Are you considering investing in neuromarketing solutions, and would you like to know the latest about neuromarketing vendors and methodologies? The NMSBA took a snapshot of the industry in 2021, answering your questions about the field. What developments have taken place recently? And to what extent has the Covid-19 pandemic impacted these developments?
Electroencephalography (EEG) is one of the most widely used methodologies employed by neuromarketers today.
How it works
EEG measures brain activity by detecting and amplifying faint electrical signals, informally called brainwaves, that are emitted continuously by the brain. These electrical signals are the means by which our brain communicates and synchronizes activity across different anatomical regions. Variations in brainwave activity are indicators of changes in cognitive processing. Two metrics are commonly used to measure brainwave frequencies: power measures the amount of brainwave activity occurring at a particular frequency over a period of time; coherence measures the consistency or correlation of brainwave frequencies across different parts of the brain. Greater power means greater activity or energy in a given area at a given frequency; greater coherence between regions often means the regions are communicating as part of a cognitive process.
Pros and cons
It is the only brain measurement technique that can capture brain activity at the speed of cognition. It measures brain activity directly, rather than indirectly, through behaviors and choices. EEG equipment has become more affordable, portable, and wireless, opening up new possibilities for mobile, in-store, and virtual reality studies. Some limitations of EEG are that metrics can be challenging to understand and difficult to interpret. Designing, running, and interpreting the results of EEG studies requires PhD-level expertise. Repeated measures are required to separate signals from the background noise of unrelated brain activity (the “signal-to-noise” problem). This can make it difficult to measure responses to novel stimuli such as new products or packaging. Neither is EEG a very suitable technique for measuring electrical activity originating deep within the brain, such as in the emotional and memory centers, because those signals become too faint and dispersed before they reach the surface of the scalp.
Functional Magnetic Resonance Imaging (fMRI) is one of the more accurate methods of tracing patterns in the brain.
How it works
When a specific part of the brain is more active, it needs more oxygen. The blood flow to this area will therefore increase. fMRI detects changes in oxygenated blood flows in response to cognitive tasks, correlating to neural activity, but it doesn’t measure neural activity itself. Researchers use fMRI data to address fundamental questions about the nature of consumer decision-making, consumer experiences, and value-learning. These insights are used to complement explicit responses, to help companies understand both the conscious and the non-conscious factors influencing consumer behavior.
Pros and cons
Use of fMRI is non-invasive and it gives more detail on e.g., what a person is feeling, compared to, for instance, EEG. fMRI works well for more static stimuli, such as packaging design, campaign slogans, payoffs and outdoor messaging. Because of its low temporal resolution, however, it is less suitable for the measurement of dynamic stimuli, like video, TV shows, commercials, online user experience. In such cases, it is interesting to see the brain responding moment-to-moment.
Eye-tracking measures eye movements and gaze.
How it works
Eye-tracking has a natural appeal because it is intuitive: We all know our eyes automatically follow what interests us, threatens us, or attracts us. Variations in eye movements, including speed of movement, duration of fixations, patterns and frequency of blinks, and patterns of visual searching behavior, are all relevant to how a person is responding to a visual stimulus like an ad, a video, a website, or a store shelf. Most dedicated eye-tracking systems use infrared (IR) light to find and trace eye movements and other important measures such as pupil size and distance for the stimulus source, both for in-lab systems and mobile eyewear. Webcam-based systems are different in that they can only access the visible light captured by the webcam. Webcam-based eye-tracking is still less precise than dedicated IR systems, both temporally and spatially, but new developments in gaze recording technology are shrinking the performance gap between IR-based and IR-free eye-tracking.
Pros and cons
Eye-tracking is relatively inexpensive, scalable, and delivers results in short timeframes. There are some limitations, however. By itself, eye-tracking cannot tell you why someone is looking at something. It can tell you what they see, but not necessarily what they perceive. It cannot tell you whether visual attention is accompanied by positive or negative emotional valence. Another important limitation is that objects in the periphery (not recorded by this method) can still have a significant effect on reactions and subsequent behavior. These issues can be addressed by combining eye-tracking with other methods. One popular combination is eye-tracking with facial expression coding, which can be used to measure emotional valence in conjunction with fixations, gaze paths, and pupil response during exposure to marketing stimuli.
Response time studies leverage the fact that human brains take time to think.
How it works
We take less time to think when we encounter things that are familiar or expected and we take more time when we encounter things that are novel or unexpected. Response time studies take advantage of this difference in processing speeds to provide a window into how concepts and attitudes are connected in long-term memory. “Strength of association” refers to the degree to which things “go together” in long-term memory. From that simple foundation, response time studies have expanded into a powerful set of tools to explore many mental phenomena, including how brands and products are connected to each other and related ideas in consumers’ minds.
Many different response time tests and implicit association measures have been developed over the years. Three of the most popular techniques used in neuromarketing today are semantic priming, affective priming, and the Implicit Association Test (IAT).
Pros and cons
The ability of response time studies to expose implicit attitudes is its biggest advantage as a technique in the neuromarketing toolkit. Numerous studies have shown not only that implicit attitudes can differ significantly from explicit, self-reported attitudes, but they can also sometimes produce better predictions of actual choices and behavior, especially when people are acting impulsively or under pressure.
