Applied Neuromarketing: Using Psychophysiological Testing to
Detect and Measure Positive Emotional Responses in Quantitative
Marketing Research
By Nathan Tothrow, Chief Marketing Officer
Piedmont Advantage Credit Union
Abstract
The emerging field of neuromarketing offers the potential to give us a better understanding of the
connection between marketing and brand associations in the consumer’s mind. It has already
been demonstrated that emotions, especially positive emotions, can play a critical role in the
process of creating effective associations. And while neuropsychological research can identify
which areas of the brain are active while processing marketing messages, it is
psychophysiological research which holds the most promise for identifying the advertising
messages with the greatest potential for generating those positive emotional reactions. Tests
conducted over the summer of 2010 profiled the physiologic reactions of subjects who watched a
demo reel of television commercials and comedy shows. The results showed correlations that
suggest positive emotional reaction profiles could be isolated by age, sex and other demographic
groups.
Over the summer of 2010 a series of psychophysiological experiments was conducted to test the
possibility of detecting and measuring the emotional reactions of subjects who watched a series
of television commercials. Using physiological data from a digital polygraph and a video reel
containing segments of comedy shows and commercials, the subjects’ emotional responses were
recorded and averaged mathematically. The results showed striking distinctions between the
reaction profiles of males and females among the same age demographic, and suggested the
potential for success with additional trials.
To understand the relevance of these tests and their results, it is helpful to introduce the
background of the study and an overview of the testing structure.
The field of neuromarketing was once just a novelty to marketers. But recent research on the
neural connections between consumers and the choices they make has begun to yield actionable
results. In a 2004 neuromarketing study at the Baylor College of Medicine, researchers used
functional MRI scans to show how increased activity in the hippocampus could successfully
predict product choices made by participants. The study also suggested that marketing and
advertising in the form of cultural influences could affect those choices by creating positive
associations with a specific brand (McClure et al, 2004).
If marketers know that branding and other cultural elements can influence the buying choices of
consumers, they have a vested interest in developing advertising and other product messages that
will create the positive emotional experiences that lead to those associations.
Traditionally, marketers and advertisers have relied on focus group research to test their
assumptions and to gather qualitative feedback on consumer responses to their products and their
advertising. But this type of research is not without flaws that can introduce confounding biases
into the results. In an attempt to draw out thoughtful responses and insights, focus groups are
typically loosely structured and the free flow of ideas and suggestions is encouraged. However,
not everyone responds well to this format. Participants in focus groups have reported feeling
group pressure to conform to prevailing ideas, or that they were deterred from participating in
discussions by stronger personalities in the group. Such activity can bias the results by excluding
potentially pertinent input from being included in the final analysis.
A potentially effective—or at the very least, powerfully supplemental—alternative to focus
group research involves the use of psychophysiological testing techniques that measure the
positive or negative (i.e., valenced) emotional response a subject would have to samples of
marketing and advertising materials. Based on the same physiological signatures used in
polygraph examinations, such tests can be used to detect the occurrence of valenced emotional
reactions.
Considerable research has been done in the past few decades exploring the relationship between
emotions and human physiology. Paul Ekman at the University of California San Francisco and
Dacher Keltner at the University of California Berkeley have demonstrated how autonomic
innervation of the facial muscles can produce expressions unique to specific emotions (Lewis
and Haviland-Jones, 2000). Emotion specificity is a theory based on the work of psychologist
William James that suggests distinct emotions can be linked to a specific set of physiological
reactions. While researchers are still working to identify and isolate precise measures of emotion
specificity, there is enough data available to show that emotions can be broadly classified as
positive or negative (i.e., valenced) by their physiologic signatures.
Physiologic data acquisition for the experiment was performed with a five-channel Stoelting CPS
II polygraph that can measure heart rate (HR), blood pressure (BP), skin conductivity (SC),
finger pulse volume (PL) and respiration (RL). Shawn Edwards, a division manager at Stoelting,
had provided basic training on the hardware and software with the CPS II system and supplied
ongoing guidance on minor software modifications to increase the effectiveness of the data
analysis. Although the CPS II was originally designed for administering polygraph examinations,
it borrows from years of technological research in the field of physiological data acquisition.
