouse into one single platform

"The IT system has become more complex, and we are not able to develop specialized skills in every area. For the applications that will be migrated to the cloud, we are overcoming constraints on the deployment and operation of the infrastructure, whilst opening up the possibility of using a large range of ready-to-use technical components.”

Philippe Le Cam, Executive Director, U Iris

Improving the performance of Système U's data information system

U Iris, Système U's IT division, manages all of the IT tools across the 1,600 retail outlets. Having streamlined its software at a regional level, it implemented a cloud-based modernization strategy. The aim of this is to create more agile, innovative software that can easily adapt to combat challenges in the retail sector at a low cost, whether that be in terms of customer experience, optimization of the supply chain, or even exploring and capitalizing on new business opportunities for U stores.

"The IT system has become more complex, and we are not able to develop specialized skills in every area. For the applications that will be migrated to the cloud, we are overcoming constraints on the deployment and operation of the infrastructure, whilst opening up the possibility of using a large range of ready-to-use technical components," explains Philippe Le Cam, Executive Director of U Iris.

Google Cloud: Freeing innovation from technical constraints

The cooperative, which uses Gmail and has also been using Google Workspace since 2013, naturally turned to Google Cloud ahead of trialing its public cloud system in 2020. Barely a year later, in June 2021, the cooperative decided to implement its strategy with its first project: a major redesign of its data architecture in the cloud. After evaluating the solutions on the market, including carrying out an extensive comparative study with an expert in the area of cloud data, Système U is opting for BigQuery for clearly defined reasons, as explained by Philippe Le Cam: "Our ambitions for transformation go way beyond the data. With BigQuery, Google Cloud has positioned itself as a partner in our transformation. For future applications that will be based in the cloud, we expect to use all of Google Cloud's services."

Google Cloud is supporting U Iris directly through its Professional Services and is also assisting the project team with its migration. "This support helps us to deliver faster and is an excellent accelerator in our transformation. It brings real added value," adds Philippe Le Cam.

"BigQuery is able to manage both of our data approaches: structured data and raw data. In doing so, it simplifies our architecture, optimizes performance through the scalability of the cloud, and reduces our operating costs by about 40%, since we will be using managed services that we do not have to maintain or update. Our teams will also be able to fully devote themselves to tasks that bring real added value to their businesses.”

Philippe Le Cam, Executive Director, U Iris

Reducing data operating costs

Alongside its data warehouse for reporting, U Iris stores its raw data in a data lake on Hadoop in order to meet increasing analysis demands from the businesses. Adding an additional layer of solutions with varying degrees of technicality allows businesses to complete more detailed analyses. Aimed at a wide range of users, these two databases essentially store the same data. Put another way, the two databases take up unnecessary storage and resources, essentially processing the same data twice. Thanks to BigQuery, it is now possible for Système U to streamline this process over time, by merging its data lake and data warehouse into one single platform.

“BigQuery is able to manage both of our data approaches: structured data and raw data. In doing so, it simplifies our architecture, optimizes performance through the scalability of the cloud, and reduces our operating costs by about 40%, since we will be using managed services that we do not have to maintain or update. Our teams will also be able to fully devote themselves to tasks that bring real added value to their businesses,” explains Philippe Le Cam.

More efficient businesses

In the same way, the IT teams will have a wide range of technological building blocks at their fingertips that they can use to respond faster to the needs of the businesses.

As they will no longer have to develop a technical component before they can use it, the teams will become more responsive and agile because changing a component will have a smaller impact: no prior training, no time lost to design the component, etc.

Finally, Système U's IT system can directly optimize the use of servers (and therefore energy) depending on its uses, thanks to the serverless aspect of BigQuery when adding a server. The new technology will also streamline recovery processes. Any disaster-recovery-plan tests carried out on the current infrastructure can take up to a week due to the complexity of the technologies that need to be implemented to process the data in Hadoop.

Performance for facing new challenges

On the businesses side of things, benefits will come from significantly simplified usage and improved independence. Analysts and data scientists can work with their data directly in BigQuery with Dataiku, rather than managing multiple different technologies, which they do currently. Performance is also improved, with some processes running up to 95% faster based on the tests completed so far. This performance also extends to data ingestion, a particularly critical point according to Philippe Le Cam: "Volumes are continuing to increase and capturing data is becoming more and more difficult. The windows for ingesting data cannot be extended indefinitely. With process times reducing from 15 minutes to nine seconds, we are no longer concerned about our ability to ingest data in BigQuery."

Système U even plans to ingest data in real time (streaming), a feature that is not possible with its current architecture but is offered by BigQuery, and this opens up new opportunities for the cooperative. The cooperative expects to mostly use the feature to ingest the data from receipts in real time.

"In retail, knowing your data is like having a gold mine in the palm of your hand. Thanks to this architecture, we will not only be able to improve customer experience, but also optimize our stock, forecast orders, help stores to increase their productivity… the applications are endless. The variety of the Google Cloud network—from ready-to-use ML services to Vertex AI to geographic information systems (GIS), or even data-sharing to create new services by sharing anonymized data with third parties—will help us to accelerate our transformation.”

Philippe Le Cam, Executive Director, U Iris

A network of services that will multiply the potential of the data

With migration of its data and processes fully underway, Système U is already looking at other potential areas of profitability. The cooperative intends to capitalize on the Google Cloud ecosystem's wealth of resources to get the most out of its data, with a particular focus on machine learning (ML) services and other geographic and data-sharing features. This development is even easier to attain, and Google Cloud services are already set to implement and perfectly integrate this into Système U's new infrastructure.

"In retail, knowing your data is like having a gold mine in the palm of your hand. Thanks to this architecture, we will not only be able to improve customer experience, but also optimize our stock, forecast orders, help stores to increase their productivity… the applications are endless. The variety of the Google Cloud network—from ready-to-use ML services to Vertex AI to geographic information systems (GIS), or even data-sharing to create new services by sharing anonymized data with third parties—will help us to accelerate our transformation," concludes Philippe Le Cam.

Tell us your challenge. We're here to help.

Contact us

About Système U

Système U, comprises almost 1,600 retail outlets in France and around 73,000 employees. Its origins go all the way back to 1894—over a century ago—when the very first retail groups began to emerge. From convenience stores to superstores, the development of each retailer in the group relies on the services provided by three central entities: U Enseigne (retail, marketing, finance, human resources management), U Log (supply chain management), and U Iris (IT).

Industries: Food & Beverage, Retail & Consumer Goods
Location: France