A 100% digital bank, FLOA specializes in financial services that facilitate day-to-day transactions for consumers and merchants. A leader in fractional payment in the French market, the company offers a wide range of bank cards, innovative financing solutions (instant mini loans), and payment options (FLOA Pay). With over three million customers, it employs around 300 staff and funds products and services to the value of over 2.5 billion euros each year. FLOA won the « Élue Service Client de l’Année » (Customer Service of the Year award) in 2022 for the second year in a row.
In 2019, FLOA took a strategic turn by setting up a data-centric organization based on a unified data architecture built on the Snowflake platform. Today, it can be proud of choosing a solution that makes a significant contribution to developing and optimizing its activities.
Reorganizing the company around data: The Snowflake Data Cloud for innovation and performance
For historic reasons, FLOA’s IT infrastructure depended on three distinct information systems that didn’t talk to each other much, if at all. This situation wasn’t optimal for data usage, particularly since one part of the data architecture was outsourced and based on aging technologies.
In 2019, FLOA decided to set up its own information system with the Azure Cloud. It created a project team and then, in 2020, an independent department dedicated to data and responsible for designing and coordinating an architecture based on the Snowflake Data Cloud which would consolidate all its information sources, whether they originated from its own internal systems or those of its partners.
FLOA’s goal at the time was ambitious: it wanted to develop a data strategy that responded to the company’s different needs for analysis, predictive modeling, and data sharing, and was looking for a technology partner to break down silos and detect fraud and risks in real time, to optimize the customer experience and effectively lead the company.
To have the resources to achieve this dual aim—a data-driven company and automated real-time detection—FLOA decided to favor the technologies emerging at the time, including the Snowflake platform, Dataiku, and DataGalaxy.
FLOA was not in a situation of migrating existing systems but rather of creating a data platform from scratch. The Snowflake platform met its expectations perfectly. The FinTech needed to base its approach on a modern outlook that offered high performance and flexibility, allowing it to handle large data volumes with, for example, over two million scoring processes a month.
« Snowflake was the logical solution as it entered territory where other solutions couldn’t so easily go. Aware of our accelerating needs, we put our trust in the promising growth and development of the Snowflake Data Cloud » explained Sébastien Robert, Chief Data Officer at FLOA. « Thanks to the possibilities offered by the platform, we were able to benefit from the recast of our information system to rethink the place of data within the company. It had to be central, secure and accessible, irrespective of the profile of the contributor and their specific needs. »
Industrializing the identification of fraud or risk profiles
In addition to this reorganization around data, FLOA was also seeking to automate key processes for its critical activities, for example, the detection of fraud or risk profiles. To facilitate the identification of suspicious or atypical behavior, the company used Snowflake and Dataiku to implement fully industrialized processes based on data science. In more concrete terms, the online banking facility could easily rely on advanced predictive analyses, personalized according to the client journey set up with its partners, and to the product purchased.
It takes FLOA on average 150 milliseconds to calculate a score on our various journeys by executing one of the thirty predictive models that we’ve developed, each model being adapted to a type of request and each model itself able to respond differently based on the input parameters for the request. Truly critical performances for a company that controls and agrees up to 60,000 fractionated credit payments per hour on Black Friday!
« We have a data platform that means we can automate the analysis of large data volumes in order to increase the speed at which the data is processed and improve our responsiveness. For example, we recently automated the provision of a report which previously required ten days of work per month to make vital data needed by the company readily available. To give you another example, it also means we can identify patterns of suspicious transactions and set up a continuous training system » explained Sébastien Robert, Chief Data Officer at FLOA.
With instances of attempted fraud increasing by 30% a year, FLOA can be proud of having reduced fraud by 30% a year thanks to its industrialized processes.
Always better and faster with Snowflake
When it chose Snowflake in 2019, FLOA had a very specific objective: « We wanted to onboard a partner capable of evolving with us over time. It’s clearly the case that now, thanks to Snowflake, we’re doing better, working faster and achieving things that we couldn’t before », summed up Sébastien Robert.
From a quantitative standpoint, FLOA has drastically reduced the development time on certain projects with the Snowflake platform. Typically, where previously two to three months were needed to set up a dashboard to monitor activity, just a few days is now all that’s needed. The democratization of access to data has led to a change in culture within FLOA with working methods that have evolved and gained in efficiency. Because beyond these essential technical enablers, this rapid transformation has required a genuine entrepreneurial strategy around Data.