Keynote Open : And if the data reconciled us?

Dans nos entreprises (comme dans la société en général) la data est un vecteur méconnu de réconciliation. Comment nous organiser humainement, techniquement et scientifiquement pour devenir data-driven ?

How do data bring together IT and BUs?

Four years ago, when he joined Open, Stéphane Messika realised that scientific, expertise and technical divisions had to be combined in order to be able to analyse, understand and use data. However, these entities have different cultures within the company. So, which operational model can businesses use to better exploit data? This ReadyForIT conference, featuring Natixis and the Engie group, offered some answers.

Which operational model do Engie and Natixis use to exploit data?

At this conference, Guiraude Lame spoke for the bank Natixis, and Gérard Guinamand on behalf of Engie.

Which organisation should be adopted?

Guiraude Lame is responsible for Natixis’s “Data for Business” project, which is based on three components: traditional CTO governance, IT facilitation with infrastructures and data scientists, and a data business dimension where the challenge is to connect with BUs and their needs.

At Engie, Gérard Guinamand occupies a position with a systemic approach. He is director of the “Data at Engie” program aimed at making data a true company asset on par with factories, client portfolios and employees. His position is also based on three dimensions:

  • Strategy and governance: in concert with business unit (BU) CEOs to help them define a data strategy in line with their business strategy, and chief data officers in each BU
  • Scientific: with a community of data scientists throughout the group
  • Technical: the most centralised part because it led to the creation of a data technology ecosystem

How does the sector of activity transform data processing?

As a bank, Natixis’s first asset is data, because data are everywhere and stand for banking products. Guiraude Lame is responsible for helping his colleagues understand that they can go beyond regulatory compliance to increase the value of data through new products or new commercial approaches.

The data context is different, however, in an energy company like Engie. It processes data in a fragment of its activity. The group’s CEO is responsible for this subject; he is convinced that data will allow the group to transform the business model that is necessary today. This is because energy companies are moving from a role of producer to a service sale model.

Engie’s goal now is to teach and help customers to produce their own energy, consume it more wisely, conserve it, and sell or buy it. From the base assets of electrons or molecules, the group now focuses on data, which are becoming the new engine of its activity.

Which infrastructure should be used?

Natixis needs a data lake infrastructure in order to match, use and store data. This is a very long-term strategy that requires first moving forward on the infrastructure side before turning to the business aspect. But without the data lake, the bank would not have made progress, especially since many value-added use cases require matching data: massive collection and structuring work.

Engie is a very decentralised group. A “top-down” data lake on a single platform would not work in its case, because CEOs of the BUs would never agree to put their data in a group structure without knowing what will be done with them next. So, the company adopted a “bottom-up” approach where each entity creates its own local data hub allowing it to store its data in the cloud. Everything is both visible and accessible.

About governance

Banking regulations oblige Natixis to establish data governance to guarantee traceability from start to finish. Although the group initially adopted data governance simply to comply with a regulatory requirement, it has now put in place the entire infrastructure necessary to achieve this goal.

Engie’s content management system reveals use cases and describes them in the light of the value that is produced. The CMS allows the group to share use cases within itself in a logic of sharing and reuse. It is a management tool for the data community.

What methods should be used to evaluate use cases?

Natixis assists its business units in their data process by conceiving, qualifying and identifying the value of use cases. Guiraude Lame explains: “If the BUs don’t address these issues, and if BU people aren’t the ones steering these initiatives, […] they’ll be useless and won’t deliver any value.”

Gérard Guinamand is “convinced that to move forward, things have to exist, and be visible and measurable.” At Engie, two elements have been put in place to this end:

  • To reveal use cases, the group has created a data office in each BU led by a chief data officer. This department has to commit each quarter to producing a certain number of use cases. To do this, it uses the design thinking method, which helps people from many horizons to work differently.
  • The group has also designed a “data garage” in each BU to transform business use cases into algorithms and solutions.

Through these two elements, tools and expertise are centralised, and the search for uses is decentralised.

Speakers: Stéphane Messika, Open; Gérard Guinamand, Engie; and Guiraude Lame, Natixis