Guest contributors
Business cases
Experts voice
Quick Poll
At a glance
PDF version

Subscribe to
Bull Direct:


n°18  |  September   2007
Guest contributors
“Combining technical and business expertise”
Interview with Gilles Chauveau
IT Director of EDF’s Business and Domestic Customer division
The EDF Group, one of the leaders in Europe’s energy industry, is an integrated energy provider with activities covering all aspects of the sector: energy production, transport, distribution, trading and sales. With more than 40 million customers worldwide, EDF achieved a consolidated turnover of €58.9 billion in 2006.
As a systems integrator, Bull has been actively contributing to the implementation of Business Intelligence (BI) at EDF.

You have deployed an ambitious business intelligence solution. Can you tell us why?
We first started thinking about the question of business intelligence about six years ago. Given the fact that the energy market was being opened up to competition, it was essential that we could anticipate how our customer portfolio might evolve, and optimize our products and services accordingly. We wanted, above all, to refine our capability to segment and predict, in order to reflect consumption profiles. We also wanted to make it easier to control the performance of our marketing and communications activities.

So what BI architectures have you implemented to meet these needs?
We moved forward in stages. In the event, we deployed a centralized data warehouse for the enterprise and small business market – which served as a test case – and then, more recently, we implemented another for domestic and private customers (handling 26.9 million customers!). These data warehouses are fed by legacy applications, and increasingly by the various mySAP utilities management modules – particularly ISU and CRM – and divided into various datamarts for special processing or restitution tasks, such as customer profiling. For historical reasons, some of our applications are still shared between the production and energy distribution businesses, and this means we have to maintain a number of specific datamarts that are not connected to the central data warehouses for the time being. Eventually, however, the target architecture will of course be unified. We should achieve this in 2008. The overall data volumes involved are vast: around ten terabytes for enterprises and small businesses, and a dozen for domestic customers (going back two years).

How do you think you will go on to develop the solution in the future?
We have had excellent feedback from users –who are in the management control, marketing and sales functions – all the regions. To take things a stage further, in the future we would like to be able to provide operational managers with detailed restitutions using even more user-friendly tools. Reporting procedures need to be on line, and made very simple. That’s our priority for the next three years.

Following on from this experience what, in your view, are the best practices that you need to follow?
Above all, I think it is important to establish a unified database with a centralized approach to business intelligence, even though we have to recognize that there is no ‘pure and perfect’ model, and that there will always be 20% of cases that will require specific solutions. We also have to take care that performance indicators do not start to multiply: we want to select the ten key indicators, which will be really useful for each profile. All this, while of course trying to find the right compromise as to what is useful, and what the IT Department can reasonably provide! Because BI solutions can be complex and expensive, we always need to keep in mind the relationship between the cost and the value that’s delivered. We also need to continue working towards getting these solutions deployed by teams that can command both technical expertise and business know-how. This is a rare combination, and one that is difficult to maintain over time. At EDF, we have put in place a structure that capitalizes on these skills, which we call the ‘AGRC’, which acts as a business intelligence center of expertise. Finally, we have to monitor data quality. False data will yield false results. It is important to convince operational teams and managers of the importance of data collection and management. These are a vital part of the enterprise’s resources!

Contact  |  Site map  |  Legal  |  Privacy