Copyright © ESOMAR 2008
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
Getting anima ted abou t emo tion
The new frontier
David Penn
LOOKING FOR EMOTIONAL ENGAGEMENT
The story so far
We’ve been getting quite animated about emotional
engagement for some years now, but we still cannot
agree on how to measure it. What almost everyone
does agree on, however, is that conventional quantitative
research methods – with their emphasis on conscious,
cognitive and rational response – fall short when it
comes to measuring emotion.
Gerald Zaltman (2003) suggests that marketing
researchers put so much emphasis on the cognitive and
rational because it is easier to think about the consumer
that way and to frame questions that way. Perhaps, but
it may also be because the inception of survey research
(as we know it today) coincided with the rise of mid–20th
century cognitive science and predated the neuroscientific
revolution at the end of the century.
Cognitive Science was mainly concerned with the mind’s
conscious mental processes – rather than with the
unconscious or the emotional – and was the dominant
paradigm until, in the late 20th century, the work of
neuroscientists like Antonio Damasio and Joseph Le
Doux challenged its assumptions about how the mind
works. Neuroscience starts from the premise that we
cannot understand the mind without first understanding
underlying brain structures and processes. In other
words, thinking and feeling are not separate from the
brain, but neurally embodied in it.
Like it or not, our feelings, our reasoning, our conscious
and unconscious thoughts are framed by our emotions
and bodily responses. It is therefore not enough to ask
people how they consciously feel about something,
because emotion occurs below the level of conscious
cognitive response. We can say how we feel, but we
find it very difficult to say why we came to feel that
way because the processes by which our thoughts and
feelings are created are mainly invisible to us.
The limitations of verbal response
Most conventional methods of measuring “how
consumers feel” are still based on expressed/verbal
response. Typically researchers ask people how
they feel and then ask them to explain why they feel
that way. Moreover, the measures which conventional
research uses to measure emotion are languagebased
and literal – the problem being that respondents
need to consider their response before answering. Yet
experience tells us that the more we think and consider
our response the further we get from our emotions,
because we start to post rationalise. Hence verbal
questioning (both quantitative and qualitative) struggles
to measure emotion, because it seems to mainly engage
our cognitive brain – the one that (consciously) analyses,
reflects, calculates and makes decisions – rather
than our emotional brain, that reacts spontaneously,
immediately and intuitively.
The limitations of verbal questioning in measuring
emotions has been further brought into focus by recent
important work into how advertising works. Binet and
Field (2007) assessed 880 (UK) case studies from the
IPA’s rigorous effectiveness awards scheme, concluding
that “communications models that use emotional appeal
are more likely to yield strong business results than
rationally based models (information and persuasion)”.
They also went on to draw the startling conclusion that
current conventional advertising pre-testing may even
reduce effectiveness; advertising that had favourable
pre-testing did significantly worse (in terms of business
Copyright © ESOMAR 2008
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
performance) than that which did not. The authors
conclude “the data are clearly not suggesting that pretesting
is entirely worthless, but do cast considerable
doubt on the ability of such research to reliably pick
winners”.
Unsurprisingly, some have lost faith with conventional
research altogether and opted for techniques (such
as brain imaging) that identify which parts of the brain
are active when we feel emotion. There was much
excitement, for example, a few years ago when a
completely new kind of market research study hit the
headlines – at the Baylor Institute in Houston, neuroscientists
used fMRI to monitor brain response during a
test of taste and brand preferences for Pepsi vs. Coke.
The neuroscientists (McClure et al, 2004) discovered that
in a blind product test (where preference for Pepsi vs.
Coke was balanced) only those parts of the brain relating
to sensory judgement were active. When respondents
were told what they were drinking, however, preference
switched in Coke’s favour, and a part of the brain (the
Hippocampus) associated with emotional response, also
became active. Brain imaging showed that knowledge
of Coke branding activated areas of the brain associated
with emotional judgement, whilst knowledge of Pepsi had
no corresponding effect. The study’s authors concluded:
‘‘Our finding suggests that the hippocampus may
participate in recalling cultural information that biases
preference judgements.’’
Excited journalists were soon hailing the birth of “neuromarketing”
and writing articles with titles such as “In
Search of the Buy Button” or even “Pushing the Buy
Button”. A new era seemed to be dawning, in which
conventional research would give way to techniques
based on brain imaging and physiological response.
So, what has happened, four years on from the birth
of neuromarketing? There have been some advances
and some worthwhile studies – particularly in the area
of media engagement – and we should also note the
re-emergence of bio-metrics alongside neuromarketing.
Based on measures of skin conductance, heart rate
and respiration, bio-measurement actually dates from
the early 20th century, but modern neuroscientific understanding
combined with advances in technology have
made its application to modern market research possible.
But can these techniques tell us something that conventional
research cannot? Actually, when it comes to
emotion, neurological/biological measurement can
generally tell us what, but not why. Mast and Zaltman
(2006) nail the problem perfectly with their observation
that ‘‘emotion without cognitive appraisal is really just
arousal’’. In other words, simply observing a neurological
or biological response is not the same as understanding
an emotion – because one is about physiology, the other
about cognitive appraisal. Indeed, most of the proponents
of neurological/biological approaches admit that their
techniques are not alternatives, but complements to
conventional research, either quantitative or qualitative.
Why? Because they have no way of understanding what
a respondent is feeling at the time of observation other
than by asking a conventional question. All of this is
somewhat ironic (and circular) when we consider that
neuromarketing and bio-metrics developed because
conventional research was not up to the task of
measuring emotion.
How do we go forward?
Neuromarketing is, essentially, about using brain imaging
techniques to locate activity in the brain and then
correlating that observation with mental response – to
a brand, product or piece of advertising. The problem
is that the interactions within the brain are so fast and
complex that it is not that easy to say that because a
particular part of the brain ‘lights up’ we are feeling
a particular emotion or thinking a particular thought.
In neuroscience there is a distinction between the
cognitive brain (the frontal cortex), which (consciously)
analyses, reflects, calculates and makes decisions,
and the emotional brain (the Limbic system), which
reacts spontaneously and intuitively. There is, however
constant two-way traffic between the two – any
information processed by the cognitive brain is sent
back to the Limbic system to be evaluated emotionally;
the reaction then goes back to the frontal cortex,
where it is interpreted as feeling.
Copyright © ESOMAR 2008
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
If our two ‘brains’ do not work in isolation and there
is constant communication between them, then surely
the most fascinating challenge for marketers must be
at the frontier between the two, where pre-conscious
impulses emerge, blinking into the light of consciousness,
as verbalised thoughts and feelings. Surely we need to
look for the links between the cognitive and emotional
brains – particularly those pre-cognitive processes that
underpin our conscious thoughts and utterances –
allowing us to make sense of things even before we
even give them conscious consideration.
By shifting the focus of survey research towards the
pre-cognitive (and away from cognitive or physiological
responses), we can start to focus on the mental
shortcuts that allow us to construe meaning without
having to think too deeply. It is highly likely that these can
provide the basis for an entirely new way of measuring
emotion – one that relies neither on cognitive response
nor on (complementary) physiological measures.
A simple example may be our instantaneous reaction
to certain facial expressions: for example, when we see
someone frown or smile, our brains interpret meaning
intuitively, without the need for considered thought.
One of the most promising means of understanding
pre-cognitive response is the cognitive linguistic
theory of metaphor.
According to this theory, metaphors can create emotional
meaning instantly and intuitively because they are
actually neural connections which create meaning
automatically in our unconscious mind, well below the
conscious linguistic surface. They arise because we
often (habitually) experience a particular feeling at the
same time as a bodily sensation or experience, thus
attaching a metaphorical meaning to that experience. The
emotion becomes conflated with the physical experience,
which becomes its metaphor, as in intimacy is closeness.
