What are the technical and strategic issues of data exploitation in business?
Data is the new black gold for businesses, a new asset that needs to be integrated, understood and protected. Beyond data exploitation, a number of essential steps need to be taken into account. While company management often talks about return on investment, they don’t necessarily ask the right questions: what data are we talking about, how do we collect, store and exploit it, what value can it have and what ethical and legal issues are related to it?
How to collect and qualify data based on uses and objectives
Dawex offers a data marketplace where supply meets demand and where businesses choose the partners with whom they will exchange data. When we speak of data, we’re not referring only to personal data, but also of all those data that the company produces and that affect its businesses (industrial, product, distribution, supply chain, R&D, marketing and sales). Data exploitation within companies began to be a topic of discussion with the emergence of Big Data.
Recently, data have acquired value outside of the company as well, within its ecosystem. It is important that management consider how data are collected and stored, how they are logged, and what can be put at the disposal of users.
Data could save lives, for example. In the health field, data are multi-source, coming from individuals, patients or health professionals (pharmacies, general practitioners, specialists and hospitals). These sources are highly structured and meet specific protocols including the notion of consent, among other things.
However, the same sector uses completely heterogeneous data from health applications or connected objects. These data are structured differently, or come from health professionals who use a certain level of structure and medical vocabulary.
To exploit these data for analysis purposes in health, it is important to use comparable information. In other words, the same language, ontology and vocabulary have to be used in order to compare individuals. This means that work needs to be done when accessing and collecting information in order to be able to truly exploit the data.
This exploitation can be for purposes other than monetary ones, because the data improve care or help to anticipate certain cases. Even if this does not systematically save lives, it could still improve patients’ quality of life. So, the medical sector faces challenges on all levels in accepting data: regulatory, ethical and technological.
How is the value of data defined?
Data represent a relatively new asset. Just ten or fifteen years ago, it would have been difficult to identify the data we are talking about today, and there are now a number of questions about the exploitation of data:
- Can data be used from a technical and legal viewpoint, in particular in the area of health? Personal data has to be collected in a consistent manner, and the people concerned must be informed of why they are being collected. It is also necessary to be able to explain the history of a data set from a technical viewpoint. These problems also arise for other types of data, generated by industry.
- What do these data refer to? In health, data possess elements of value and sensitivity that are not addressed in the same way. Furthermore, revealing a person’s state of health would be problematic from a legal standpoint. Beyond violating medical secrecy, a data breach revealing this information would also be detrimental to a company’s reputation.
- What does the data set look like: is it homogeneous and does it provide relevant information?
So, companies are dealing with an object whose use is not all that simple, since it has to take these three dimensions into account.
What is the economic potential of the data?
Companies in a data exploitation process can generate revenues that are not merely marginal because there are elements of value related to the technical or legal characteristics of the data set. So, their data have very real economic potential.
However, ROI can suffer when personal data are processed poorly. The same is true when data are irrelevant or when they contain biases, which can be very costly.
Data having elements relevant to the functioning of a market are more expensive than a smaller data set, or one concerning a less relevant sector of information. Moreover, data are also of interest in data science, since their structure can contain information and opportunities for exploitation.
We can also adopt a macro-economic approach when considering data’s economic potential and ask the following question: like trademarks, do data constitute a factor improving performance of the combination of capital and labour?
About data protection
We find ourselves today in a prehistoric data period where data sets have value, but arbitrations do not exist like on other markets. However, regulation won’t be long in coming.
Last year, the European Commission suspended a draft regulation on non-personal data, because companies were already busy applying the GDPR. Nevertheless, it issued guidelines on the exchange of non-personal data. So, this market will soon be regulated.
What’s more, many are demanding the protection of sensitive personal data. It is important to find a middle ground between data protection and usefulness, because excessive protection would hinder risk taking that’s essential to innovation.
This is an important subject in health: when there is a large volume of data that could be used to analyse trends, it is not always necessary to have perfect data, and satisfactory results can be obtained through increasingly intelligent tools. From this type of data, we obtain analytical and descriptive – or even predictive – value as regards the health of large populations, on the epidemic scale.
Here, the concept of protection becomes less important, because the sector needs fewer sensitive data to arrive at a pertinent analysis. In a context of innovation or of analytical weight on specific and precise conclusions (in health, in life or death situations), the concepts of data origin, qualification and transformation become all the more important.
Michèle Arnoe believes that "The GDPR fits in this concept [risk-benefit] and it’s a culture shock for some. It’s great, because it opens so many possible opportunities, but there needs to be oversight.”
Last, the monetisation of data is one way of taking control back from the GAFAM, by imposing European values. The GAFAM’s economic models were created at a time without regulations. This is an advantage for European companies who now understand the legal frameworks and regulation.
However, before creating a data "exchange", many data transactions will have to be recorded and prices projected. Sustainable models related to data exploitation will have to be set up.
Speakers: Paul-Olivier Gibert, AFCDP; Laurant Lafaye, Dawex and Michèle Arnoe, IQVIA