Stoelting has a dedicated neuroscience research division and Edwards was enthusiastic about
offering assistance to this project.
The software for the CPS II was originally developed to detect emotional signatures of deception
in polygraph examinations. Designed and written by Dr. John Kircher of the University of Utah,
it uses raw scores of emotional response to calculate a mathematical mean across a selected
group of indices. A Z-score (measure of standard deviation) for each tested segment can then be
deduced. It is these Z-scores that were used to calculate degrees of significance in increases or
decreases in key physiological responses.
The laboratory environment was constructed following the recommendations of practicing
psychophysiological researchers (Curtin, Lozano and Allen, 2007). Test participants were shown
a prepared demo reel that integrated a collection of short comedy show clips (control segments)
integrated with a series of 30-second commercials (relevant test segments). By simulating an
actual viewing environment, the emotional reactions elicited in the experiment were more likely
to be representative of those experienced by consumers.
For each subject, scores in each physiologic measure were recorded. The mean was calculated
and the Z-score for each commercial was determined. Using the standard levels of significance
(+/- 0.4 sigma) used in polygraph examinations, each physiologic channel could be scored “up”
or “down” for each tested segment. The flexibility of the software also allowed for the arithmetic
mean to be calculated from the averaged scores of the comedy (control) segments as well. This
would allow for testing relevant reactions against the average reaction to other tested
commercials as opposed to testing the reaction against the overall viewing environment.
Each test generated a chart that graphically showed the activity of each physiologic channel
against a timeline coinciding with the test reel. This allowed a real-time review of how the
subject reacted relative to the individual timeline of each segment or commercial. But the
numerical scoring of each commercial’s Z-scores allowed for a closer, more quantitative analysis
of the subject’s reaction to the test material.
While far from conclusive, the results gathered from these initial trials indicated that males and
females were indeed showing different reactions to the same test material. Using test data from a
number of physiological researchers who tracked increases and decreases in key measures
against the experience of reported specific emotions, it was possible to identify which
commercials generated the most positive response from each subject.
In addition to the Z-score numerical analysis that was the basis of the experiment, the data were
also subjected to a probability analysis normally reserved only for polygraph examinations.
The CPS II software included a separate analysis based on Bayes’ theorem that calculated a
percentage probability of deception for a relevant question. This analysis was introduced at
random on a commercial that was created to target male consumers of a specific demographic.
The results showed a strong disparity between male and female viewers with the male scoring
0.0013 for deception and the female scoring 0.99 for deception. These scores can be interpreted
to indicate that the male viewing this commercial exhibited the same emotional reaction profile
that could be expected from a person who is answering a question with complete honesty, while
the female was experiencing the emotions expected of a person being highly deceptive. The
producers of this spot had subjective ideas that their commercial would be successful with males,
but this probability analysis generated quantitative backing for that hypothesis.
The implications from these initial tests suggest that it is possible to build a testing paradigm that
can accurately determine the probability that a sample of marketing materials can elicit a positive
emotional response from the targeted consumer. Additional trials are scheduled for fall of 2010
and will include tests designed to detect vagal nerve activity which some researchers have linked
to sensations of deepened trust and confidence (Keltner, 2007).
Curtin, John; Lozano, David and Allen, John. “The Psychophysiological Laboratory.” Handbook of Emotion
Elicitation and Assessment. Ed. Coan, James and Allen, John. Oxford University Press. New York, NY.
2007.
Keltner, Dacher. Course Lecture, “Bodily Changes and Emotion 1.” PSYCH 156 Human Emotion. University of
California. Berkeley, CA. 1 Oct. 2007.
Keltner, Dacher and Ekman, Paul. “Facial Expression of Emotion.” Handbook of Emotions, Second Edition. Ed.
Lewis, Michael and Haviland-Jones, Jeannette. The Guilford Press. New York, NY. 2000.