As Zaltman (2008) comments: ‘‘In many ways, deep
metaphors and emotions are siblings. Both are hardwired
in our brains … people experience them at some basic level
worldwide.’’ He argues that ‘‘… because metaphors and
emotions work hand in hand … it may be impossible to
understand the latter without the former.’’ That is certainly
also the view of Zoltán Kövescses (2000), who says:
‘‘Emotion language is largely metaphorical in English and
in all probability in other languages as well … to capture
the variety of diverse and intangible emotional experiences.
Methodologically, then, this (metaphorical) language is not
only a reflection of the experiences, but also creates them.
Simply put, we say what we feel and we feel what we say.”
The power of emotions in driving our behaviour is,
according to Kövesces, made plain in his “master
metaphor” for emotion – emotions are forces –
illustrated by the some of the following sentences:
• He was seized by emotion
• She was overcome by emotion
• His emotions ran away with him
Damasio (1999) observes: “We are about as effective
at stopping an emotion as stopping a sneeze’’. Indeed,
the use we make linguistically of the emotions are
forces metaphor suggests that this is something
we know and express intuitively. Each of the above
sentences convey the idea of emotion as an almost
irresistible physical force.
FINDING A SOLUTION
Going from the literal to metaphorical
It is deeply ironic that market researchers inhabit a
world of brands and concepts, yet, in measuring it, so
easily default to the prosaic and literal. After all, most
brands and advertising are metaphorical in some way,
yet even today, in 2008, most market research studies
are text-based and literal, not visual or metaphorical.
The problem with textual questions is that they usually
require cognitive processing, and respondents need to
consider their response before answering. The more
we think and consider our response, the further we get
from our emotions: the emotion may still be present,
but it becomes blended with our thinking and more
prone to rationalisation.
Clearly, we need to look for approaches that encourage
an intuitive and spontaneous, rather than a thoughtCopyright
© ESOMAR 2008
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
through response. Cognitive linguistics suggests that
metaphors have the power to express emotion more
vividly than literal language because they can evoke
an emotional response directly, without the need for
conscious, rational consideration. Furthermore, because
many metaphors are cross-cultural in their application,
they provide the basis for a universally relevant
measurement tool.
Using metaphors to express emotion
Metaphors seem fundamental to being human; even
the way we think about ourselves is fundamentally
metaphorical. For example, we talk of struggling to
gain control over ourselves, of our higher self battling
with our lower self, trying to find one’s true self.
When people talk about having ‘‘relationships’’ with a
brand, or a brand being like an ‘‘old friend’’ they are
using a metaphoric mode of speech to convey their
feelings (emotions) about it.
Take a close look at the following everyday sentences:
• I have a very warm relationship with her.
• I feel close to him.
• I jumped for joy.
• He’s a very distant sort of person.
• I’m feeling down today.
• He’s a very cold person.
• Things are looking up.
In each sentence, emotion is expressed by means of a
metaphor: warmth for affection, closeness for intimacy,
jumping (off the ground) for excitement and so on.
Perhaps the reason that we use metaphors so frequently
to describe how we feel is that it is very difficult to
describe these emotions without using a metaphor.
Most importantly, we seem to grasp the meaning of
metaphors intuitively. For example, when someone is
described as “distant”, our brain ‘knows’ (unconsciously)
that we are not talking about geographic proximity. Yet
how does it know this? Because literal communication
mostly involves the cognitive brain – which consciously
processes information – whereas metaphorical
communication seems to speak to the unconscious
parts of the brain, evoking images and creating meaning
without our being aware of what is happening. Thus
metaphors may appear to consist of words that appear
on the linguistic surface, but, underneath this surface,
their meaning is so immediate and powerful that there is
no need to compute or deduce it.
Singerland (2005) uses the example of ‘‘digging one’s
own financial grave’’ to illustrate the importance of
metaphor in evoking emotional reaction. It is a well
understood metaphor meaning that bad financial
decisions cause financial failure, but why import the
metaphor of grave digging at all? It does not seem to add
anything in terms of helping us to understand the literal
point; rather, its purpose is to evoke (negative) emotion,
to inspire (or help us imagine) the negative emotional
reactions associated with death, corpses and so on. In
other words, the metaphor adds emotive power to the
literal meaning.
Which metaphors are important to emotion?
The reason that metaphors are so powerful, according
to leading cognitive linguistics experts such as Lakoff and
Johnson and Kövesces, is that they are actually neural
connections, which create (automatic) meaning below
the conscious linguistic surface. The key idea here is
that metaphorical thought is based on bodily experience
and correlated neural activity in the brain. This leads to
the hypothesis that metaphorical meaning may become
attached to certain sensory experiences if we habitually
experience a particular feeling or emotion at the same
time as that experience.
The idea that metaphorical meaning is grounded in
our physical experience has led cognitive lingustics
experts to suggest that the metaphors which arise from
this process are universal in their application. Why?
If metaphor is based on the way the body and brain
function and human beings are alike at the level of this
functioning, then most of the metaphors that people use
must be fairly similar, if not universal. There is indeed
a great deal of evidence to suggest this, particularly for
the so-called primary metaphors, which seem to be
part of our unconscious mind, acquired automatically
and unconsciously, in our formative years, through the
conflation of subjective and sensory experiences.
Copyright © ESOMAR 2008
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
The Theory of Primary Metaphor derives from fundamental
insights by Christopher Johnson (1997) and Srini
Narayanan (1997), later developed into an integrated
theory by George Lakoff and Mark Johnson (1999).
It was Christopher Johnson (1997) who first argued
that primary metaphors are learned (in young children)
through the conflation of subjective and sensory
experiences. Thus, it is argued, affection is typically
correlated with the physical experience of warmth,
because most young children experience warmth
when being held affectionately in their mother’s arms.
Objectively, warmth is not even similar to affection, yet
the two become conflated, and amongst very young
children, the experience is undifferentiated, so that
the feeling and the sensory experience are felt to be
the same. As they develop, children start to separate
them, although the primary metaphors persist, leading
the infant, in later life, to speak of ‘‘a cold response’’ or
‘‘a warm relationship’’. Narayanan’s theory explains that
the connections made during conflation are ‘hardwired’
(instantiated) into our neural circuitry – producing a
stable, conventional system of primary metaphors that
tend to remain in place indefinitely and are independent
of language.
Visualising metaphors – going beyond language
We usually think of a metaphor as a set of words,
but, as we have seen, they actually exist beneath the
linguistic surface. Metaphors are essentially concepts,
which represent one image or idea in terms of
another, and therefore do not have to be expressed
linguistically. Humans understood emotions, such as
fear and affection, long before they had the capacity
to describe them linguistically, and so are able to
represent them through visual metaphors. We can
do this because so much of our thought is based
on images (not words). Indeed, most stimuli actually
reach the brain through the visual system, and so
we constantly translate verbal information into visual
imagery. Hence if we read or hear a word, what we
generally see (in our mind) is a visual representation
of the word, not the word itself.
Metaphors lend themselves to visual expression
because they are inherently conceptual: they are about
representing one thing in terms of another. If the visual
metaphor is instantly recognised, it will create meaning
automatically, without the need for conscious reflection.
Visual metaphors thus provide a powerful route into
emotion, avoiding the need for thinking or rationalisation
on the part of the respondent.
