An array of techniques have cropped up in recent years, with the purpose of interpreting signals that our body gives to try and ‘measure’ people’s mindset, for example electroencephalography (EEG) sets, facial coding, or voice emotion measurement with the claim they can tell what participants like or dislike and whether they have a stronger affinity towards one product or the other. Some claim these techniques can measure emotions and can predict which way consumers will skew.

This is problematic. Not only because the measurements taken with those so-called “neuromarketing” techniques are not accurate, but also because they are based on a flawed assumption: that emotions are “things” we can measure.

We will explain why we can’t measure emotions, but first we will look at the use of neuromarketing techniques for market research and why they are not appropriate. Katy Irving, global head of the behavioural science unit at HRW (HRW Shift) tells us why these techniques are not right for market research applications.

KI: If we look at EEGs first, they are headsets with sensors placed on the skull that measure conductivity relating to different regions of the brain.  Medically, they help diagnose abnormal brain activity. Commercially, these headsets are promoted as a pragmatic neuroscience measure. However, members of HRW that have used medical EEG sets academically and compared the accuracy between those and commercial sets have raised concerns. It is important to note that medical EEGs use a multitude of electrodes spread across the scalp affixed with a gel placed at the interface between the skull and the set, which helps measure conductivity more accurately.

In contrast, ‘dry’ commercial EEGs are much less receptive to brain conductivity because they lack this gel, and they have fewer electrodes. So, we really are not confident that any commercial providers of EEGs have proven any type of validity to coordinate the brain regions that they have measured.  We have proven this in a self-funded study we ran where we used commercial EEG sets to test the ‘emotional impact’ of a campaign on infectious diseases, that we had in fact developed with a client previously. The EEG results predicted a campaign failure. However, we knew from the real-life data that it had a significant impact on screening: 12% increase after campaign launch. So, we don’t recommend the use of EEG in any market research applications because of our concerns regarding inaccuracy and lack of clarity in terms of what’s being measured.

Alexandra Petrache, PhD Neuroscience, behavioural science analyst at HRW: That’s interesting, Katy. During my PhD, I experimented with recording the electrical activity straight from brain cells in brain slices. To get reproducible results or even to be able to record activity at all, the conditions of the recording must be extremely controlled, from temperature, to mechanical stressors, to making sure the solutions are “right” and mimic the brain environment.  And this is not considering the experience necessary to perform the technique accurately, to understand and analyse the data. Keeping this in mind, the suggestion that an EEG sensor on the scalp that far away from the cells themselves would be capable of accurate measurements does not seem plausible. What do you think about facial coding or voice emotion measurement?

KI: Facial coding is a technique usually conducted with a webcam (but it can be done manually) which records facial movements of respondents presented with a stimulus and then translates those movements using an algorithm into 7 facial expressions. The outputs claim to be measuring emotion. However, several accuracy issues can be flagged: around the software, webcam and, most importantly, around whether facial expressions are an indication of emotion. On top of this, from our experience, the stimuli presented need to generate strong facial responses.


These techniques are not just inaccurate, but what they claim to be measuring does not exist the way we perceive it to.

Can they actually measure emotion? And even if they could measure a response that arises following a stimulus, how does one measure whether that response shows pleasure or displeasure? And how are emotions characterized?

First, one must ask themselves what an emotion is.

We have asked Professor David Thomson, founder and Chairman of HRW to share his thoughts.

DT: This boils down to how people choose to define emotion and what emotion is. But first, the term “neuromarketing techniques” has been largely misused as an expression. It implies that there are areas of the brain responsible for specific discreet emotions, areas that tell us when someone is sad or happy, which is simply not the case. In fact, highly competent people have carried out metanalyses which tell us that such correlations between brain regions and emotions don’t exist (Siegel et al, 2018).

Not only that, but emotions are not clear processes that manifest in the same way for all individuals. They don’t even manifest in the same way for the same individual and they are context dependent. That means, they are heavily influenced by the context in which we are- for the same person, anger could manifest very expressively one day or very quietly another day. Likewise, cultures shape emotion- different cultures manifest happiness or sadness or disgust in different ways, which are all on a spectrum and again, they manifest differently in different occasions even for the same individual.

What is generally agreed is that an emotion is a multi-component “thing”: we experience one thing and may express another. Alongside this there are neurophysiological factors at play. However, if you were to carry out any sort of measures you would not be capturing all of the emotion and how it could be influencing our behaviour. How can we do that when we ourselves cannot tell what we feel and how we feel all of the time?

Importantly, we should ask ourselves whether emotions are ‘natural’ and whether they stick to the discreet category of anger, disgust, fear, surprise…


Keeping in mind that emotions are not easy to measure and characterise, what should clients bear in mind when thinking about neuromarketing?

DT: First and foremost, we need to remember that human behaviour is driven by reward- we seek pleasure and avoid pain. Neuromarketing sounds great, because it’s about engineering our product to better appeal to people’s psychology and that is great, but when we take a step back, can we really do that?


By Katy Irving, Alexandra Petrache, and David Thomson


Apply Now!

Get in touch:

    I would like to receive news and information from HRW
    I agree to the HRW privacy policy

      Sign up to our newsletter

      If you wish to hear about our latest blogs, podcasts, webinars, and team news, simply enter your email to sign up to our monthly newsletter.


      We use cookies to give you the best possible experience. We can also use it to analyze the behavior of users in order to continuously improve the website for you. Data Policy

      analytical Essential