Quantitative Research: We are not as unpredictable as we think | DMA

Filter By

Show All

Connect to


Quantitative Research: We are not as unpredictable as we think


We all like to think that we act autonomously. Our consciousness – the inner narrative that makes us feel that we are free agents in the world – gives us the impression that we are in control of our actions, our thoughts and our beliefs. But as our understanding of human behaviour becomes more comprehensive, and the systems that aid this understanding more complex, it is more apparent that we aren’t as unpredictable as we think. This article will take a look at the notion of free will, machine learning and quantitative research and how the demonstrable predictability of humans can be implemented in your marketing strategies.

In 1983, psychologist Benjamin Libet conducted quantitative research that raised important questions about the nature of free will in humans. In short, through his methodology, he demonstrated that your brain is already making decisions about what to do before ‘you’ are aware you are making a decision in the first place. The experiment raised doubts about ‘conscious free will’ which has been hotly debated since the release of Libet’s seminal paper. It raises an important question – are we actually freely acting on the world around us? Or are we products of our environment and predictably reactive to marketing campaigns, political campaigns, and the media?

If you think about it, the very fact that marketing exists stands as a testament that our autonomy is in question. If it didn’t we wouldn’t have marketing campaigns that resonant with most people – and those that don’t. Moreover, the expansion of neuromarketing techniques in creative testing, brand proposition testing and wider marketing strategies are beginning to show that we often react in a very similar way to each other. For that reason, some organisations have begun testing their brand strategy by using brain imaging techniques (e.g. fMRI, EEG) and eye tracking.

Furthermore, the emergence of machine learning has also began to raise some questions about the uniqueness of the individual. Machine learning is a discipline in computer science that gives computers the ability to progressively improve on a task using data, without explicit commands given to it by humans. Progression in this area has allowed machines to begin to understand human behaviour. Have you ever heard your friends or colleagues have a conversation that sounds like:

“I think my phone is listening to what I say, yesterday I was talking about cat food and then a day later I get advertisements for Felix…”

It can seem strange that Google could possibly know that you might need cat food. Especially when you never buy cat food online in the first place. However, your phone is not actually listening to you. Why it seems that way is partly a result of something called the Baader-Meinhof phenomenon (a cognitive bias in which recent names, words etc. that have recently come to one’s attention suddenly seem to appear everywhere) and, more importantly, the fact that Google can predict what sort of things you will be interested in, by your demographics and behaviour. They don’t need to listen to your conversations to know that you might want cat food, because they can create models that are able to predict your interests and needs, by learning from the behaviour of other people who are like-minded.

quantitative research

The massive developments in artificial intelligence (AI) in recent years are responsible for this. For example, Minority Report-style research in Germany created AI that was able to predict what would happen in the next few minutes of a scene in video with 40% accuracy. Everywhere, machines are learning to predict human behaviour and consequently building evidence that the feeling that you would have acted differently to your peers is perhaps not as infallible as you would like to think.

However, while we are not so different from the arbitrary groups that we may be assigned to, the complex factors that affect opinions and behaviour at the group level make implementing successful marketing strategies difficult. As discussed in our blog on testing marketing campaigns, in this new technological age, the opinion of the masses change at such a rate that it can be difficult for marketers to keep up. With high-tech AI still in its infancy, a more accessible way to create data-driven marketing campaigns is to pre-test them using research. By digging deep into data; into the different socio-economic groups and demographics that are available, you can get your marketing strategies right the first time. With access to the vast amount of data on your consumers, you can use quantitative research to predict performance of your marketing campaigns and learn how to increase your ROI.

This article was written by Tom Burke, Insight Executive at fastmap. To learn more about how our research can help inform your marketing strategies don’t hesitate to get in touch with David Cole, Managing Director, fastmap on +44 (0) 20 7242 0702 or david.cole@fastmap.com.

Hear more from the DMA

Please login to comment.