Business School Professor Sandra Matz Explores the Implications of Psychological Targeting

In an age where our digital interactions reveal more than we think, understanding how our data is used to influence behavior is essential. On October 15, Columbia Business School Professor Sandra Matz gave an eye-opening talk at the Santiago Center on the emerging field of psychological targeting, a discipline that reveals how our digital footprints expose intimate aspects of our psychology and can be used to shape decisions—from what we buy to how we vote.

The event was opened by the Center’s director, Antonio Campaña, who welcomed the audience with an overview of the Center’s mission to foster cross-border knowledge exchange. He then introduced Francisca Lund (CBS’15), a member of the Columbia Business School Alumni Association of Chile — co-sponsors of the event — who provided insights into Professor Matz’s research. Matz, a social scientist with a background in psychology and the social sciences, has spent the last decade exploring the intersection between psychology and digital footprints to help individuals and businesses make more informed and ethical decisions. Lund also highlighted Matz’s forthcoming book, "Mindmasters: The Data-Driven Science of Predicting and Changing Human Behavior", which synthesizes over a decade of research and is set to be released in January 2025.

A Deep Dive into Psychological Targeting

Matz’s presentation began with a personal anecdote about growing up in a small village in southern Germany, where neighbors knew almost everything about each other. She compared this experience to today’s digital world, where although it may seem that privacy is maintained, data traces left through digital interactions provide anyone with access to that data a remarkably detailed view of an individual’s life. “It's not data tracing we’re talking about, but the inferences we can make about a person from it,” she explained.

Matz uses the term psychological targeting to describe the process of inferring an individual’s psychology, personality, preferences, and behavior based on their digital data, and to what extent these insights can be used to change the way people think, feel and behave. “It’s not that we’re trying to peek into people’s psychology, it’s that often the goal is to push people in a certain direction —whether it’s what to buy, how to vote, etc.”

The Three Big Questions: How, So What, and What Now?

Matz structured her talk around three essential questions: How does psychological targeting work? What power does it give those who can tap into it? And what should we do about it?

In reference to the first point, she stated that by observing data from people’s smartphones, it can be determined how a person socializes with others, how often they go to bars and experience environments in which there are others around, and so on: “What someone does online, directly translates into something that is deeper in a psychological level.” Now, when it comes to how accurate are these predictions can be, she explained that data traces are incredibly intimate, “It’s not just a GPS record or something that you tweet —it really gives others information of who you are on the inside,” she assured.

While growing up in a close-knit village meant knowing a lot about neighbors, today's digital world allows for even greater—and more intimate—insights into people’s lives. Matz explained how the concept of psychological targeting has evolved with the rise of data-driven technologies, particularly Generative AI, which has revolutionized access to psychological profiling. She noted that while this used to be the domain of data scientists with huge datasets who built models with a limited scope, AI tools like ChatGPT basically allow anyone with access to social media profiles to make highly accurate psychological predictions. “This means that a 15 year old can predict the psychological profile of millions of people just by scrapping the web. This democratizes access, but is certainly a bit terrifying.”

Now, if we think about why all of this matters, Matz explained that it’s because of the power to influence behavior through personalization. For instance, she detailed how companies like Netflix and Amazon use personalized recommendations to shape user behavior. “Netflix invests over $150 million annually in recommendation systems, and over 75% of the content viewed on the platform comes from those recommendations. Similarly, 35% of Amazon’s sales are generated by its recommendation engine,” Matz noted, emphasizing how personalization has become a powerful tool for business success. “They spend a fortune in personalized recommendations and make a fortune based on them,” she closed.

Ethical Considerations and Future Implications

As Matz shifted to the ethical implications of psychological targeting, she raised critical concerns about privacy, agency, and manipulation. “It’s not just about losing privacy, we’re also losing agency and self-determination, because the choices that we make are no longer really our own if someone is pulling the strings behind our back. We see this playing out in different contexts,” she remarked, referring to high-profile incidents like the social credit system in China and the Cambridge Analytica scandal. About the latter, she said: “its influence in the 2016 presidential election in the United States, is another example of how peeking into people’s psychology and potentially shifting their behavior is not just affecting one single individual; it might affect the entire course of society.”

Despite the dark side of this technology, Matz remains optimistic about its potential to do good. She referenced technology historian Melvin Kranzberg, known for developing a series of “Laws of Technology,” the first one of which says: “Technology is neither good nor bad; nor is it neutral.” “I think the last part is actually the really interesting one, because if it’s not neutral it really depends on us,” she added.  

Matz also discussed the possibility of using psychological targeting for early detection of mental health issues, citing her lab’s research into detecting signs of depression from smartphone data. For instance, if they see from someone’s GPS data that they’re no longer leaving the house, being physically active or taking as many phone calls, it could be nothing or it could be early signs of depression. “There’s a huge gap between people seeking therapy and those who can access it. Generative AI could help close that gap allowing to intervene sooner. At least for the people that don’t have access to a traditional therapist, there might be ways in which you can engage them at the right moment with the right treatment. So, I think there’s a lot of potential to use this in a positive way.”

Regulation and the Path Forward

Matz concluded her presentation by addressing the need for regulation to balance the benefits and risks of psychological targeting. She stressed the importance of shifting the responsibility for data protection from individuals to corporations through mechanisms such as stricter data privacy laws, default data protection settings, and potential “data taxes” on companies that collect excessive amounts of personal data. “It’s not costly to collect as much data as you want, even if you’re not using it. If there was something like a data tax, maybe companies would think twice whether they really need it,” she explained. However, she noted that although regulation minimizes risks of data abuse, it’s not necessarily helping people make the most of their data.

In closing, Matz stressed the need for innovative solutions that would optimize utility for individuals, such as data co-ops or data trusts, which would allow communities of people who have a shared interest in using their data, to manage their own data collectively and leverage its value in a more controlled and beneficial way. She explained that these co-ops already exist in Switzerland, where they work well helping patients with rare diseases that are not covered by pharma. Ultimately, it’s not just about minimizing risk, but also about helping people make the most of their data,” she concluded.

Watch the full video here.