Talking about data and behavioral sciences at NudgeFest

Last week I was invited to participate in the second edition of NudgeFest , a conference on Behavioral Sciences. The panel was about the mixture of data and behavior and I was lucky to share stage with Emma Bernardo Sampedro , Head of Marketing at Neovantas , Juan de Rus Gutiérrez , partner at the same company, and Victoria Valle Lara , Behavioral Economist at Sandoz .

We talked about many things, and I wanted to summarize part of the conversation based on the notes I took while we were talking.

We firstly discussed about the definition and meaning of behavioral data. It is basically information we collect about people's actions and their behavioral patterns. This can include everything from how we interact with apps on our phone, to our shopping habits, and even the way we browse the internet. The interesting thing about this data is that it offers us a window into what we really do, not just what we say we do. For example, instead of relying on someone to tell us how much they exercise, we could look at how often they use a fitness app or activity tracker. Behavioral data is very valuable because it can help design better products, services and public policies, adapting them to how people act in reality and not how we believe or wish they would act.

Behavioral data is information we collect about people's actions and their behavioral patterns.

An important point we established is that we need to be aware that most of the times we are not looking at the real behavior but to proxies, data points that indirectly tell us about a person's or group's behavior.

It could seem obvious to many of you, but I must say there is a significant difference between behavioral data and traditional data. Traditional data typically refers to demographic information, such as age, gender, educational level, and geographic location. This data gives us a static view of who a person is in sociodemographic terms.

Most of the times we are not looking at the real behavior but to proxies, data points that indirectly tell us about a person's or group's behavior.

On the other hand, behavioral data focuses on the “how” and “when”: how we interact with our environment, how we make decisions, how we use products and services, and how our habits and preferences change over time. They are dynamic and can offer deeper insights into our motivations, desires and real behaviors.

While traditional data can tell us who is most likely to buy a product (for example, women aged 30 to 40), behavioral data can tell us why they buy it, when they are most likely to buy it, what motivates them to choose one product over another, and how that behavior can change under different circumstances. This allows companies, governments and other organizations to design much more personalized and effective strategies to achieve their objectives.

However, there is a fine line between both traditional and behavioral data so take all this with a pinch of salt!

As I mentioned above, behavioral data represents a powerful tool for narrowing the gap between what people say they will do and what they actually do, commonly known as the “intention-behavior gap.” This discrepancy occurs because our actions are often influenced by factors that we are not even fully aware of, such as cognitive biases, social influences, or even the design of our physical and digital environment. Behavioral data, by capturing our actual actions rather than our stated intentions, provides a more accurate and authoritative view of our behavior. This has several important practical applications:

• Product and service design: Understanding real user behavior can lead to the development of more intuitive and easy-to-use products and services that better align with people's real needs and habits or habits-to-be.
• Behavior change interventions: In public health, education, and social welfare, programs designed based on behavioral data may be more effective in promoting desirable behaviors, such as increasing physical activity or improving our diet.
• Public policies and regulations: Governments can use behavioral data to design more effective, evidence-based public policies that achieve their objectives without imposing unnecessary burdens on citizens.
• Marketing and advertising: Marketing based on behavioral data can help companies deliver more personalized and relevant messages, improving both the customer experience and the effectiveness of their campaigns.

Behavioral data represents a powerful tool for narrowing the gap between what people say they will do and what they actually do!

An example: Quantified Reading

Emma asked me about one of the examples I describe at my book, "En la mente del usuario" (in Spanish).

In the educational context, tracking whether a student has accessed a resource or how much time they have spent within an ebook is just the tip of the iceberg in terms of behavioral data. These pieces of data, while useful, offer a very limited view of student engagement and interaction with the material. We are missing a series of richer and deeper dimensions of learning behavior, which could offer much more valuable insights to improve education. Some of these include:

• Interactions within the content: Not only whether they opened the book, but which parts they read most carefully, which parts they review several times, and which parts they skip. This can indicate what content they find most interesting, challenging or relevant.
• Patterns of annotation and underlining: The notes that students take or what they underline can reveal which concepts they consider important or which they find difficult to understand.
• Related questions and searches: The queries you make in search engines or databases while studying can offer a window into your thought processes, doubts, and areas of additional interest.
• Social interactions around content: Discussions in forums, study groups, or even on social networks about course material can show how students process information and discuss it with others, which is crucial for collaborative learning.
• Study sequence and pace: How you alternate between different resources, what time of day you prefer to study, and how you allocate your study time before an assessment can offer clues about your study habits and effectiveness.

What is this data used for? How can teachers, authors or publishers take advantage of it?

• Personalization of learning: We can adapt content, challenges and support to the individual needs of students, based on their actual behavior, not generic assumptions.
• Improving educational material: Educators and content creators can identify which parts of the material are most effective and which need to be improved or explained in another way.
• Early detection of difficulties: By observing early signs of struggle or disinterest, educators can intervene before students fall too far behind.
• Encouraging effective study strategies: Understanding successful study patterns can help teach other students these strategies.
• Designing fairer and more representative assessments: By understanding how students interact with the material, assessments can be designed to better measure what they have actually learned.

The final part of the conversation went about how AI affects all this conversation. This is something I already wrote about in my linkedin profile.

It was a great conference, full of many insights and practical information I didn't know about (such as Google's intriguing books about behavior).

Thanks to Oto Whitehead and Emma Bernando for the invitation and the organization of this great event!