The visual nature of (emotional) metaphor lends itself
strongly to online research applications, and it is curious
that the first generation of online research more or
less ignored this. Early online researchers mainly
took conventional, text-heavy questionnaires and put
them online. This is particularly odd given the obvious
opportunities offered by online technology to visualise
and animate questionnaire formats. As Lanham (2007)
says: ‘‘The printed page depends on the economics
of deprival. No color, no movement, images in careful
moderation… the digital screen is the economics of plenty
(my italics). It allows competition between word, image
and sound for our attention.’’
It is disappointing how, even today, so few online
questionnaires fully exploit the opportunities of the
digital screen, or use its great potential for visual rather
than textual presentation of questionnaires. Yet digital
technology and Web 2.0 have given us the opportunity
to ask questions in novel and interactive ways that
really engage respondents. Perhaps the problem is
that researchers still regard online as a means of data
collection, whereas they should look on it as a medium.
DEVELOPMENT OF THE METAPHORIX™
APP ROACH
Assembling the building blocks
It seems that researchers have been debating and
discussing the importance of measuring emotion for
some years now, yet there has been relatively little
concrete progress in terms of either new measures or
metrics. The remainder of the paper is about how to
turn theory into practice, and describes how Conquest
Research developed and validated a new system for
measuring emotional response, called Metaphorix™.
Copyright © ESOMAR 2008
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
The foregoing discussion of emotion, metaphors and
Web 2.0 has highlighted a number of problems and
deficiencies in conventional approaches to measuring
emotion. These are summarised in table 1, and alongside
each problem is a proposed solution.
The Metaphorix™ project puts these ideas and hypotheses
into practice. It is about using animated visualisations
of metaphor to plug directly into consumers’
immediate emotional response (to brands, advertising
and other marketing stimuli), thus avoiding the over
rationalisation inherent in traditional approaches. All
Metaphorix™ measures are thus interactive online
animations that respondents can use intuitively – without
having to think or consider their response.
Development work on current animations commenced
in 2007, and comprised a series of depth interviews and
online observation sessions in which respondents were
introduced, online, to a range of ideas and executions for
visualised metaphors. A number of requirements and
guidelines emerged from this phase that inspired and
informed the development of the overall project.
1. The metaphors must be intuitively understood
Cognitive linguistic theory would suggest that if we
are to use metaphors for measuring emotions, they
should be intuitively understandable. If the metaphor
needs to be ‘processed’ (i.e. thought about) it defeats
the object of encouraging a spontaneous and, above
all, non-thought through response. Unsurprisingly, it
emerged from development work that visualisations
based on primary metaphors proved most effective in
this respect.
Thus, in developing Metaphorix™, considerable emphasis
was given to selecting metaphors that are both intuitive
and have potentially universal application. Clearly those
(primary) metaphors linked with subjective emotional
states – such as love, affection and intimacy – have
greatest relevance to the measurement of emotion. In
each case a particular (sensory) experience correlates
with the emotional judgement to create the metaphor.
(See figure 1.)
Thus if we consider the construction of the primary
metaphors illustrated in figure 1, it can be seen that
each represents an emotional state, in terms of a
specific sensory experience. For example, affection is
represented in terms of warmth, intimacy in terms of
physical proximity, happiness/sadness in terms of up/
down and so on. Each one finds also expression in
familiar linguistic expressions, for example:
• ‘‘I feel really cold towards him…’’
• ‘‘ I have a very close relationship with her …’’
• ‘‘ I feel very down today…’’
Ta ble 1
Problem Hypothesised Solution
When we think and consider we get further away from our
emotions
Encourage intuitive, spontaneous and non–thought through
response
When we talk about emotion we use metaphorical rather than
literal language
Allow respondents to express their emotions via metaphor
We think mostly in images not words, but conventional MR is
text based
Use measures that are visual and non-verbal, rather than text
based
Conventional online research doesn’t engage with respondents
as well as it should (hence compromising data quality)
Use a seamless visual interface and level of interactivity that
engages the Web.2.0 generation.
Use of language-based verbal scales (and in particular use of
top boxes) differs across cultural boundaries, making crosscultural
brand evaluation difficult
Visual animations based on primary metaphors transcend
language barriers and aid our interpretation
Copyright © ESOMAR 2008
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
2. The visual interface must fit seamlessly with
the medium
In the course of the development work, it became clear
that respondents engaged most rapidly and intuitively
with well-executed and well-finished graphics. Having
a high quality respondent interface is not therefore
merely a question of aesthetics, but a crucial means of
gaining and maintaining respondent engagement with
the interview process. Furthermore it has a functional
benefit, because the quality of the online interface
is crucial to how easily respondents actually use it.
Online users are now accustomed to an extremely high
quality of interface and react negatively towards poorly
executed visual graphics, dismissing them as ‘badly
drawn cartoons’ or ‘amateurish’. Poor graphics and
‘clunky’ interactivity actually get in the way, because they
encourage respondents to evaluate the medium rather
than the message, and defeat the object of what we
are trying to achieve – an instantaneous and intuitive
response.
We need to recognise that, from a consumer perspective,
online is not a ‘means of data collection’ but a medium.
What we present via that medium needs to be consistent
with the other experiences that the respondent may
have with, for example, Facebook, You Tube or with
any number of other high quality sites that they may
visit in a Web 2.0 environment. Just as poorly executed
TV commercials stand out (for all the wrong reasons),
poorly executed online studies impair the quality of
interaction/response – taking people further away
from their emotions and into unhelpful (cognitive)
judgement of the medium itself.
For these reasons, considerable investment was put in
to creating high quality animations for Metaphorix™, that
deliver an online experience equal to the best that can be
found in the Web 2.0 environment. A creative company
in London’s Soho was commissioned to produce the
designs and a specialist animation house appointed to
execute them.
3. Interactivity turns respondents into participants
Metaphorix™ is, essentially, a Web 2.0 product and the
essence of Web 2.0 is user interactivity. It is about
opening up access and allowing users to co-create. Tim
O’Reilly (who invented the term), describes it as ‘‘going
beyond the page metaphor of Web 1.0 to deliver rich
user experiences.” User engagement and involvement
are, therefore, not benefits of Web 2.0, they are the very
essence of it.
One of the guiding principles of the Metaphorix™
approach is that respondents should be enabled to
interact and co-create. It became evident that most
respondents share this aspiration, but many want more
than this – they actually want to enter the process by
projecting themselves into it. Encouragingly, this desire
is entirely consistent with some of the Metaphorix™
measures that ask the respondent to express his feelings
towards a brand through movement/ spatial relationship.
figure 1
some primar y me taphors
Copyright © ESOMAR 2008
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
For example, the visualisation of intimacy is proximity
invites a respondent to demonstrate how close they
feel to a brand by moving themselves towards or away
from it.
4. The importance of avatars
Whilst relatively few people actually visit Second Life
there is quite widespread familiarity with the idea
behind it: that a participant can represent himself by
means of an avatar and project himself into a scenario.
Early testing of Metaphorix™ indicated that this was a
highly engaging and effective idea, which has particular
relevance to and significance for the measurement of
emotion. After all, qualitative research practitioners
have, for some decades now, been using projective
techniques to help penetrate the unconscious mind by
eliciting feelings below the surface of conscious, rational
response.
Thus, in the first stage of the Metaphorix™ interview,
respondents are asked to choose an avatar to represent
themselves within each animation. The choice of avatars
has been specifically designed to represent gender
and racial diversity so that respondents can choose a
representation of themselves with which they feel both
empathetic and comfortable. The avatars validated for
use in Metaphorix™ are illustrated in figure 2.
International testing indicates that these avatars are
widely applicable across different markets, including
China. Table 2 demonstrates that respondents – interviewed
across four countries and three continents –
were comfortably able to pick an avatar from the set
of choices.