McClure, Samuel; Li, Jian; Tomlin, Damon; Cypert, Kim; Montague, Latane and Montague, Read. “Neural
Correlates of Behavior Preference for Culturally Familiar Drinks.” Neuron. Vol. 44 (14 October). pp. 379-
387. Cell Press. 2004.
Detect and Measure Positive Emotional Responses in Quantitative
Marketing Research
By Nathan Tothrow, Chief Marketing Officer
Piedmont Advantage Credit Union
Abstract
The emerging field of neuromarketing offers the potential to give us a better understanding of the
connection between marketing and brand associations in the consumer’s mind. It has already
been demonstrated that emotions, especially positive emotions, can play a critical role in the
process of creating effective associations. And while neuropsychological research can identify
which areas of the brain are active while processing marketing messages, it is
psychophysiological research which holds the most promise for identifying the advertising
messages with the greatest potential for generating those positive emotional reactions. Tests
conducted over the summer of 2010 profiled the physiologic reactions of subjects who watched a
demo reel of television commercials and comedy shows. The results showed correlations that
suggest positive emotional reaction profiles could be isolated by age, sex and other demographic
groups.
Over the summer of 2010 a series of psychophysiological experiments was conducted to test the
possibility of detecting and measuring the emotional reactions of subjects who watched a series
of television commercials. Using physiological data from a digital polygraph and a video reel
containing segments of comedy shows and commercials, the subjects’ emotional responses were
recorded and averaged mathematically. The results showed striking distinctions between the
reaction profiles of males and females among the same age demographic, and suggested the
potential for success with additional trials.
To understand the relevance of these tests and their results, it is helpful to introduce the
background of the study and an overview of the testing structure.
The field of neuromarketing was once just a novelty to marketers. But recent research on the
neural connections between consumers and the choices they make has begun to yield actionable
results. In a 2004 neuromarketing study at the Baylor College of Medicine, researchers used
functional MRI scans to show how increased activity in the hippocampus could successfully
predict product choices made by participants. The study also suggested that marketing and
advertising in the form of cultural influences could affect those choices by creating positive
associations with a specific brand (McClure et al, 2004).
If marketers know that branding and other cultural elements can influence the buying choices of
consumers, they have a vested interest in developing advertising and other product messages that
will create the positive emotional experiences that lead to those associations.
Traditionally, marketers and advertisers have relied on focus group research to test their
assumptions and to gather qualitative feedback on consumer responses to their products and their
advertising. But this type of research is not without flaws that can introduce confounding biases
into the results. In an attempt to draw out thoughtful responses and insights, focus groups are
typically loosely structured and the free flow of ideas and suggestions is encouraged. However,
not everyone responds well to this format. Participants in focus groups have reported feeling
group pressure to conform to prevailing ideas, or that they were deterred from participating in
discussions by stronger personalities in the group. Such activity can bias the results by excluding
potentially pertinent input from being included in the final analysis.
A potentially effective—or at the very least, powerfully supplemental—alternative to focus
group research involves the use of psychophysiological testing techniques that measure the
positive or negative (i.e., valenced) emotional response a subject would have to samples of
marketing and advertising materials. Based on the same physiological signatures used in
polygraph examinations, such tests can be used to detect the occurrence of valenced emotional
reactions.
Considerable research has been done in the past few decades exploring the relationship between
emotions and human physiology. Paul Ekman at the University of California San Francisco and
Dacher Keltner at the University of California Berkeley have demonstrated how autonomic
innervation of the facial muscles can produce expressions unique to specific emotions (Lewis
and Haviland-Jones, 2000). Emotion specificity is a theory based on the work of psychologist
William James that suggests distinct emotions can be linked to a specific set of physiological
reactions. While researchers are still working to identify and isolate precise measures of emotion
specificity, there is enough data available to show that emotions can be broadly classified as
positive or negative (i.e., valenced) by their physiologic signatures.
Physiologic data acquisition for the experiment was performed with a five-channel Stoelting CPS
II polygraph that can measure heart rate (HR), blood pressure (BP), skin conductivity (SC),
finger pulse volume (PL) and respiration (RL). Shawn Edwards, a division manager at Stoelting,
had provided basic training on the hardware and software with the CPS II system and supplied
ongoing guidance on minor software modifications to increase the effectiveness of the data
analysis. Although the CPS II was originally designed for administering polygraph examinations,
it borrows from years of technological research in the field of physiological data acquisition.
