Consumer behavior is predictable. What we see and miss, how we respond, what we think, what we forget and remember, and what we buy — it’s all far away from chaotic and unpredictable. In this age where a part of marketing is focusing on hyper-individualized strategies, perhaps something is being forgotten: that cultural microtrends and individual trajectories are not all there is.

So let’s start with a basic assumption: in addition to being cultural minds, we’re also biological creatures. Our brains evolved mainly for a life where the environment was very different. In some ways, the prehistoric world was more predictable. At least compared to today’s daily changes and disruptions. So when we go to the store, use our phones, or watch Netflix, we bring our evolutionary past. Our evolved brain is always part of who we are, and how we respond, feel, think, and act. This makes us predictable.

Unconscious brain activity predicts choice

A couple of decades ago, neuroscience research made a dramatic breakthrough. By inspecting brain responses up to simple choices such as pressing a button, neuroscientist Benjamin Libet found brain responses that preceded and thereby predicted the action a second before the choice. At the time, the discussion that followed focused more on whether humans have free will. But we completely missed a major point: the brain responses in Libet’s study showed that a person’s intentions could be told before she acted!

More recently, neuroscience studies by Stanford University’s Brian Knutson showed that individual consumer choice could be predicted several seconds before the actual choice. In the brain imaging study, participants were asked to first look at a product for 4 seconds, then were given a price for that product for 4 seconds, and finally asked to choose whether they would buy the product for that price (also 4 seconds). During the passive product viewing phase, brain responses in deep structures such as the nucleus accumbens were highly accurate in predicting what people would choose. This means that brain activity 8-12 seconds before the actual choice could predict what people would actually choose. When the researchers asked the participants about when they felt like making up their minds, they typically said something like “First I saw the product and then the price, then I weighed things a bit back and forth and then made up my mind.”

In my own work, a similar trend was found, but even more specific. Instead of just predicting what people would choose, we were able to predict the amount of money that participants would actually pay for the product. As soon as we are able to track brain activity related to valuation, we can predict choice and pricing.

A small group can predict a market

While individual choice prediction was being unraveled, a far more powerful prediction was emerging: the ability to predict market responses from a small sample of people. In several studies now, researchers have shown how brain activity in a sample of 30 participants or less can predict market effects.

As mind-blowing as that sounds, it all comes back to the initial thought: that several types of human responses are shared. This means that instead of focusing on individual responses, another path looks at shared and common responses. When groups converge in their responses, it is highly predictive of the responses of people outside that limited group.

For example, in a study from Lucas Parra’s lab, it was found that convergent brain activity in a small sample could predict cultural responses such as Twitter behavior and Nielsen TV ratings. Boksem and Smidts in Holland found that a certain type of brain activity in a small sample watching movie trailers was predictive of box office sales of the same movies. The aforementioned Brian Knutson reported that brain responses in a small sample could predict both financial market responses and crowdfunding responses.

This new research area goes under the heading of “neuroforecasting” — the ability of brain responses in a group of people to predict market responses. This relies heavily upon the previously mentioned common human biology: we all jump in our chairs to a jumpscare. Some jump more, some less, but it’s a very coherent behavior. As soon as our responses are coherent, even in a smaller group, it is highly predictive of how thousands and millions of people will respond.

From neuroforecasting to automatic predictions

Recently, there has been a final step on this interesting journey: the emergence of machine learning (ML) and artificial intelligence (AI). This has been driven by both a strong development in the fields of mathematics and statistics, as well as a substantial increase in the processing power of computers. But beyond self-driving cars and Netflix recommendations, a whole new breed of AI models is winning out.

This breed of AI models is modeled on the predictability of human responses. The more coherent human behavior is, the more likely it is that AI models will learn to predict such behaviors. Take visual attention: can you predict where people will look at a painting? An ad? A TV show? It can be a daunting task. But when given enough high-quality data from eye-tracking studies, an AI model can actually learn to predict human attention. At the last count, the best-in-class model had an accuracy of 95%.

That is, the model can predict where we as consumers will look at an ad, just by feeding it the image itself. It has already learned where we humans look and what we ignore.

‍

‍

Neurons AI is a visual attention prediction AI with 95% accuracy compared to hardware-enabled eye-tracking. Instead of running an eye-tracking study for weeks, this accurate prediction was created in 10 seconds.

Attention is only the beginning. New AI models are emerging, predicting emotional responses, memory, and ultimately choices. Just from the image or video itself.

What predictive AI mean for the customer experience

What does this mean for customers that they are predictable and that there are AIs emerging that accurately predict their (mass) behavior? After all, this new trend moves away from more problematic AI problems that are focused on the individual, to broader behavioral patterns.

To answer this question, one premise is needed: that the emerging AI solutions mean democratization and an increase in the availability of the tools that designers and agencies can use. Instead of costly and time-consuming projects to understand customers, predictive AI models can produce results in seconds and help users optimize their content in a day, instead of weeks or months.

With that in mind, two consequences can be eyed:

• Communication relevance — we will experience advertising, apps, websites, and other touchpoints as less intrusive and more relevant. With the power of available AI solutions, designers and agencies will reduce the waste in their production.
• Less noise — with 82% of online ads being ignored (Goo Technologies), up to 48% of products failing on the market (Merle & Cooper), and 97% of websites failing (Forrester), the consumer journey is littered with noise! AI-driven tools will allow a better accuracy in driving attention, engagement, and choice

While the AI revolution is upon us, the vast majority of attention tends to focus on the solutions that bring headlines. But the AI revolutions bring much more than autonomous vehicles and sophisticated chatbots. In addition, attention should be brought to the many smaller yet potent AI solutions being brought to the fore. These may turn out to be the true revolutionary changes that AI brings to us. Indeed, as Steve Jobs would say that there’s an app for everything, we will soon say “There’s an AI for that.”

‍