METAPHORIX™ IN ACTION
How to bring visual metaphors to life
Metaphorix™ currently comprises a range of five
metaphoric visualisations, all of which are presented
as interactive online animations that respondents can
use intuitively, without having to think or consider their
response.
figure 2
the me taphorix TM ava tars
© Copyright 2008 Conquest Research & Consultancy Ltd. All rights reserved.
The MetaphorixTM tool and brand name MetaphorixTM are the exclusive property of Conquest Research & Consultancy Ltd.
Copyright © ESOMAR 2008
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
Within each animation, respondents can move their selfrepresenting
avatar to show how they feel about a brand,
a piece of advertising, or any other marketing stimulus.
Each animated visual scenario represents a metaphor
associated with an emotion: for example, proximity/
closeness as a metaphor for intimacy (see figure 3).
The emotion is represented by a visual space in which
the respondent places his avatar in relation to another
person, an object, or value, such as a brand. The
respondent is given a non-leading instruction such as:
‘‘move your avatar to show how you feel about Brand
X’’. A key element of the respondent interface is that the
animations are free-flowing, allowing the avatar to move
seamlessly across the entire animation, without the
respondent being aware of any scaling or calibration.
The Metaphorix™ tool uses visualised metaphors
(of sensory experience) to represent five emotions/
feelings. These are summarised in table 3.
Q12. And which of these phrases best describes how you found picking a character to represent you in the survey?
Base: All respondents Language Gender
Total UK US Italy China Male Female
Unweighted Base 1228 306 306 308 307 612 616
Weighted Base 1228 306 306 308 307 614 614
% % % % % % %
It was fine - I had no problems choosing an avatar
that I wanted to represent me
77 74 76 85 71 80 74
It was ok - none of the avatars represented me
well, but I could choose one
22 24 23 14 28 19 26
Not good - none of the avatars were suitable to
represent me
1 1 1 1 * 1 *
Ta ble 2
The Ava tars Work Interna tional y
figure 3
the proximi ty me taphor in ac tion
© Copyright 2008 Conquest Research & Consultancy Ltd. All rights reserved.
The MetaphorixTM tool and brand name MetaphorixTM are the exclusive property of Conquest Research & Consultancy Ltd.
Copyright © ESOMAR 2008 10
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
For example, the metaphor affection is warmth is represented
visually by temperature (from hot to cold), since
the sensory experience that underpins the metaphor is
warmth. In each case there is no need for the animation
to have numbers or words, but behind the visual animation,
response can be scaled and calibrated quantitatively.
An example of output for the Proximity metaphor is
shown in figure 4, which shows the presentation and
calibration of data (using a 10 point scale), for three
different brands.
Figure 4 indicates that respondents are quite likely to
use the extremes of the scale – if they feel strongly
about the brand(s) in question – uninhibited by having
to give/choose a semantic response. Indeed, one of
the key findings to emerge from the validation exercise
is that metaphoric measures enable consumers to
express themselves more fully, which has the effect
of ‘pulling apart’ responses where standard measures
tend to produce a flat-line.
Ta ble 3
Emotion Visual metaphor used
Affection Temperature
Intimacy Proximity
Excitement Being off the ground
Happiness Facial expression
Empathy Approaching/Coming towards you
figure 4
proximi ty to bran d
Copyright © ESOMAR 2008 11
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
VALIDATION OF METAPHORIX™
The testing programme
Metaphorix™ was launched internationally, in May 2008,
after completion of an extensive (iterative/multi-country)
validation programme. It comprised over 5,000 interviews,
focussed on major international brands, and was
designed to test the following hypotheses:
1. That animated visual metaphoric scales better
capture consumers’ emotional response to brands
and advertising than conventional measures.
2. That this approach can be employed universally to
measure engagement with brands and advertising.
3. That these new metaphorical measures work effectively
alongside conventional verbal scales to enhance
insight into the contribution of emotion to building brand
share.
By exploring response to a wide range of brands and
advertising – some emotional, others more rationally
based – a balanced test design highlights where
conventional research falls short of capturing emotion,
and also shows how visual metaphoric measurement
can fill the gap left by conventional measures.
The two key phases of the programme were as follows:
Test 1 assessed 18 TV commercials, and comprised
3681 respondents, interviewed online in the United
Kingdom during March/April 2008.
The research approach and rationale is illustrated in
figure 5. Three matched cells were needed, for each
brand and advertisement, to establish the benefit of
using visual metaphors over a conventional approach
(to measuring emotion). The test design thus allowed
robust analysis of:
• The impact of exposure to advertising versus a
non-advertised cell (who saw only a brand logo).
• The effect of combining Metaphorix™ measures
with verbal scales versus a matched cell, which
was exposed to the same stimulus, but given only
conventional verbal scales.
figure 5
tes t i: overview of me tho dolog y
Copyright © ESOMAR 2008 12
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
Test II comprised 1,228 respondents interviewed online,
during May 2008, in China, USA, Italy and the United
Kingdom. It focussed on response to major international
brands across four categories within each country,
using both conventional and Metaphorix™ measures.
Each respondent assessed only one brand from each
category with order of exposure rotated using a balanced
incomplete block design. The test was designed to
explore consumer relationships with pairs of high equity
brands – for example, Coke and Pepsi, Nike and Adidas.
The methodology is illustrated in figure 6.
WHAT ARE THE BENEFITS OF METAPHORIX™
OVER CONVENTIONAL MEASURES?
The programme of testing revealed a number of
benefits of incorporating the Metaphorix™ approach
into brand and advertising evaluation. These benefits
are described below:
1. Enabling consumers to express themselves fully
The results of the validation programme indicate that
visualised animations allow respondents to express
their feelings – about brands or advertising – more
freely than do conventional verbal/semantic scales. In
particular, it appears that respondents are more likely
to use the top end of scales and less likely to use the
middle responses.
Figure 7 shows the comparison of results (from Test 1)
of a Metaphorix™ Brand Warmth scale (10 point) with
a comparable verbal scale (where 10 = I feel extremely
warm towards brand / 1 = I feel extremely cold towards
brand). Test 1 comprised matched independent samples
and, in each cell, respondents were asked to express
their feelings towards a brand, after seeing some TV
advertising, one cell using a metaphoric Brand Warmth
scale and the other the comparable verbal scale. Each
cell was matched to the other in terms of demographics
and brand usership, and was shown exactly the same
advertising stimulus.
The results indicate that respondents are significantly
more likely (when evaluating advertising metaphorically)
to use the top end of the scale and significantly less likely
to use the middle (4 - 7) responses.
This pattern of response is also evident internationally:
results from Test II (conducted across four countries)
indicate that the same pattern of response is evident
figure 6
tes t ii : overview of me tho dolog y
Copyright © ESOMAR 2008 13
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
for measures of both Empathy and Proximity (see
figure 8) when respondents are asked to assess major
international brands, such as Subway and McDonald’s,
Apple and Microsoft.
Why is “freedom of expression” important when
measuring emotional reactions to brands or to marketing
communication? The results of the IPA study (outlined
earlier in this paper) highlighted the difficulty of picking
‘winners’ using conventional advertising pre-testing,
particularly for advertising that speaks mainly to the
emotions. Conventional pre-testing typically measures
emotional engagement via scalar responses (on
dimensions such as likeability, enjoyment or involvement),
and this can produce an undifferentiated response,
making it difficult to pick out winners. This effect
was apparent in the test of 18 TV commercials in the
United Kingdom (Test 1) described earlier. Each of the
commercials submitted to this test focussed on nonrational
(emotional) communication.
