By Aurélien Gour, Data and Digital Partner, Micropole Group.
Many people are unaware that the key to success in many digital transformation projects is data management. Today, different businesses invest large sums of money expecting concrete results as soon as possible, rather than putting data under control through appropriate governance. It is clear that many projects that are already well advanced are confronted with data quality issues and difficulties in identifying the key people (those responsible for this data and able to validate that this information is reliable and usable). The overall observation is that the lack of control over this data impacts not only IT, which is exhausted trying to interpret and rectify it, but also the business, which sees these projects slip, or the expected qualitative objectives become distant. So what are the keys to ensuring data control and therefore the success of a digital transformation project?
First of all, let's take a look at some of the pitfalls frequently observed in this type of project, using the example of Hadoop platform implementations:
- Lack of traceability of the sources of information injected into a Data Lake.
- Lack of clarification of the governance to be put in place around the Data Lake. This point is often dealt with too late. In any case, it is important to define the stakeholders and processes to be put in place to enrich, manipulate and access the data in the Data Lake.
- The rush to deliver MVPs (Minimum Viable Products) : companies want to see products quickly, so as to launch their industrialization. However, problems linked to data reliability, knowledge and traceability mean that the transition from POC mode to industrialization is often problematic.
So how do we deal with these pitfalls?
1 - CONSIDER THE BENEFITS OF DATA GOVERNANCE AT THE EARLY STAGE OF THE PROJECT
Faced with the difficulty of mastering the sources of data available to customers, the problem often arises of identifying one or more people responsible for data within the company. Hence the need to ask the right questions upstream:
- How can these people be identified?
- How can the right people be made aware of the benefits of data governance, so that they can instill it in their teams?
- And finally, what governance strategy should be implemented to ensure that projects benefit from it, as in the case of the Data Lake, for example?
Indeed, in a Data Lake environment, a lot of information and raw data (texts, videos, images, Excel) are stored. But if this "big" storage reservoir is at the heart of the project, only a data governance will effectively allow to :
- Accelerate all digital initiatives: all digital projects need data close to production, so that they can move from "prototype" mode to "industrialization" as quickly and easily as possible, in order to make new products available. The aim ofData Governance is to ensure that data is reliable, comes from the right sources, and is easily interpreted and understood by project and business teams.
- Master the knowledge around data: to share a common language on data, a base of knowledge and know-how (MOE, MOA), then to be able to characterize the service commitments with respect to the data and the various actors. This type of project also allows to control the propagation of data, from its constitution (collection, calculation) to its use (traceability).
- Manage the risks related to data subject to regulation: the objective here is to ensure that the uses of the data comply with the company's obligations (personal data, regulatory reporting, etc.). Hence the need to qualify the level of sensitivity of the data, to classify it and to define the requirements in terms of security.
- Managing data quality and developing the right tools: the challenge is twofold: firstly, you need to identify the right people to define all the quality rules needed to establish data reliability; then you need to identify where in the IS it is most appropriate to implement tools to monitor and industrialize governance. This makes it possible to guarantee the consistency, relevance and reliability of data, and to assess and manage its quality. Managing data is definitely a guarantee of quality, relevance and consistency within the information system.
By covering all of these benefits, all the means will have been put in place to successfully carry out the digital transformation project. And this will benefit all IS stakeholders.
2 -GOOD PRACTICES TO MAKE DATA GOVERNANCE COME TO LIFE
Another finding of digital transformation is that data governance is all too often confused with project governance. This is a mistake that should be avoided, as data governance actually concerns all the company's data, far beyond the projects that use this data.
Furthermore, it is common that when a data governance committee is set up, companies only consider data through the lens of whatever projects they are working on. These committees thus set up operational and strategic objectives, whereas the main goal of governance is to make data reliable wherever it is. Governance cannot be reduced to a strictly "project" vision. If, on the other hand, the company wants to become Data Centric, then data governance must be clearly placed at the center of the company's strategy and teams must be motivated to learn how to have a transverse vision of their data.
Finally, for the proper implementation of this strategy, don't forget about change management! It is important to first identify the few stakeholders who will trigger this data governance and whose objective will be to instill a change of mindset within the organization. When a company wants to move from a product-centric or customer-centric strategy to a data-centric strategy, the paradigm shift implies a real change in mindset. To conclude, digital transformation means new offers, better knowledge of data and changes in mentality. To guarantee the success of projects, data governance must now be an integral part of the company's strategy.
