13 February 2019
Mr. Franck ELLA, Engineer, graduated from the Ecole Centrale d'Electronique de Paris. Since 2016, he has been responsible for Data Systems at Axa Direct France, for France and Spain. He has nearly 20 years of experience in the field of data management (Relational Database Management Systems, Busines Intelligence, Big Data) and in various sectors: software publishing, energy, industry, logistics, e-Commerce, banking and insurance. He spent 8 years at Microsoft France as back office engineer and advisor on Data Technologies before joining the Cap Gemini Group as an expert consultant on Microsoft's flagship technologies in Database and Business Intelligence, with frequent interventions with many key account customers such as GDF SUEZ, LCL or Airbus. In 2011, he created his own IT services company called "AKOK Technologies", specializing in database and business intelligence expertise and architecture. During this five-year adventure, he worked for major companies in the Paris region such as BNP Paribas, Véolia, Geodis Calberson, Vente Privée and AXA Direct France. He is a member of DAMA France, French representative of DAMA International (The Global Data Management Community).
1: We are talking more and more about the outbreak of digital data or "Big data". What does this mean?
First, the context of Digital Data needs to be clarified. Indeed, all of the digital data produced since the beginning of time until the end of 2018, now corresponds to the mass of those generated every minute. About 90% of global data has been generated in recent years with the use of the Internet and social networks, and one estimate is that 1.7 megabytes of data will be generated every second by one person in 2020. This quantum explosion of digital data has forced the IT industry to find new orders of magnitude for capturing, searching, sharing, storing, analyzing, and presenting data. Thus was born the "Big Data" concept, the need for storing an unspeakable amount of information on a digital basis.
2: Is this mass of data useful for decision making? Should we not optimize the management of this data?
Several studies have shown that less than half of a firm's structured data is used in decision-making and less than 1% of unstructured data is analyzed or taken into account for decision making. In addition, over 70% of employees have access to data that they should not be aware of. Thus, besides the technological aspect, a major stake for companies is the establishment of an effective strategy of governance of all their data. This corresponds to the organizational set, the procedures and tools to be put in place in order to optimize the use of data at each stage of their life cycle. It is important to note that this governance should not be applied at the level of a product or service, but at the level of the enterprise.
From a practical point of view, data governance is based on the following five pillars:
1. The definition and evolution of the management rules.
2. The organization of the roles and responsibilities of the different actors.
3. Cost, value and legacy management.
4. Control, through the establishment of mechanisms, to ensure the proper application of the established rules.
5. Data risk assessment.
The Chief Data Officer is the supervisor of the implementation of data governance in the company. They are responsible for data management and must put in place the processes that guarantee their state, governance and origins: internal, external, private and public.
Effective data governance must incorporate a defensive strategy that includes:
• The integrity of data circulating in the enterprise, via authoritative data sources (single source of truth),
• Compliance with prevailing regulations (e.g. the General Data Protection Regulation, GDPR, in Europe) and their monitoring,
• The use of data analysis to detect and limit fraud and the implementation of a piracy prevention system.
In addition, effective data governance must include an offensive strategy that applies in particular to Big Data and focuses on:
• Assistance in achieving business objectives,
• Improving competitiveness, profitability and customer satisfaction by analyzing, modeling, transforming and enriching data, and facilitating decision-making using interactive dashboards,
• The merging of internal and market data, especially to create new products and services.
3: You talked about structured data and unstructured data in the context of Big Data. How does 'Big Data' fall into these two categories?
First, it should be noted that structured data is generally data stored in relational database management systems or in data warehouses, whereas unstructured data is generated by e-mail, documents (PowerPoint, Word, etc.) or via media such as images, audio or video files.
Big Data systems, such as Microsoft Data Lake, are built from structured data. However, the large number of unstructured data (about 80% of digital data) is of great interest both in their variety and in their volume and, although more difficult, their analysis can significantly increase customer knowledge. Therefore, more and more Big Data tools make it possible to use this type of data.
Implementing Big Data solutions, therefore, makes it possible to process these two different types of data.
4: Do these strategies apply to all sectors? Would it be possible to have sectoral specificities as in finance for example?
Depending on its industry, competitive and regulatory environment, the company will invest more in one strategy than another. Thus, in the medical sector in which is an environment where quality and data protection are paramount, the defensive approach will take precedence over the offensive approach. In the banking and insurance sectors, being highly regulated, thereby requiring strong data protection, the dynamism of the markets in which they operate will require a good balance of offensive and defensive approaches. The retail sector, which is a less regulated sector where personal data is less sensitive, will favor the offensive approach, in order to answer the important needs of responsiveness to the competition and to the evolution of the market.
5: Based on your experiences, could you share some economic benefits that would result from good data governance?
The economic benefits of good data governance are diverse. For instance, through a good defensive strategy, there can be a prevention of payment of fines that can reach huge amounts in case of violation of private data. For example, (4% of the group sales or 20 million euros, with regard to the GDPR). We can also cite the improvement of competitiveness, customer satisfaction and therefore sales through an effective and relevant offensive strategy.
6: For Africa, how do you see data governance, what are the implications for competitiveness?
The field of data governance is quite new, and very few companies can boast of having finalized the implementation of a true governance system integrating all the subsets defined in Data Management Body of Knowledge (DMBOK®) listed below by DAMA International. Africa still has the opportunity not to be left behind, with the same implications as the other continents – namely, the control, protection and profitability of its data.
The ultimate goal of data governance can be summed up by this quote from Carly Fiorina, former executive president, and chair of Hewlett-Packard Co: “The goal is to turn data into information, and information into insight.”