For this article, Irma Doze spoke with Florent Maire, Head of People Analytics Reporting at Societe Generale, about the bank’s journey toward stronger people insights, better data foundations and a predictive approach to HR decision-making. Their conversation offers a clear look at how a large global organisation transforms not only its reporting capabilities but also the culture around data.
Societe Generale is one of Europe’s major banking groups, employing more than 100,000 people across regions, business units and functions. With such scale, consistency in HR-data is both critical and challenging.
Florent leads the central People Analytics reporting team in Paris, supported by a functional team of around ten specialists in Bucharest. Together, they create and maintain core HR dashboards used across the organisation.
From the very beginning, the focus was on building a solid foundation. Many organisations try to move fast with analytics, only to discover that poor data quality undermines trust. Societe Generale deliberately took a different route. “We spent several years building the HR data lake and the governance around it. This is what allows everything else to happen.”
Instead of one unified HRIS, the organisation operates several Workday instances, configured differently across countries. Integrating these systems into a group-wide HR data lake was a significant undertaking, but it created one reliable source for reporting. Today, all central dashboards feed directly from this data lake.
This foundational work enabled another major shift: HR leadership now uses a single, consolidated view. As Florent shares, the CHRO relies on the group dashboard in discussions with the Executive Committee. This creates clarity and alignment across business units, and it sends a strong signal that the organisation is committed to facts and consistency.
Creating a People Analytics Community across the organisation
Building strong reporting tools is only one part of the story. Adoption requires engagement, communication and the ability to bring people together. Because Florent is not hierarchically responsible for local HR analytics practitioners, he needed another way to create alignment.
The solution was the People Analytics Community, launched this year. Florent mapped all analytics profiles across business units and support units, identified who was working on analytics locally and brought them together in a structured forum. “It is a platform for everyone to learn from each other,” he explains. “Local teams bring their own challenges and best practices. We share group-level updates in return.”
This approach encourages shared ownership of data quality, definitions and reporting standards. It also provides a space where local improvement ideas can be elevated to the group level. Over time, the community will likely become a home for analytics translators as well, helping bridge the gap between HR professionals and technical teams.
The impact is visible. As the group dashboard gained authority, local teams increasingly rely on central insights. Instead of maintaining parallel reporting structures, they now work collectively toward a shared understanding of the workforce.
Moving from descriptive to predictive insights
After strengthening its descriptive reporting capabilities, Societe Generale is taking the next step: predictive analytics. One of the first major projects in this space focuses on resignation risk. “We knew our dashboards were strong, but they were descriptive. We wanted to anticipate what might happen next,” Florent says.
The central reporting team partnered closely with the HR data science team for this initiative. While the data science team brought modelling expertise, Florent’s team ensured that the output could be industrialised and embedded in the existing dashboard landscape. This partnership was essential. Predictive models often fail not because the math is wrong, but because they never reach the end user in a usable form. By combining modelling and BI expertise, the two teams created an integrated predictive resignation dashboard that allows HRBPs and HRDs to explore risk levels and the most influential factors.
The development process took around nine months. It required detailed data validation, careful UX testing and multiple rounds of feedback with HRBPs. “We wanted to make sure that people could find the right information, interpret it correctly and feel comfortable using it in conversations with the business,” Florent explains. The first rollout took place in France, where data quality is strongest. Plans for international expansion will accelerate as the organisation converges toward a more unified HRIS landscape during 2026.
Challenges along the way: data, translation and trust
Every major analytics initiative encounters obstacles, and this project was no exception. Florent explains that one of the first challenges was determining which data could be used reliably. The team initially considered using the declared reason for resignation, but quickly discovered that this field was too inconsistent to support predictive modelling. They shifted toward behavioural and structural data points that offered stronger quality.
A second challenge was translating HR needs into actionable metrics. HRBPs clearly recognised the importance of understanding resignation risk, yet expressing precisely which indicators they needed proved difficult. The team spent considerable time converting these business questions into analytical constructs that would work in a model.
Finally, internal validation and governance added complexity. Predictive analytics requires far more inspection than descriptive dashboards, and the validation cycles were lengthy but essential to ensure responsible use and build trust across the organisation. Together, these challenges highlighted that scaling predictive analytics responsibly depends on clarity, patience and close collaboration.
Engaging HR professionals and encouraging data-driven conversations
Technology alone does not create impact. People must feel equipped and confident to use data in their daily work. Societe Generale pays attention to this. The predictive dashboard rollout included onboarding sessions for all HRBPs and HRDs in France. They received early access, training and space to ask questions. This helped them understand not only the tool, but the conversations it could unlock.
Florent emphasises that adoption is a journey. The dashboard was launched only recently, and the team will continue to track usage, gather feedback and refine the product. As more HR professionals use predictive insights in discussions with leaders, the organisation will discover new opportunities for improvement.
Leadership sponsorship as a catalyst for change
A turning point in the people analytics journey came when the organisation appointed a new CEO and CHRO. Both leaders brought a global perspective and a strong commitment to data-driven decision-making. “When the CHRO expects one single answer from everyone, people align quickly,” Florent notes.
The leadership shift also encouraged a move away from a historically France-centric approach. Today, the organisation aims for global consistency without losing sight of local realities. This combination of top-down sponsorship and bottom-up community-building creates the right environment for analytics to thrive.
Lessons learned from Societe Generale’s people analytics journey
Across the conversation, Florent shares several insights for organisations advancing their own people analytics capabilities. Taken together, these lessons highlight that people analytics succeeds when organisations combine solid data structures with a human approach to collaboration and learning.
- Invest early in data foundations. If data quality is poor, dashboards will not be trusted or used. Strong foundations are non-negotiable.
- Build clarity through governance. Definitions, ownership and access rules prevent confusion and maintain consistency across the organisation.
- Partner across teams. Analytics requires BI, data science, central HR and local teams working together.
- Engage the business. Onboarding, training and open forums support understanding and adoption.
- Treat predictive analytics as a long-term capability. Models need governance, iteration and trust.
What resonates most from this conversation, as Irma Doze observed, is how consistently Societe Generale invested in clarity and alignment before advancing its analytics capabilities. Many organisations rush into predictive modelling without the foundations to support it, while Societe Generale chose the more sustainable path: build trustworthy data, involve people through a community and scale responsibly. It shows that meaningful people insights arise not from tools alone, but from the discipline to create shared understanding, the courage to challenge data quality and the willingness to collaborate across teams.