Understanding behavioural responses

To explore these types of data, one must consider the observational studies. These are techniques widely used in consumer behaviour research, by virtue of their focus on the context in which the phenomenon under analysis is immersed. The use of these methods is therefore appropriate whenever it is assumed that the “context” variable plays a fundamental role in determining and giving value to the consumption experience (Adler & Adler, 1994;Jorgensen, 1989).The observational research techniques are distinguished according to the level of participation of the researcher in the studied phenomenon: either pure observation or participatory observation. While pure observation requires researchers to remain extraneous observers to the phenomenon, analysing it at a distance without taking part in it, participatory observation, on the other hand, requires the researcher to come into contact with the phenomenon merely by observing it. Both techniques are useful for exploring the consumption experiences but for different goals. While pure observation can be adopted to gain behavioural data that could be easily treated as quantitative variables for descriptive and causal research, participatory observation provides rich and detailed information, aiming at exploring customer experiences. Indeed, it is a method of study of long tradition, justified by grounded theory, used by anthropologists to study human groups, cultures, and societies. Exploration of customer experience greatly benefits from participatory observation in any of its stages, especially because it overcomes the barriers of the communicability of knowledge (Troilo, 2006), generating deep knowledge about emotional interactions with contexts. Participatory observation can therefore be a powerful tool for collecting primary data (Grove & Fisk, 1992) aiming at drawing an overall picture of what happens in the investigated context, to then distinguish within it the behaviours or the specific phenomenon of his/her interest. Participatory observation is especially useful to fully understand consumer experiences overall. When consumers are not aware of what they feel, and they do not want to share their emotions with others, behavioural studies can be of help. However, behavioural responses have a limit: they are ambiguous (Holbrook, 1980). These investigations are not easy to interpret because the same behavioural or psychophysiological indicator can be considered a proxy for many different experiences. In addition, behavioural methods pose the need to understand the degree of association, the sequence of events investigated, the effects of possible intervening variables, and errors in the measurements of variables (Bagozzi, 1977; Baker, 2002).

Understanding sensorial responses

Through the analysis of the sensorial responses, brand managers can fully understand consumer emotions because the latter are the source of the former. Two kinds of studies can be useful to understand the sensorial responses of customer engagement: (1) studies on the physiological reactions of emotions at the peripheral level; and (2) studies on the physiological reactions of emotions at the central level. Both are powerful in detecting what the individual is not aware of or what he or she is unable to verbalize or codify explicitly and voluntarily (Zaltman, 2003). They are regarded as the new frontier of research into emotions and the consumption experience, but their initial adoption dates back to the 1970s. Recently, the neurological investigation techniques applied to marketing studies, also referred to as neuromarketing, have allowed us, at least in part, to explore new dimensions of customer experiences. Research techniques on physiological reactions attempt to measure the unconscious reactions of individuals to marketing stimuli, believing that this kind of data can be more interesting than the story that the consumer provides of his/her experience. In particular, this type of application is used to study emotions and cognitive processes that involve memory, areas in which traditional research applications seem to show the greatest limits.

Studies on physiological reactions at the peripheral level focus on the sensory reactions of individuals, both with respect to stimuli from the outside world and to those from within. Over time and with advances in technology, research on peripheral physiological reactions has evolved, developing new and more sophisticated measurement systems. Biofeedback provides accurate measures of spontaneous, and unconscious and immediate, reactions of the individual, which can be analysed via a mix of biology, psychology', and electronics, especially in stress conditions.

The sensory-emotional analysis based on the reactions of the body ranges from the study of posture, hand movements, and physical proximity to the study of pupil dilation, to the analysis of the voice, to the analysis of the heartbeat, blood pressure, pulse, breathing, and temperature. For instance, facial expressions, as originally widely studied by Paul Ekman and Wallace V. Friesen, who developed the facial action coding system (FACS) in the 1960s, are the result of one or more action units; i.e., the smallest and most visible single actions of the facial muscles, combined together as a physiological manifestation of primary emotions. For example, primary emotions, as expressed by the movement of facial muscles, are sadness, happiness, surprise, anger, fear, anxiety, and disgust. The movement of the muscles also indicates whether or not the emotion is authentic; as for the smile, authenticity and simulation use different muscles. Disney films illustrate how physiological reactions can be useful in designing new products (Box 5.2).


