Matching advertising images to personality traits can find the most suitable image for a particular customer and help win sales, finds a new study co-authored by Dr David Stillwell of Cambridge Judge Business School.
Consumers pay far more attention to
images than words, so the people who design television commercials or print ads
have long agonised over picking the perfect image.
But the shift from one-to-many mass
communication to personalised one-to-one digital communication could soon pose
the question: “What is the most appealing image for this particular consumer?” That’s the focus of a new study,
co-authored by Dr David Stillwell of Cambridge Judge Business School, just
published in the Journal of Consumer
The study finds that by matching
the features of images with personality traits, it is possible to find not only
a suitable image for the general
population but the most suitable image
for a particular customer – and, importantly, this influences their purchasing
intentions towards the brand behind such an image.
This “additional ‘personal
touch’ could turn out to be crucial in building strong and successful long-term
relationships with consumers,” says the study. “Making a company’s
website or marketing campaign even a little bit more appealing than that of the
competition might be the deciding factor in whether a consumer buys from one
company or the other” – and this “could lead to meaningful gains”
if matching images to personality is implemented at scale by multinationals and
other big marketers.
“Marketers have always sought
to personalise communication, but this usually focuses on what is communicated in terms of the product advertised or the
element featured on a company’s website landing page,” says study
co-author David Stillwell, University Lecturer in Big Data Analytics &
Quantitative Social Science at Cambridge Judge Business School. “Our
findings point out the big potential of customising how this content is communicated by matching images to personality.”
It’s long been known that some
image characteristics have the widest appeal to average consumers (they favour
cold colours and symmetrical compositions, for example), and demography narrows
that down further (preference for saturated colours is stronger in men than
women). Yet the study notes that demographic variables “very quickly reach
their limits of discriminating between consumers in meaningful ways: it seems
rather unrealistic to expect all women of a certain age to have the same
The study therefore goes beyond
demographics to focus on psychological differences – as measured by the
well-known Five Factor Model personality test that measures Openness,
Conscientiousness, Extroversion, Agreeableness, and Neuroticism.
The first part of the study
recruited 745 participants, and selected 1,040 professional images from image
library Shutterstock.com taken from dozens of categories that include “nature”,
“buildings” and “people”. Computer algorithms then
extracted 89 features per each image, which are split into four main
categories: colour (including hue, saturation and value), composition (such as density
of edges, aspect ratio and blurring of background), texture (granularity,
contrast and homogeneity) and content (focusing on nine popular objects such as
cats, people, cars and bicycles).
In total, these participants gave
38,740 ratings on how much they liked the images, and then completed a 50-item
Five Factor Model personality test.
In terms of personality
characteristics: openness correlated positively (among other features) with the
colours blue and black; conscientiousness with the colour red and the presence
of at least one person; extroversion with the colour pink, low saturation, and
sharpness in the inner part of the image; agreeableness with the use of light
and the colours brown, green pink, purple, red and yellow; neuroticism with
image size and aspect ratio, and the display of cats (while showing negative
correlation with the colour brown and the presence of faces).
“This pattern suggests that
people high in neuroticism prefer natural images and images with no people,”
the study says. “This preference for calm and minimally stimulating scenes
without people is in line with the general attributes of neuroticism, including
envy, loneliness, anxiety, and fear.”
Later parts of the study recruited
an additional 868 participants to evaluate images in the Shutterstock.com
categories “holiday”, “beauty” and “phone” – and
also evaluated the participants’ attitudes toward particular brands (“appealing/unappealing”)
and their intent to purchase from those brands (“never/definitely”).
“The results demonstrate the
value of using image-person fit by showing that people do not only like
matching images more but also report more favourable attitudes and purchase
intentions towards brands that use matching images,” the study says. “This
spillover effect of image perceptions to brand perception is crucial for
marketers who want to use the mechanism of image-person fit to increase not
only the appeal but also the effectiveness of their marketing campaigns.”
The study – entitled “Predicting the personal appeal of marketing images using computational methods” – is co-authored by Sandra C. Matz of Columbia Business School; Cristina Segalin of the Department of Electrical Engineering at California Institute of Technology; David Stillwell of Cambridge Judge Business School; Sandrine R. Müller of the Psychology Department at the University of Cambridge; and Maarten W. Bos of Disney Research in Pittsburgh, Pennsylvania.