Ethical Decision-making in Data Mining: Try the Rights Approach

6 min readJun 22, 2020

Introduction

Companies are collecting as much data as they can to make smarter business decisions and provide better products and services to their customers. And they should. But sooner or later ethical questions arise on what data should be collected and how it should be used and for what purpose. A framework for ethical decision-making can guide leaders in the process of developing the right data-mining principles.

As it turns out, there are several frameworks that can be applied. In this article, I argue that the Rights Approach to ethical decision-making is the best framework for businesses working with customer or user data.

I am using the Royal Dutch Football Association as a case study. By not thinking carefully through the ethics of their data mining practices, they turned from a football association into a marketing database.

KNVB: From a Football Association to a Marketing Database

The UEFA Euro 1988 final: the football[1] match that showed the world the talent of giants such as Marco van Basten and Ruud Gullit. The “Orange team” defeated the Soviet Union, in what would be that nation’s last European Championship (the Soviet Union collapsed two years later). Although the Dutch never reached the same high, The Netherlands was turned into a football nation.

In the decades that followed, amateur football clubs blossomed, with 2,986 clubs[2] counted in 2017 (equal to ~0.4 club per city or town)[3]. All these clubs are governed by the Koninklijke Nederlandse Voetbalbond (The Royal Dutch Football Association, or “KNVB”), the largest sports association in the Netherlands with 1.2 million members (7% of the population).

The KNVB has a reputation for being an innovator and is one of the leading forces behind the government-sponsored Sports Innovator Program.[4] But innovation requires investment, and after disappointing results in international championships in 2002 and 2004, the organization was operating at a loss, driven by a rapid decline in sponsor revenues. Management was under pressure to turn things around, and it did not take long before someone suggested that the answer may lie in its members. More specifically, in the data on its members.

Starting in 2004, the KNVB set out to collect as much data on its members as it could, to build a rich database including demographics, contact information, and transaction data. Although the KNVB’s member base counts about 1.2 million members, its database includes 3.6 million people or roughly 20% of the Dutch population.

In 2014, the KNVB hired three specialized data consultancies — 40beats, SAS, and Crystalloids — to optimize its database for marketing purposes. Member data was being combined with third-party data such as census data collected by the Dutch governmental institution ‘Central Bureau of Statistics’ (CBS), to make personas even more complete.

It turns out that the KNVB commercialized its data by selling it to companies like Heineken, The National Lottery, Nike, Coca-Cola, and KPN (a Dutch Telco), who used it for digital and direct marketing campaigns.[5] Teenage KNVB members report being called by advertisers (using hidden numbers) trying to sell lottery tickets. Encouraging teens to gamble seems hardly ethical and far off from the raison d’être (reason of existence) of a sports association.

As a result, this endeavor now risks turning into a privacy scandal and PR nightmare.

What approach to ethical decision-making in this user-data mining context should the KNVB have taken?

Five Approaches to Ethical Decision-Making

The Silicon Valley-based Markkula Center for Applied Ethics (part of Santa Clara University), describes five popular approaches to ethical decision-making: the utilitarian approach, the fairness or justice approach, the common-good approach, the virtue approach, and the rights approach.[6]

The utilitarian approach

The utilitarian approach was first fully articulated in the 19th century by English philosophers Jeremy Bentham and John Stuart Mill.[7] The focus of this approach is on the outcome and tells the decision-maker to choose the action that yields the greatest well-being of the greatest number of people. In the case of the KNVB, it is not clear that this approach would have changed the course of action. Paul Decossaux, KNVB’s Commercial Director, explained they operated on the belief that these data contracts would ultimately provide its members with more relevant promotions addressing their needs. (A view that sounds familiar to the view shared by another organization that is consistently in the user-privacy spotlights: Facebook.)

The fairness or justice approach

It was Greek philosopher Aristotle who laid the groundwork for the fairness or justice approach. The key question in this framework is: How fair is an action and is everyone treated in the same way? [6] But the question of equal treatment of people hardly seems the right (or at least most important) question to ask in a user-data mining context. If you treat all users in the same bad way, it does not make it an ethically acceptable action.

The common-good approach

The common-good approach is another view pioneered by the Greeks. It was Plato who was at the roots of the framework, later to be fine-tuned by Aristotle and modernized by French philosopher Jean-Jacques Rousseau.[8] This approach stresses that decision-makers ought to be guided by what is best for the people as a whole and should protect the vulnerable. Although noble, this approach is unlikely to have changed the actions of the KNVB. Taking another example, Facebook: a specific belief about what is best for the people as a whole, such as a global social network, is more likely to permit actions without much scrutiny, as long as the decision-maker beliefs it contributes to their higher goal of the common good. It, therefore, disqualifies as a good approach in a user-data mining context.

The virtue approach

The virtue approach describes that actions should be governed by a set of excellent traits, such as honesty, courage, compassion, and integrity.[9] As such, this framework embeds ethical decision-making in a layer of good intentions. In the KNVB case, there are no reasons to question the good intentions of the decision-makers in its organization. Nonetheless, its actions are ethically questionable. The virtue approach, therefore, disqualifies as an appropriate framework for the assessment of ethical issues in the context of user-data mining.

The rights approach

Eighteenth-century German philosopher Emmanuel Kant focused on individual rights and the right to choose for oneself. His approach to addressing ethical questions came to be known as the rights approach and is centered around the idea that “People are not objects to be manipulated; it is a violation of human dignity to use people in ways they do not freely choose”.[6] I would like to argue that this is the approach organizations using user data for commercial ends should use to guide their ethical decision-making process.

Why the Rights Approach is the Right Approach

There are two main reasons why I think the rights approach is the best framework for ethical decision-making for businesses.

First, the rights approach focuses heavily on the individual’s rights. In our digital world, we tend to talk about metadata and forget that there are actual human beings who are influenced by data mining decisions. In the user-data mining context, marketers aim to predict which users are most likely to buy a product or service and allocate marketing efforts accordingly. Datasets contain millions of records, and decision-makers like to think of this as “big data”. Reality is, however, that predictions are made on a user-level, thereby potentially violating individual user’s rights.

Secondly, the approach implies that it is unethical to use people’s data in ways they did not freely choose. The KNVB should have asked users for explicit consent to sell their data to commercial parties. Misuse of data would be a reason for employees to speak up and for privacy authorities to assess violation of privacy.

This is exactly what is the privacy watchdogs are currently doing in the Netherlands.

You can read this and other stories on www.thalein.com