Understanding customer reactions to products is a priority for every customercentric company. The motion picture industry is no exception.

In 2017, Disney launched a facial-expression tracking system to identify audience reactions to movies (Deng et al., 2017). The system is based on a new algorithm, called "factorized variational autoencoders" (FVAEs), which use deep learning to automatically translate faces (and potentially any kind of image of a complex object) into a series of numerical data (latent representation or encoding). Disney Research, the California Institute of Technology (Caltech), and scholars at the Simon Fraser University in Canada collaborated in creating this new algorithm. The analysis works as follows:

  • 1. Gathering data. Infrared high-definition cameras capture motions and facial expressions in the dark that individuals make when watching movies. Specifically, data were gathered while watching 150 showings of nine Disney movies, including Big Hero 6, The Jungle Book, and Star Wars: The Force Awakens. These shows took place in the natural environment. Indeed, a 400-seat movie theatre was equipped with four infrared cameras. The latter captured 68 landmarks per face at a rate of two frames per second. A total of 3,1 79 audience members took part in the study.
  • 2. Converting images into numbers. These images of complex objects are then broken down into a set of numbers via the FVAEs. Each of the latter represents specific features that have been previously codified. For instance, one number captures the intensity of the smile, another one gets the diameter of the eyes, and so on.
  • 3. Using metadata. The data transformed via the algorithm - i.e., the sets of numbers - are then included into a network of metadata, so that links between several kinds of data can be created. For instance, the links between different physiological reactions expressed at several moments of time or expressed by different individuals can be analysed together. The final data set comprehends approximately 16 million bits of data, yielding some 16 million individual images of faces.
  • 4. Analysis and prediction. Once the neural network learns the patterns, and this step needs only a few minutes of observation, it can predict individual reactions, such as laughs and smiles, before their manifested expression.

Particular attention is paid to the relationship between emotion and pupil size. It is, in fact, the most uncontrollable emotional reaction of individuals and, for this reason, is considered more reliable for measurement purposes. This biofeedback is gathered through eye tracking, which is an increasingly common research instrument.

Further, the level of hand sweating is an emotional expression. To measure this phenomenon, a measurement called “galvanic skin response” (GSR) was developed, which evaluates the electrical conductivity of the surface of the skin of the hands. As excitement increases, the sebaceous cells of the skin of the hands increase the secretion, which, being salty, turns into a better electrical conductor than the skin in its normal situation. Consequently, the increase in the conductivity of the skin surface of the hands and the amount of electric current measured in ohms can be used as indicators of emotional excitement.

Studies on the physiological reactions of emotions at the central level comprehend several techniques, referring to two categories (Zaltman, 2003): (1) the techniques of inaction in the answer; (2) neuroimaging techniques. The first category is widely applied in psycholog)' studies, and only recently has it found space in marketing studies. These are useful techniques for analysing consumption experiences that are more imbued with the image of the social self, being focused on the reluctance of individuals to express their thoughts and their emotions. They study the span of time that passes from the question to the answer. The lesser or greater length of this period of time indicates the presence of an element of disturbance in the thoughts and emotions of the interviewees. Two main techniques belonging to this category are priming and the implicit association test.

With regard to the neuroimaging techniques, brand managers can create images from the activities of neurons in the minds of individuals. That is, it is possible to visualize the activity of the brain while individuals are involved in the activities that brand managers want to study. For example, Harvard neuroscientist Lawrence

Farwell (1995) studied the waves generated by the brain 300,000ths of a second after receiving a significant stimulus, to understand when an individual tells the truth. He then developed brain fingerprinting, a machine that uses the murmurs, memory, and encoding-related multifaceted electroencephalographic response methodology. Neuroimaging techniques seem to be particularly promising for studying the attitudes of individuals towards corporate communication and brand image. At least six important techniques belong to this category: electroencephalography (EEG); high-resolution electroencephalography (HEEG); positron emission tomography (PET); magnetoencephalography (MEG); functional magnetic resonance imaging (fMRI); and functional diffuse optical tomography (fDOT).

Regardless, the type of physiological response in use is difficult to interpret for any brand because the same indicators are proxies for many different experiences (Holbrook, 1980).

< Prev   CONTENTS   Source   Next >