Antoine Jeanjean, contact@opt2a.com

To use the expression used by our Niortais friends from Kereon Intelligence, who recently came to give a (very good) presentation in Bordeaux about Data Literacy: 'data can now be compared to gold'! Data has become so important to companies, organizations, administrations and governments that data is the real black gold of the 21st century.

Cet article a été publié sur Linkedin en Français le 28 Octobre 2019 - Tous droits réservés



And like any gold rush, we are witnessing a real 'increase in pressure' on these subjects related to Data or AI:

• - The media is responsible for publishing numerous daily articles that mix notions of AI, Statistics, Mathematics, ... this in turn misleads the reader (who is usually uneducated on the topic) by approximations or dubious shortcuts;
• - The startup ecosystem has helped to create a funding bubble, especially using the term “Artificial Intelligence” in order to get large amounts of funding. This is sometimes done without having a revolutionary autonomous expert system but only simple linear regression. This discredits the discipline for companies which do, in fact, offer these systems but do not have the same visibility;
• - The training industry is also concerned and some doubtful private training programs which aim to earn large profits are given by profiles who represent themselves as statistical experts or mathematicians. These companies and individuals promise earnings of 50k in a few months to people who know very little or nothing on the subject;
• - The consulting market extends its offer to these new businesses and creates new AI offers without having the associated consultants for the mission. This is because the term “AI” is fashionable and easy to sell to managers who do not understand these topics;
• - Complex and unsuitable new software or platforms are sold for millions of euros because, once again, the managers who buy them do not understand the ins and outs of these huge projects;

We are rushing through these complex subjects which require a long-term strategy and the appropriate resources.

We get to the heart of the problem: training, which is key to the success of these data management projects. Solid understanding of these data subjects and a good mastery of data flows will allow you to:

• - know the information of your organization (types of data, freshness, granularity,...)
• - locate information (data mapping, data warehouses location...)
• - secure information (user rights, GDPR, data governance policy,...)
• - make the information available (unused data, data organization issues…)
• - use information (with decision-support tools )
• - calculate new information (forecasts, KPI, rich data)
• - ultimately make good decisions USING your data. Setting up efficient data processes will help your organization to optimize the potential of this 'gold in stock'.

What if we managed data stocks, such as our physical stocks, in order to catalyze change?

Currently, the data is not yet valued in the balance sheet of a company where the stock of goods nevertheless appears. We can expect that the data stock will one day (soon) be valued for accounting purposes by the finance departments. But they have often not yet fully understood the underlying potential that an accounting valuation of each byte stored in the database would entail. For example, X records of such a data table are valued at Y € in the annual balance sheet, according to its valuation potential (scarcity, complexity, ...). At the end of the year, the company considers all of its data stocks and obtains a complete valuation of its warehouses, which it can then put on the balance sheet.

When such a change has taken place (a story of a few years), we will then see a turning point in the valuation of data in companies, because, like a physical stock:

• - If data is stolen ==> financial risk for the activity;
• - If data is of poor quality ==> financial risk for the activity;
• - If data is out of date ==> financial risk for the activity;
• - If a company is purchased ==> a full audit of the integrated data will be necessary.

The Operations Director (generally in charge of the physical stock) will have to either coordinate with the Data Director or take the role of lead for the management and valuation of digital data stock.

Because the following financial stakes will be enormous:

• - Specialized companies (data marketplace) are starting to emerge so that you can resell your data, even if it is not the core business of your activity. For example, the promising French startup Dawex;
• - You will have to learn to tidy up / structure / organize / make available and improve your data over time, under penalty of not respecting the accounting standards that will be created. This variable valuation of your 'data assets' will be no doubt also taken into account during fundraising or IPOs and the auditors will undoubtedly see something to complain about;
• - Your objective will then be to maximize the value of these new stocks and this will sometimes go as far as making you change the way you run your business. We even see companies that sell a service at a loss, in order to then resell the data associated with the service they have made available to you. The valuation of the data is then integrated into the business model. The more customers they have, the more data they have and the more relevant and marketable data they will then have, justifying the initial investment in the initial product or service;
• - We can therefore see that we are at the beginning of a new era concerning this subject 'valuation of data' which is no longer only the prerogative of Data Lab or isolated team. It is a subject that will take root in the very strategy of organizations. A key point for managers in all fields: learn about these complex subjects, document yourself and do not stand still, because changes will be profound in all departments of your organization.

This article follows my participation at Café PIGMA on October 10, 2019 in Bordeaux. To find out more about this day and find all the speaker videos, click here. If you are interested in Data / AI / Operational Research subjects, the Digital Aquitaine Association organizes monthly meetups. Find the full program on the association's website: www.iadatascience.fr. The Nouvelle Aquitaine annual conference is named Dataquitaine: www.dataquitaine.com/

Antoine Jeanjean, contact@opt2a.com

Antoine Jeanjean (PhD)