The results are shown in figure 9, which compares
response to the 18 commercials using a conventional
5-point enjoyment scale with a Metaphorix™ Happiness
measure. Overall levels of response were, on average,
higher using the metaphorical measure, but, most
importantly, there are a number of cases where
response is significantly higher than one might expect.
Of most interest is the ‘winning’ commercial (the award
winning ‘Gorilla’ commercial for Cadbury’s Dairy Milk,
a major UK confectionery brand) which is significantly
ahead of the next best commercial on the metaphorical
measure, whereas on the conventional measure it
scores at a similar level to three other commercials.
Moreover, the percentage gap between this and the
bottom commercial is 42 percentage points on the
figure 7
bran d warm th
conven tional vs . me taphorix TM scale (uk )
Base: 3681 respondents in UK
Copyright © ESOMAR 2008 14
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
conventional measure, and 63 percentage points on
the metaphorical scale.
These results support the hypothesis that the visual/
metaphorical approach allows respondents to express
themselves in a more unfettered, spontaneous manner,
which helps to discriminate between different advertising
executions, particularly those – like Cadbury’s Gorilla –
that appeal directly to the emotions. Response to the
commercial for Brand Z is also particularly interesting
in this respect. Brand Z is a leading UK beer brand,
aimed at young men and using the creative theme of
male bonding in its advertising. Again, the intent of the
advertising is to appeal to the emotions rather than
communicate a rational benefit. Respondents of this age
and gender are notoriously difficult to interview about
their emotions, and it seems likely that the conventional
verbal scales constrained their response in this instance,
whilst the visualised metaphoric scale seemed to
encourage a ‘freer’ and less constrained response.
(See figure 9.)
It can be also be seen in figure 9 that, in a number of
cases, there was no significant difference between
the scores on the metaphoric and conventional scales,
indicating that where difference does arise, it is because
the metaphorical approach is better able to capture
emotional response.
2. More sensitive measure of brand engagement
shifts
The ‘test and control’ methodology of Test I also enables
an assessment of the ability of metaphoric measures
to detect underlying movements in brand engagement.
In the control cells, respondents saw no advertising
and were asked to rate the brands (represented by
their logos) using both Metaphorix™ measures and
figure 8
conven tional vs . me taphorix tm scales - in terna tional com parison
Base: 1227 in UK, USA, China and Italy (approx 300 per country)
Copyright © ESOMAR 2008 15
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
conventional brand measures, such as brand disposition
and purchase propensity. In the parallel test cells,
respondents were shown TV commercials (for the same
brands) and asked to rate the brands using the same
mixture of metaphorical measures and conventional
brand measures. Comparison of movements between
the two cells can thus tell us how well each approach
identifies brand engagement shifts engendered by
exposure to advertising.
The results show (see figure 10) that conventional
persuasion measures produced only two significant
shifts between the test and control cells – for brands
B and M – one moving positively and the other moving
negatively. The metaphorical measures were generally
more sensitive to movements in brand engagement
engendered by exposure to the commercials. Six of the
commercials (33% of the total) produced a significant
positive shift on at least two of the three Metaphorix™
measures (Empathy, Warmth and Proximity) versus the
control cell, with no corresponding movement on the
conventional measures.
3. Making the interview more engaging
In previous sections, we discussed the importance of
engaging respondents with the interview process. Thus,
at the end of each interview, respondents were asked
to say how much they enjoyed taking part in it, and to
give their comments. Comparison of response in cells
where Metaphorix™ measures were used versus those
where only conventional measures were employed,
allows examination of the effect of including such scales
on respondent engagement and survey enjoyment.
Figure 11 shows that, in the United Kingdom, levels of
survey enjoyment were significantly enhanced (over a
parallel conventional study), with approximately 60%
saying “I really enjoyed it”. A similar outcome was
evident in the United States, and in both China and Italy,
levels reached c. 70%. The levels of survey enjoyment
observed across different countries suggests therefore
that the Metaphorix™ approach has wide cross-cultural
acceptance and applicability.
figure 9
adver tising engaemen t: conven tional vs . me taphorix tm measures
Base: 3681 respondents in UK
Copyright © ESOMAR 2008 16
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
figure 10
com parison of bran d engagemen t shif ts, pos t adver tising :
persuasion shif ts vs . me taphorical scales
figure 11
surve y en joymen t: conven tional vs . me taphorix tm
Copyright © ESOMAR 2008 17
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
4. Predicting and diagnosing brand success
Can metaphoric scales predict the success of a brand?
For two years prior to launching Metaphorix™(in May
2008), Conquest Research incorporated experimental
visual metaphoric measures (alongside more conventional
questions) into its online brand health and advertising
studies in the United Kingdom and elsewhere. Over 120
brands have now been assessed across more than
20 categories.
These tracking data allow us to test the hypothesis that
metaphoric scales are predictive of brand success by
observing the relationship between a dependent and
independent variable. The dependent variable is brand
share/first choice brand, and the independent variable is
the proportion that placed the same brand at the top
end (top three deciles) of a visual brand warmth scale.
This relationship (illustrated in figure 12) shows that brand
warmth is a statistically significant predictor of brand
share/first choice. The high correlation (The Pearson
Correlation Coefficient is 0.76 and p>0.01) provides strong
evidence for the effectiveness of metaphoric scales
as both a measure of emotional engagement and as a
predictor of success in the market.
Figure 12 also indicates that successful brands tend to
enjoy high levels of emotional engagement. Generally, the
higher the level of engagement, the higher the market
share, which is perhaps unsurprising, but figure 12 also
indicates that high engagement does not always create
successful brands. Thus there are outlying brands that
have a high level of warmth, but low brand choice; and
those with low warmth and high choice.
figure 12
me taphors can pre dic t bran d suces
Copyright © ESOMAR 2008 18
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
What does this tell us? Where engagement is high relative
to brand share, it may be because recent experience
or knowledge of the brand is too patchy or too weak to
forge a strong link between the brand and what it offers.
Alternatively, it may be that the brand is not marketing
itself strongly enough; some brands have a huge
emotional potential, which is not fully exploited by its
marketers. Engagement is not enough to drive sales –it
has to be supported by a good price, product offer and
widespread availability.
Where there is low brand engagement, but high brand
share, it may be because the brand is offering an
attractive set of product features, price, deals, etc., but
is not creating enough engagement with its customers.
Engagement cannot usually be bought through price,
deals or promotion – it usually requires sustained, brandled,
communications activity.
It is also possible to combine metaphorical measures
of engagement with more conventional ones (such as
brand consideration) to segment consumers in terms
of their rational and emotional commitment to different
brands. Table 4 shows an analysis of two brands
from the UK cell phone network market. Brand A has
converted a high percentage of those who are warm
toward the brand, but has a high percentage of Cool
Considerers, who are not particularly positive. There is
probably a need to forge a stronger emotional connection
with the brand before these consumers would choose
it. For Brand B, on the other hand, the opportunity may
be to build a relationship with the Warm Prospects, who
already view the brand positively, but may not yet have
been given a reason to buy. A campaign focussing on
rational benefits might well resonate with this group.
5. A cross cultural measure of engagement with
brands and advertising?
Encouragingly, it appears that most the benefits of
metaphoric measurement, described above, are available
to researchers worldwide. We have seen that levels of
respondent engagement – across markets as diverse
as Italy and China – are consistently high when visual
metaphoric scales are used. The hypothesis that visual
metaphorical measures work across cultures is further
supported by very positive open-ended comments,
which suggest that respondents in different cultures
enjoy, understand and engage with the interview
process. Moreover, the high correlations observed
internationally (see figure 13) between Metaphorix™
measures (of Proximity, Warmth and Empathy) and
conventional brand choice/consideration scales suggest
that visual metaphorical measures are predictive of
brand success in a range of international markets.
