What’s Behind Behavioural Data Science?
Using behavioural data to understand consumer preferences is not a new concept for marketers.
However, in the past marketing researchers have struggled with the reliability of their data; namely, survey data that is limited to a sample of respondents at one time, survey design flaws, incorrect analytical techniques, and much more.
Behavioural data, on the other hand, presents marketers with inexpensive, accurate data about what consumers actually do. But how does collecting behavioural data differ from behavioural data science? Let’s take a look.
What is Behavioural Data Science?
Behavioural data science, to better model and predict behaviour, combines techniques from the behavioural sciences (psychology, sociology, economics, etc.) with a plethora of computational approaches ranging from the computer sciences, data-centric engineering, statistics, and more.
All of these disciplines (in addition to many others) help inform behavioural data science to more sensibly understand human, algorithmic and systems behaviour in regards to increasing quantities of data.
Behavioural Data Science isn’t behavioural analytics
Behavioural data science isn’t only about uncovering insights into the behaviour of consumers (which is under the behavioural analytics umbrella – and often confused with behavioural data science). Behavioural data science considers the interactions between humans and technology and how humans and algorithms can harmoniously co-exist going forward. Some of these considerations include:
- How can behavioural data science measure and improve people’s welfare at scale?
- Can we use behavioural data science to better understand machine and algorithmic behaviour?
- How can understanding human behaviour help attain better economic and social outcomes?
Behavioural data science addresses theoretical and empirical challenges related to human behaviour through the development of innovative and impactful approaches. Not only does behavioural science help people to better understand why they make the decisions they make, but also how they can optimise their behaviour to achieve better economic and social outcomes.
The 3 strands of Behavioural Data Science
Behavioural data science takes into account three important elements: human behaviour, algorithmic behaviour, and lastly, systems behaviour. These strands each provide a range of methodological tools to better integrate human intelligence with artificial intelligence and beyond.
Human behaviour strand
This strand of behavioural data science demonstrates how methods used in psychology, behavioural science, and other soft sciences can be enriched by data science approaches to describe human behaviour utilising large datasets.
Algorithmic behaviour strand
The algorithmic behaviour strand combines a range of algorithmic approaches from the hard sciences (statistics, mathematics, etc.) which can be used to assess and predict behaviour. Similarly, this approach deals with machine behaviour and the behaviour of algorithms themselves, as they also display biases and behavioural regularities.
Systems behaviour strand
Lastly, the systems behaviour strand looks at methods that enable modelling complex systems, markets, networks, cultural differences, and culture in an assortment of conditions.
The future of Behavioural Data Science
At large, the field aspires to recognise how to immerse human values into the heart and behaviour of AI systems. In this way, as AI systems continue to advance, humans will retain the ability to verify their integrity, answerability, and system resilience. This will ensure that the AI systems and their data will operate under successful, digitally-powered yet human-centric communities.
Behavioural Data Science not only has the potential to revolutionise our predictive capabilities but can also enable the secure flow of information and power between institutions across the globe. The science hopes to certify that AI systems will elevate, rather than threaten, our individual and collective prosperity.
