Data governance has more or less pervaded companies of the business world. This is not self-evident, because while it may seem easy to collect data today, understanding it and using it to the best is a challenge that fully justifies the role of the CDO.
We have entered for a long time a data society. As early as the 17th century, the scientific approach instituted the obligation to verify physical laws by applying them to the numbers produced by the measurement of the phenomena they govern. The following centuries saw, with the development of statistics, the generalization of this approach to social and economic phenomena. The present era has succumbed to the promise that an analysis or prediction, based on a measured and therefore objective fact, would have a truth value superior to intuition, reputed to be misleading. Who among us does not own a watch or a connected scale? Companies are overwhelmed by numbers and statistics to justify every decision they make. In his book “La gouvernance par les nombres” (Governance by numbers), Alain Supiot gives us a striking analysis that shows the extent of the historical, social and cultural mechanisms at work.
The ease with which current artificial intelligence technologies produce the right decisions from a mass of numbers, digital data, is unexpected and impressive. It seems capable of propelling humanity beyond the history it has lived until now.
It’s all playing out on the business front. Data seems familiar, yet in some ways it is elusive. Take, for example, Customer identity data. Nothing more common. And yet, at the very least, you need the last name, first name, date of birth and city of birth to minimize the chances of having namesakes. Under the RGPD, the date and city of birth are not justified by the purpose of use, unlike the delivery or postal address, which could allow, in compensation, to distinguish namesakes. In some cases, the cell phone proves to be a more reliable criterion when it is not a prepaid phone card that changes number with each renewal.
More complex data, such as measuring the performance of a business activity, is subject to comparable difficulties. Take, for example, a company that tracks the Margin of a product per customer, which is very common since the Margin determines the company’s ability to invest and grow in the future. It deducts from the selling price, all the costs incurred by the purchase or manufacture of the product, those related to the delivery or provision of the product, and those related to the acquisition and maintenance of the relationship with the customer.
If the calculation model seems clear, the selling price fluctuates, depending on the channels, the seller’s negotiation margin, and the weightings applied to the indicator itself, which is generally an average, and therefore subject to uncertainty. The same is true for costs, notably the cost of acquiring the customer and the costs of managing the relationship. To cope with this, the company conducts detailed cost studies every 2 or 3 years, which produce estimated costs, known as standard costs, in order to manage its margin. All these data are estimates which, at a given moment, for a given customer, can differ significantly from reality. On the other hand, the weightings, which always express a conscious or unconscious intention, can penalize certain products and benefit others.
Data-driven corporate governance is therefore far from being an exact science. The role of the CDO is not only to implement data processing technologies, but also to understand the gaps between the objectives pursued by the company and the uses it makes of its data, the impacts on employees and on products, and to help remedy the drawbacks.