Ta ble 4
Bran d Warm th vs . Rational Consi dera tion : Drive Matrix
Brand Engagement
BRAND A BRAND B
DRIVE MATRIX COLD LUKEWARM WARM COLD LUKEWARM WARM
1st Choice 0% 4% Committed
10%
0% 0% Committed
4%
Would Consider 1% Cool
considerers
17%
Warm
Prospects
8%
1% Cool
considerers
8%
Warm
Prospects
12 %
Not in
Consideration
Set
22% 33% 6% 32% 42% 2%
Consideration
Copyright © ESOMAR 2008 19
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
Clearly a lot more work remains to be done, but the
results of the Metaphorix™ project thus far provide
strong initial evidence both for the effectiveness of visual
metaphor as a means of measuring emotion, and for
employing online metaphoric scales universally, across
cultures.
CONCLUSION
Let’s get animated about emotion
It is high time we stopped talking about measuring
emotion and started doing something about it. The
Metaphorix™ approach – based on the applying the
cognitive linguistic theory of metaphor to the world of
Web 2.0 – suggests that animated visual metaphors can
work where most conventional approaches fail. Given
the burgeoning understanding of the emotional unconscious,
it is astonishing that, even now, in 2008, most
conventional market research is text-based and literal,
not visual or metaphorical. Most conventional research
actually divorces respondents from their emotions – the
emotion may still be present, but it becomes more prone
to rationalisation as they consider their response.
Whilst neurological/biological measures seem to
provide an exciting alternative, even their strongest
proponents accept that they are at best a complement to
conventional approaches, not a replacement for them.
There is strong evidence both for the effectiveness of
visual metaphor as a means of measuring emotion,
and for employing online metaphoric scales universally,
across cultures. These measures seem to go beyond
purely rational response, freeing respondents to express
feelings and emotions which conventional (verbal)
approaches miss or suppress. They can thus better
capture emotional response and, when employed
alongside rational measures, provide a commercial
insight into brand health denied to conventional
approaches alone.
Marketing is all about imagination and metaphor, yet
we so easily default to prosaic and literal means of
measuring it. By sticking with the conventional paradigm,
online research has thus far failed to exploit the most
exciting opportunity that Web 2.0 offers – to turn
respondents into participants by fully engaging them with
the interview process. It is time to seize the opportunities
that the medium can offer and get truly animated
about emotion!
figure 13
in terna tional me taphorix tm corela tions wi th bran d choice /consi dera tion
Copyright © ESOMAR 2008 20
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
References
Binet, Les and Field, Peter (2007). Marketing in the Era of
Accountability. WARC, Henley on Thames, Oxon.
Damasio, Antonio (1994). Descartes’ Error. GP Putnam’s Sons, New
York
Damasio, Antonio (1999). The Feeling of What Happens. A Harvest
Book, Harcourt Inc, San Diego, CA,
Franzen, Giep and Bauwman, Margot (2001). The Mental World of
Brands. WARC, Henley on Thames, Oxon.
Grady, J. (1997). Foundations of Meaning: Primary Metaphors
and Primary Scenes. Ph.D. dissertation, University of California,
Berkeley.
Johnson, Christopher (1997). Metaphor vs. Conflation in the
Acquisition of Polysemy. In M.K. Hiraga, C. Sonha, and S. Wilcox,
eds. Current Issues in Linguistic Theory 152. John Benjamins,
Amsterdam.
Johnson, Mark (1987). The Body in the Mind: The Bodily Basis of
Meaning, Imagination and Reason. University of Chicago Press,
Chicago.
Johnson, Christopher (1997). Learnability in the Acquisition of
Multiple Senses: SOURCE Reconsidered. In J. Moxley, J. Juge, and
M. Juge, eds., Proceedings of the Twenty-Second annual Meeting
of The Berkeley Linguistic Society. Berkeley Linguistics Society,
Berkeley.
Kövesces, Zoltán (2000). Metaphor and Emotion: Language, Culture
and Body in Human Feeling. Cambridge University Press.
Kövesces, Zoltán (2002). Metaphor in Culture (2005). Cambridge
University Press.
Lanham, Richard. A (2006) The Economics of Attention: Style and
Substance in the Age of Information. University of Chicago Press.
©2006
Lakoff, George and Mark Johnson (1999). Philosophy In The Flesh.
Basic Books, New York, NY.
Le Doux, Joseph (2002). Synaptic Self. Penguin Books, London.
Le Doux, Joseph (1996). The Emotional Brain. Touchstone, Simon
and Schuster, New York.
Mast, Fred and Zaltman, Gerald (2006). Anatomy of Engagement
in Definitions about the Anatomy of Engagement, The 52nd ARF
Convention, 2006.
Mc Clure, Samuel M; Li, Jian; Tomlin, Damon; Cypert, Kim S;
Montague, Latane M and Montague, P.Read. (2004). Neural
Correlates of Behavioural Preference for Culturally Familiar Drinks.
Neuron, Vol. 44, 379-387. Cell Press.
Narayanan, Srini (1997). Embodiment in Language Understanding:
Sensory Motor Representations for Metaphoric Reasoning About
event Descriptions. Ph.D. Dissertation, Dept of Computer Science,
University of California, Berkeley.
Pinker, Steven (2008). The Stuff of Thought. Allen Lane, London,
England.
Singerland, Edward. G. (2005). Conceptual Blending, Somatic
Marking and Normativity. Cognitive Linguistics 16-3. Pages 557-564
Zaltman, Gerald (2003). How Customers Think. Harvard Business
Press, Boston.
Zaltman, Gerald (20008). Marketing Metaphoria. Harvard University
Press
The Author
David Penn is Co-founder and Managing Director, Conquest
Research, United Kingdom.
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
Getting anima ted abou t emo tion
The new frontier
David Penn
LOOKING FOR EMOTIONAL ENGAGEMENT
The story so far
We’ve been getting quite animated about emotional
engagement for some years now, but we still cannot
agree on how to measure it. What almost everyone
does agree on, however, is that conventional quantitative
research methods – with their emphasis on conscious,
cognitive and rational response – fall short when it
comes to measuring emotion.
Gerald Zaltman (2003) suggests that marketing
researchers put so much emphasis on the cognitive and
rational because it is easier to think about the consumer
that way and to frame questions that way. Perhaps, but
it may also be because the inception of survey research
(as we know it today) coincided with the rise of mid–20th
century cognitive science and predated the neuroscientific
revolution at the end of the century.
Cognitive Science was mainly concerned with the mind’s
conscious mental processes – rather than with the
unconscious or the emotional – and was the dominant
paradigm until, in the late 20th century, the work of
neuroscientists like Antonio Damasio and Joseph Le
Doux challenged its assumptions about how the mind
works. Neuroscience starts from the premise that we
cannot understand the mind without first understanding
underlying brain structures and processes. In other
words, thinking and feeling are not separate from the
brain, but neurally embodied in it.
Like it or not, our feelings, our reasoning, our conscious
and unconscious thoughts are framed by our emotions
and bodily responses. It is therefore not enough to ask
people how they consciously feel about something,
because emotion occurs below the level of conscious
cognitive response. We can say how we feel, but we
find it very difficult to say why we came to feel that
way because the processes by which our thoughts and
feelings are created are mainly invisible to us.