Stoelting has a dedicated neuroscience research division and Edwards was enthusiastic about
offering assistance to this project.
The software for the CPS II was originally developed to detect emotional signatures of deception
in polygraph examinations. Designed and written by Dr. John Kircher of the University of Utah,
it uses raw scores of emotional response to calculate a mathematical mean across a selected
group of indices. A Z-score (measure of standard deviation) for each tested segment can then be
deduced. It is these Z-scores that were used to calculate degrees of significance in increases or
decreases in key physiological responses.
The laboratory environment was constructed following the recommendations of practicing
psychophysiological researchers (Curtin, Lozano and Allen, 2007). Test participants were shown
a prepared demo reel that integrated a collection of short comedy show clips (control segments)
integrated with a series of 30-second commercials (relevant test segments). By simulating an
actual viewing environment, the emotional reactions elicited in the experiment were more likely
to be representative of those experienced by consumers.
For each subject, scores in each physiologic measure were recorded. The mean was calculated
and the Z-score for each commercial was determined. Using the standard levels of significance
(+/- 0.4 sigma) used in polygraph examinations, each physiologic channel could be scored “up”
or “down” for each tested segment. The flexibility of the software also allowed for the arithmetic
mean to be calculated from the averaged scores of the comedy (control) segments as well. This
would allow for testing relevant reactions against the average reaction to other tested
commercials as opposed to testing the reaction against the overall viewing environment.
Each test generated a chart that graphically showed the activity of each physiologic channel
against a timeline coinciding with the test reel. This allowed a real-time review of how the
subject reacted relative to the individual timeline of each segment or commercial. But the
numerical scoring of each commercial’s Z-scores allowed for a closer, more quantitative analysis
of the subject’s reaction to the test material.
While far from conclusive, the results gathered from these initial trials indicated that males and
females were indeed showing different reactions to the same test material. Using test data from a
number of physiological researchers who tracked increases and decreases in key measures
against the experience of reported specific emotions, it was possible to identify which
commercials generated the most positive response from each subject.
In addition to the Z-score numerical analysis that was the basis of the experiment, the data were
also subjected to a probability analysis normally reserved only for polygraph examinations.
The CPS II software included a separate analysis based on Bayes’ theorem that calculated a
percentage probability of deception for a relevant question. This analysis was introduced at
random on a commercial that was created to target male consumers of a specific demographic.
The results showed a strong disparity between male and female viewers with the male scoring
0.0013 for deception and the female scoring 0.99 for deception. These scores can be interpreted
to indicate that the male viewing this commercial exhibited the same emotional reaction profile
that could be expected from a person who is answering a question with complete honesty, while
the female was experiencing the emotions expected of a person being highly deceptive. The
producers of this spot had subjective ideas that their commercial would be successful with males,
but this probability analysis generated quantitative backing for that hypothesis.
The implications from these initial tests suggest that it is possible to build a testing paradigm that
can accurately determine the probability that a sample of marketing materials can elicit a positive
emotional response from the targeted consumer. Additional trials are scheduled for fall of 2010
and will include tests designed to detect vagal nerve activity which some researchers have linked
to sensations of deepened trust and confidence (Keltner, 2007).
Curtin, John; Lozano, David and Allen, John. “The Psychophysiological Laboratory.” Handbook of Emotion
Elicitation and Assessment. Ed. Coan, James and Allen, John. Oxford University Press. New York, NY.
2007.
Keltner, Dacher. Course Lecture, “Bodily Changes and Emotion 1.” PSYCH 156 Human Emotion. University of
California. Berkeley, CA. 1 Oct. 2007.
Keltner, Dacher and Ekman, Paul. “Facial Expression of Emotion.” Handbook of Emotions, Second Edition. Ed.
Lewis, Michael and Haviland-Jones, Jeannette. The Guilford Press. New York, NY. 2000.
McClure, Samuel; Li, Jian; Tomlin, Damon; Cypert, Kim; Montague, Latane and Montague, Read. “Neural
Correlates of Behavior Preference for Culturally Familiar Drinks.” Neuron. Vol. 44 (14 October). pp. 379-
387. Cell Press. 2004.
No comments:
Post a Comment