Ethical Considerations When Companies Study – And Fail to Study – Their Customers

Abstract

Experimentation has been a hallmark of science for 400 years, but only recently – since the advents of the computer and the Internet – have relatively quick, inexpensive experiments at scale become feasible for businesses. Today, the practice of companies studying their customers is ubiquitous. In 2011, Google ran 100–200 experiments per day on its products, services, algorithms, and designs. By 2016, it was running over 10,000 experiments per year. If you are a Bing user, on any given day, you are participating in some fifteen experiments out of the company’s 200 or so concurrent daily experiments, each of which exposes several million users to experimental conditions. And if you are a Facebook user, you are a subject in some ten experiments at any given time. Yet frameworks for thinking about the ethical conduct of research – long a mainstay of research throughout the academy and in industries whose research is subject to regulation by the Food and Drug Administration (FDA) – have lagged behind in most business contexts.

This chapter provides an overview of the different ways in which businesses study their customers, users, employees, and other stakeholders, and the reasons why they do so. It argues that a complete ethical analysis of business research requires consideration not only of the purpose, nature, and effects of research but also of the business’s alternative of not studying the effects of its products, services, and practices on stakeholders. Depending on various criteria, a particular business study – even one conducted without study-specific informed consent – can be unethical, ethically permissible, ethically laudable, or even obligatory. Although business research is now ubiquitous, the fact that individual, study-specific informed consent is usually infeasible in this context means that a careful consideration of a study’s risks and expected benefits is called for. As the chapter explains, the content of federal regulations that govern risk-benefit analyses of most academic and some industry research – the so-called Common Rule – is not easily translated to the business setting. But companies should consider adopting something like the process used by institutional reviews boards (IRBs) to prospectively review and oversee research, and the chapter provides some recommendations about how such company “research review boards” might operate.

Keywords: Research, Experiment, Consent, Common Rule, Ethics, IRB, Responsible Innovation, Evidence-Based Practice, Quality Improvement, Big Data

JEL Classification: K10, K23