The limitations of verbal response
Most conventional methods of measuring “how
consumers feel” are still based on expressed/verbal
response. Typically researchers ask people how
they feel and then ask them to explain why they feel
that way. Moreover, the measures which conventional
research uses to measure emotion are languagebased
and literal – the problem being that respondents
need to consider their response before answering. Yet
experience tells us that the more we think and consider
our response the further we get from our emotions,
because we start to post rationalise. Hence verbal
questioning (both quantitative and qualitative) struggles
to measure emotion, because it seems to mainly engage
our cognitive brain – the one that (consciously) analyses,
reflects, calculates and makes decisions – rather
than our emotional brain, that reacts spontaneously,
immediately and intuitively.
The limitations of verbal questioning in measuring
emotions has been further brought into focus by recent
important work into how advertising works. Binet and
Field (2007) assessed 880 (UK) case studies from the
IPA’s rigorous effectiveness awards scheme, concluding
that “communications models that use emotional appeal
are more likely to yield strong business results than
rationally based models (information and persuasion)”.
They also went on to draw the startling conclusion that
current conventional advertising pre-testing may even
reduce effectiveness; advertising that had favourable
pre-testing did significantly worse (in terms of business
Copyright © ESOMAR 2008
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
performance) than that which did not. The authors
conclude “the data are clearly not suggesting that pretesting
is entirely worthless, but do cast considerable
doubt on the ability of such research to reliably pick
winners”.
Unsurprisingly, some have lost faith with conventional
research altogether and opted for techniques (such
as brain imaging) that identify which parts of the brain
are active when we feel emotion. There was much
excitement, for example, a few years ago when a
completely new kind of market research study hit the
headlines – at the Baylor Institute in Houston, neuroscientists
used fMRI to monitor brain response during a
test of taste and brand preferences for Pepsi vs. Coke.
The neuroscientists (McClure et al, 2004) discovered that
in a blind product test (where preference for Pepsi vs.
Coke was balanced) only those parts of the brain relating
to sensory judgement were active. When respondents
were told what they were drinking, however, preference
switched in Coke’s favour, and a part of the brain (the
Hippocampus) associated with emotional response, also
became active. Brain imaging showed that knowledge
of Coke branding activated areas of the brain associated
with emotional judgement, whilst knowledge of Pepsi had
no corresponding effect. The study’s authors concluded:
‘‘Our finding suggests that the hippocampus may
participate in recalling cultural information that biases
preference judgements.’’
Excited journalists were soon hailing the birth of “neuromarketing”
and writing articles with titles such as “In
Search of the Buy Button” or even “Pushing the Buy
Button”. A new era seemed to be dawning, in which
conventional research would give way to techniques
based on brain imaging and physiological response.
So, what has happened, four years on from the birth
of neuromarketing? There have been some advances
and some worthwhile studies – particularly in the area
of media engagement – and we should also note the
re-emergence of bio-metrics alongside neuromarketing.
Based on measures of skin conductance, heart rate
and respiration, bio-measurement actually dates from
the early 20th century, but modern neuroscientific understanding
combined with advances in technology have
made its application to modern market research possible.
But can these techniques tell us something that conventional
research cannot? Actually, when it comes to
emotion, neurological/biological measurement can
generally tell us what, but not why. Mast and Zaltman
(2006) nail the problem perfectly with their observation
that ‘‘emotion without cognitive appraisal is really just
arousal’’. In other words, simply observing a neurological
or biological response is not the same as understanding
an emotion – because one is about physiology, the other
about cognitive appraisal. Indeed, most of the proponents
of neurological/biological approaches admit that their
techniques are not alternatives, but complements to
conventional research, either quantitative or qualitative.
Why? Because they have no way of understanding what
a respondent is feeling at the time of observation other
than by asking a conventional question. All of this is
somewhat ironic (and circular) when we consider that
neuromarketing and bio-metrics developed because
conventional research was not up to the task of
measuring emotion.
How do we go forward?
Neuromarketing is, essentially, about using brain imaging
techniques to locate activity in the brain and then
correlating that observation with mental response – to
a brand, product or piece of advertising. The problem
is that the interactions within the brain are so fast and
complex that it is not that easy to say that because a
particular part of the brain ‘lights up’ we are feeling
a particular emotion or thinking a particular thought.
In neuroscience there is a distinction between the
cognitive brain (the frontal cortex), which (consciously)
analyses, reflects, calculates and makes decisions,
and the emotional brain (the Limbic system), which
reacts spontaneously and intuitively. There is, however
constant two-way traffic between the two – any
information processed by the cognitive brain is sent
back to the Limbic system to be evaluated emotionally;
the reaction then goes back to the frontal cortex,
where it is interpreted as feeling.
Copyright © ESOMAR 2008
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
If our two ‘brains’ do not work in isolation and there
is constant communication between them, then surely
the most fascinating challenge for marketers must be
at the frontier between the two, where pre-conscious
impulses emerge, blinking into the light of consciousness,
as verbalised thoughts and feelings. Surely we need to
look for the links between the cognitive and emotional
brains – particularly those pre-cognitive processes that
underpin our conscious thoughts and utterances –
allowing us to make sense of things even before we
even give them conscious consideration.
By shifting the focus of survey research towards the
pre-cognitive (and away from cognitive or physiological
responses), we can start to focus on the mental
shortcuts that allow us to construe meaning without
having to think too deeply. It is highly likely that these can
provide the basis for an entirely new way of measuring
emotion – one that relies neither on cognitive response
nor on (complementary) physiological measures.
A simple example may be our instantaneous reaction
to certain facial expressions: for example, when we see
someone frown or smile, our brains interpret meaning
intuitively, without the need for considered thought.
One of the most promising means of understanding
pre-cognitive response is the cognitive linguistic
theory of metaphor.
According to this theory, metaphors can create emotional
meaning instantly and intuitively because they are
actually neural connections which create meaning
automatically in our unconscious mind, well below the
conscious linguistic surface. They arise because we
often (habitually) experience a particular feeling at the
same time as a bodily sensation or experience, thus
attaching a metaphorical meaning to that experience. The
emotion becomes conflated with the physical experience,
which becomes its metaphor, as in intimacy is closeness.
As Zaltman (2008) comments: ‘‘In many ways, deep
metaphors and emotions are siblings. Both are hardwired
in our brains … people experience them at some basic level
worldwide.’’ He argues that ‘‘… because metaphors and
emotions work hand in hand … it may be impossible to
understand the latter without the former.’’ That is certainly
also the view of Zoltán Kövescses (2000), who says:
‘‘Emotion language is largely metaphorical in English and
in all probability in other languages as well … to capture
the variety of diverse and intangible emotional experiences.
Methodologically, then, this (metaphorical) language is not
only a reflection of the experiences, but also creates them.
Simply put, we say what we feel and we feel what we say.”
The power of emotions in driving our behaviour is,
according to Kövesces, made plain in his “master
metaphor” for emotion – emotions are forces –
illustrated by the some of the following sentences:
• He was seized by emotion
• She was overcome by emotion
• His emotions ran away with him
Damasio (1999) observes: “We are about as effective
at stopping an emotion as stopping a sneeze’’. Indeed,
the use we make linguistically of the emotions are
forces metaphor suggests that this is something
we know and express intuitively. Each of the above
sentences convey the idea of emotion as an almost
irresistible physical force.
FINDING A SOLUTION
Going from the literal to metaphorical
It is deeply ironic that market researchers inhabit a
world of brands and concepts, yet, in measuring it, so
easily default to the prosaic and literal. After all, most
brands and advertising are metaphorical in some way,
yet even today, in 2008, most market research studies
are text-based and literal, not visual or metaphorical.
The problem with textual questions is that they usually
require cognitive processing, and respondents need to
consider their response before answering. The more
we think and consider our response, the further we get
from our emotions: the emotion may still be present,
but it becomes blended with our thinking and more
prone to rationalisation.
