In this regard, the purpose is an essential element that will determine the level of expectation with regard to the data. However, at the organizational level, needs can vary considerably: from human resources management to accounting, marketing, and business intelligence. Data processing takes place at many levels, without necessarily involving artificial intelligence.
Identify the data and the issues that concern it
Based on the purpose, the next step is to identify the useful data. This may include data relating to customers, suppliers, employees, or other individuals linked to business performance or processes. Beyond that, it is important to consider the issues in light of the various areas of data management knowledge, as defined by the DAMA wheel: quality, architecture, security, modeling, integration and interoperability, document management, repositories, etc.
A step-by-step approach
That said, it is understandable that implementing optimal data governance can be complex. It is not a question of tackling all these dimensions at once, but rather proceeding in stages. Once the scope of useful data has been defined, we will work on one aspect of governance with a view to demonstrating the relevance of the approach.
“This iterative approach, with concrete deliverables, allows us to move forward based on real issues, without exhausting ourselves,” explains Nicolas Vivarelli. By focusing on the essentials and gradually addressing new aspects of governance, we can demonstrate the relevance of the approach while preserving the energy needed to generate more value from the data. "
The first steps in implementing governance may be dictated by needs that vary greatly from one project to another.
“In some contexts, regulatory requirements will necessitate focusing efforts on security and quality. In others, sharing and interoperability issues will be priorities,” says Rémi Dusaud.
Empowering all teams
Data governance is not the sole responsibility of IT or a department in charge of data management. It requires a cross-functional approach across the organization, with clearly defined roles and responsibilities. “It is important to be able to track data from the moment it is entered until it is processed,” explains Rémi Dusaud. “When it comes to customer data, it is essential that the person who enters it initially a salesperson in a store, for example is aware of how it will be used. They must understand that missing or incorrect data will have repercussions elsewhere.”
In many stores, for example, sales assistants neglect certain fields when registering a customer, such as their date of birth. “In several organizations we work with, we have seen a surge in customers who are supposed to turn 125 this year,” Nicolas Vivarelli notes ironically. Not specifying the customer's year of birth can have consequences elsewhere: in marketing communications, customer analysis, and so on.
Embed governance in the corporate culture
Implementing data governance is not a one-off project or program. It must be embedded in the corporate culture.
“Good data management, by everyone according to their responsibilities, must be valued. Expectations in this area should be detailed in job descriptions, and it should not be considered a secondary task,” explains Nicolas Vivarelli. “Otherwise, employees will only pay attention to data when they find the time to do so... which is usually never.”
A continuous improvement approach
While these aspects of data management and governance must be supervised as a whole, each employee must be made accountable for the areas of data management relevant to them within the governance chain. By becoming aware of the value of data, based on a better understanding of how it is used, they will naturally be able to contribute to improving its management.
“In this way, the company will gradually gain maturity, facilitating the deployment of data-driven projects to create more value,” concludes Nicolas Vivarelli.
De cette manière, l’entreprise pourra progressivement gagner en maturité, pour un déploiement facilité des projets s’appuyant sur la donnée, afin de créer davantage de valeur