Response time tests, when properly designed and executed, can be run online at scale; they are inexpensive to conduct and straightforward to interpret; they require no sensors, specialized labs, or complex data analysis algorithms; and they can be turned around quickly. It is important to keep in mind, however, that the strength of an association is not a measure of the type of association it represents. A strong association may represent an identity (A is B), a causal relation (A causes B), a simple co-occurrence (A accompanies B), or some other relationship. Just because someone strongly associates “dog” with “cat” does not mean they think dogs are cats. Implicit associations are unconscious connections that may or may not make logical sense. They are relatively easy to identify but need to be interpreted with caution.
Biometrics include physiological responses like perspiration, respiration, and heart rate.
How it works
The term “biometrics” refers to a wide range of physiological changes that occur when human beings respond emotionally and physically to the world around them. Many biometric responses are involuntary, so they provide a window into nonconscious processes that accompany consumer choice and behavior without conscious awareness.
Neuroimaging technologies like EEG and fMRI capture CNS (Central Nervous System) activity inside the brain. Biometrics capture physical responses (muscle movements) directed by the brain through the PNS (Peripheral Nervous System). These physical actions are communicated through two subsystems within the PNS: the autonomic nervous system (ANS) and the somatic nervous system (SNS). Signals originating through the ANS impact “smooth” muscles, organs, and glands like the heart and stomach. They are relatively slow, mostly automatic, and produce body responses such as perspiration, heart
rate, breathing, and pupil dilation. Signals originating through the SNS impact the skeletal muscle system, are much faster and, importantly, are also under at least partial voluntary control. They include responses like facial expressions, eye movements, blinks, and physical actions like walking and talking. (Genco et al., 2013).
Pros and cons
Biometric measures are increasingly popular in neuromarketing due to their low cost, scalability, fast turnaround times, and intuitive metrics. Because different physiological reactions reflect different aspects of emotion (discrete, dimensional, arousal vs. valence) and cognition (attention, cognitive load, memory activation, fatigue), these measures are best used in combination with each other and with neuroimaging technologies like EEG and fMRI.
Consumer Neuroscience and AI/Machine Learning
Artificial intelligence and machine learning systems allow researchers to build better models to simulate and predict how people respond to marketing, products, brands, and shopping experiences.
How it works
In most products where machine learning is used, we have some input for which we want to generate a certain output. For instance, the input is a picture and the output is whether it features a face or not. We call this machine learning when the algorithm transforming the input to the output is not a set of predefined rules designed by a programmer but is automatically derived from a large set of training data. The algorithm that maps the output onto the input is learned automatically, by the “machine” itself, hence machine learning.
In machine learning, the algorithm that sits between input and output is an artificial neural network. This network is a collection of connected units or nodes, which loosely model the neurons in a biological brain. Each node can send and receive signals from and to other nodes it is connected to. The output that each unit sends is some non-linear function of the sum of its inputs. The weight of the connections between units determines the size of the effect that one unit has on the other. Typically, neurons are aggregated into layers, similarly to a biological brain.
Pros and cons
Like our brains, machine learning algorithms get smarter the more data they absorb. The main strength of these systems is their ability to discover patterns in complex data. They have two main weaknesses: their predictions assume the future will be much like the past, and their results do not provide explanations for why they work (or don't work). In addition, many consumers have reservations about the ethical implications of AI systems. To adopt AI, legitimate concerns about bias, consumer manipulation, and privacy need to be addressed.
Neuromarketing Service Offering in 2021
Most of the respondents in our 2021 vendor survey have seen a shift towards the use of more online services, due to the Covid-19 pandemic. Some companies have stopped using lab research and moved completely to online neuromarketing research. Online facial coding and eye-tracking research in particular have really taken flight, as well as survey research and implicit testing.
Another development is that more mainstream or standard methodologies have been refined so data can be analyzed quicker, and results can now be more easily interpreted by marketers. “We have refined our facial coding offer and digitalized more products, making them more agile and available online”, said
one respondent. Several companies have used this period of time to introduce new offerings: “We launched our predictive eye-tracking platform in June 2020, and we have shifted our focus towards a platform that is more scalable and brings neurotechnology to a broader market,” commented one vendor.
Another company has focused on innovative methodologies with the addition of smart-speaker and augmented reality surveys, as well as the use of EEG-UX, to better capture consumer experiences hands-free, in-the-moment, and in-context.
One respondent said it had introduced an online version of their platform to allow their clients to perform state-of-the-art biometric research remotely. Others have integrated existing technologies better: “The integration of EEG, eye-tracking and VR allows for a completely new approach to retail and shelf research,
that's both highly realistic and offers a high degree of experimental control and stimulus randomization”, one vendor explained.
Looking at the future
When the respondents were asked whether they expected any shifts in terms of the methodology used in the near future, most envisioned the growth of online approaches e.g., implicit and facial coding to continue, although not necessarily at the expense of other methods (such as EEG). Another development to watch is the integration of AI and machine learning as a means of building predictive success models for clients and more sophisticated metrics. Due to the changes of the past year as a result of Covid-19, such as remote working and virtual research, the need for in-the-moment and in-context research will continue to increase. Clients have found a more robust and engaged sample in addition to the convenience of online research. Research methods will continue to adapt to these new environments.
The unforeseen circumstances of the pandemic have opened a space for adaptability and innovation. With more accessibility to thought leaders and educational opportunities, clients are better informed about the types of tools and available methodologies. The neuromarketing field has matured with clients asking better questions and expecting more actionable results. Methodologies all have limitations; however, finding ways for tools to complement each other will generate useful frameworks for studying human behavior.