L'Oréal Luxe : Reconciling the democratization of data access and data confidentiality as part of a large-scale BI project

Equancy intervened in 2023 for L'Oréal's Luxury division as part of the implementation of an ambitious global performance management project.

For Florence VALLA BERTHOLET, Global Head of Data & Insights for L'Oréal's Luxury division, the key issue was data confidentiality and the rules to be put in place to manage access to a global BI platform designed for a vast and heterogeneous audience:  

Which teams, which profiles have access to which dashboards, on what geographical scale, while limiting access to certain sensitive information?

This is where the concept of Row-Level Security (RLS) comes into play; RLS is a set of functions that enable access to data to be restricted according to the rights of the connected user.

Unlike simply limiting access to reports or dashboards, RLS filters data at the level of each row. This means that each user only sees data that is authorised according to their privileges.

BI solutions such as Tableau, Qlik and PowerBI all incorporate RLS functions.

In theory, the process is simple:

1.           Identification of user roles and hierarchies.

2.           Definition of precise rules specifying the data accessible for each role.

3.           Implementation of these rules.

4.           Recipe to ensure that each user accesses the authorised data.

However, the BI project for L'Oréal's Luxury division combines various brands, confidential data (finance, HR, etc.), complex geographical perimeters with a country/plate hierarchy, and roles with varied rights. This structural complexity undermined the very essence of the BI platform as a reliable source of data, leading to a loss of user confidence in the information displayed.

Other challenges also arose:

A time-consuming acceptance phase during which it was difficult to distinguish anomalies between data and RLS problems.
The maintenance of complex rules, amplifying the paradox in a project aiming to federate data and simplify users' vision.

Faced with this situation, we proposed a bold decision: to drastically simplify the RLS in order to improve understanding of the data. This generated a great deal of internal discussion about the confidentiality of certain sensitive information and required alignment between the different teams in charge of data.

The question we are still asking ourselves as we expand the BI platform is: how can we strike a balance between the dissemination of data and its confidentiality in the field of BI? This is a challenge that we continue to explore, seeking to create innovative solutions that preserve data integrity while making it accessible and understandable to users.

Business Intelligence, far from being a simple technical solution, is emerging as a crucial point of convergence between data distribution and data security.  How do you see this quest for balance?

An article written in collaboration with Florence VALLA BERTHOLET, Global Head of Data & Insights for the Luxury division at L'Oréal and Laurie JULLIEN for Equancy,