Clearly, we need to look for approaches that encourage
an intuitive and spontaneous, rather than a thoughtCopyright
© ESOMAR 2008
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
through response. Cognitive linguistics suggests that
metaphors have the power to express emotion more
vividly than literal language because they can evoke
an emotional response directly, without the need for
conscious, rational consideration. Furthermore, because
many metaphors are cross-cultural in their application,
they provide the basis for a universally relevant
measurement tool.
Using metaphors to express emotion
Metaphors seem fundamental to being human; even
the way we think about ourselves is fundamentally
metaphorical. For example, we talk of struggling to
gain control over ourselves, of our higher self battling
with our lower self, trying to find one’s true self.
When people talk about having ‘‘relationships’’ with a
brand, or a brand being like an ‘‘old friend’’ they are
using a metaphoric mode of speech to convey their
feelings (emotions) about it.
Take a close look at the following everyday sentences:
• I have a very warm relationship with her.
• I feel close to him.
• I jumped for joy.
• He’s a very distant sort of person.
• I’m feeling down today.
• He’s a very cold person.
• Things are looking up.
In each sentence, emotion is expressed by means of a
metaphor: warmth for affection, closeness for intimacy,
jumping (off the ground) for excitement and so on.
Perhaps the reason that we use metaphors so frequently
to describe how we feel is that it is very difficult to
describe these emotions without using a metaphor.
Most importantly, we seem to grasp the meaning of
metaphors intuitively. For example, when someone is
described as “distant”, our brain ‘knows’ (unconsciously)
that we are not talking about geographic proximity. Yet
how does it know this? Because literal communication
mostly involves the cognitive brain – which consciously
processes information – whereas metaphorical
communication seems to speak to the unconscious
parts of the brain, evoking images and creating meaning
without our being aware of what is happening. Thus
metaphors may appear to consist of words that appear
on the linguistic surface, but, underneath this surface,
their meaning is so immediate and powerful that there is
no need to compute or deduce it.
Singerland (2005) uses the example of ‘‘digging one’s
own financial grave’’ to illustrate the importance of
metaphor in evoking emotional reaction. It is a well
understood metaphor meaning that bad financial
decisions cause financial failure, but why import the
metaphor of grave digging at all? It does not seem to add
anything in terms of helping us to understand the literal
point; rather, its purpose is to evoke (negative) emotion,
to inspire (or help us imagine) the negative emotional
reactions associated with death, corpses and so on. In
other words, the metaphor adds emotive power to the
literal meaning.
Which metaphors are important to emotion?
The reason that metaphors are so powerful, according
to leading cognitive linguistics experts such as Lakoff and
Johnson and Kövesces, is that they are actually neural
connections, which create (automatic) meaning below
the conscious linguistic surface. The key idea here is
that metaphorical thought is based on bodily experience
and correlated neural activity in the brain. This leads to
the hypothesis that metaphorical meaning may become
attached to certain sensory experiences if we habitually
experience a particular feeling or emotion at the same
time as that experience.
The idea that metaphorical meaning is grounded in
our physical experience has led cognitive lingustics
experts to suggest that the metaphors which arise from
this process are universal in their application. Why?
If metaphor is based on the way the body and brain
function and human beings are alike at the level of this
functioning, then most of the metaphors that people use
must be fairly similar, if not universal. There is indeed
a great deal of evidence to suggest this, particularly for
the so-called primary metaphors, which seem to be
part of our unconscious mind, acquired automatically
and unconsciously, in our formative years, through the
conflation of subjective and sensory experiences.
Copyright © ESOMAR 2008
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
The Theory of Primary Metaphor derives from fundamental
insights by Christopher Johnson (1997) and Srini
Narayanan (1997), later developed into an integrated
theory by George Lakoff and Mark Johnson (1999).
It was Christopher Johnson (1997) who first argued
that primary metaphors are learned (in young children)
through the conflation of subjective and sensory
experiences. Thus, it is argued, affection is typically
correlated with the physical experience of warmth,
because most young children experience warmth
when being held affectionately in their mother’s arms.
Objectively, warmth is not even similar to affection, yet
the two become conflated, and amongst very young
children, the experience is undifferentiated, so that
the feeling and the sensory experience are felt to be
the same. As they develop, children start to separate
them, although the primary metaphors persist, leading
the infant, in later life, to speak of ‘‘a cold response’’ or
‘‘a warm relationship’’. Narayanan’s theory explains that
the connections made during conflation are ‘hardwired’
(instantiated) into our neural circuitry – producing a
stable, conventional system of primary metaphors that
tend to remain in place indefinitely and are independent
of language.
Visualising metaphors – going beyond language
We usually think of a metaphor as a set of words,
but, as we have seen, they actually exist beneath the
linguistic surface. Metaphors are essentially concepts,
which represent one image or idea in terms of
another, and therefore do not have to be expressed
linguistically. Humans understood emotions, such as
fear and affection, long before they had the capacity
to describe them linguistically, and so are able to
represent them through visual metaphors. We can
do this because so much of our thought is based
on images (not words). Indeed, most stimuli actually
reach the brain through the visual system, and so
we constantly translate verbal information into visual
imagery. Hence if we read or hear a word, what we
generally see (in our mind) is a visual representation
of the word, not the word itself.
Metaphors lend themselves to visual expression
because they are inherently conceptual: they are about
representing one thing in terms of another. If the visual
metaphor is instantly recognised, it will create meaning
automatically, without the need for conscious reflection.
Visual metaphors thus provide a powerful route into
emotion, avoiding the need for thinking or rationalisation
on the part of the respondent.
The visual nature of (emotional) metaphor lends itself
strongly to online research applications, and it is curious
that the first generation of online research more or
less ignored this. Early online researchers mainly
took conventional, text-heavy questionnaires and put
them online. This is particularly odd given the obvious
opportunities offered by online technology to visualise
and animate questionnaire formats. As Lanham (2007)
says: ‘‘The printed page depends on the economics
of deprival. No color, no movement, images in careful
moderation… the digital screen is the economics of plenty
(my italics). It allows competition between word, image
and sound for our attention.’’
It is disappointing how, even today, so few online
questionnaires fully exploit the opportunities of the
digital screen, or use its great potential for visual rather
than textual presentation of questionnaires. Yet digital
technology and Web 2.0 have given us the opportunity
to ask questions in novel and interactive ways that
really engage respondents. Perhaps the problem is
that researchers still regard online as a means of data
collection, whereas they should look on it as a medium.
DEVELOPMENT OF THE METAPHORIX™
APP ROACH
Assembling the building blocks
It seems that researchers have been debating and
discussing the importance of measuring emotion for
some years now, yet there has been relatively little
concrete progress in terms of either new measures or
metrics. The remainder of the paper is about how to
turn theory into practice, and describes how Conquest
Research developed and validated a new system for
measuring emotional response, called Metaphorix™.
Copyright © ESOMAR 2008
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
The foregoing discussion of emotion, metaphors and
Web 2.0 has highlighted a number of problems and
deficiencies in conventional approaches to measuring
emotion. These are summarised in table 1, and alongside
each problem is a proposed solution.
The Metaphorix™ project puts these ideas and hypotheses
into practice. It is about using animated visualisations
of metaphor to plug directly into consumers’
immediate emotional response (to brands, advertising
and other marketing stimuli), thus avoiding the over
rationalisation inherent in traditional approaches. All
Metaphorix™ measures are thus interactive online
animations that respondents can use intuitively – without
having to think or consider their response.
Development work on current animations commenced
in 2007, and comprised a series of depth interviews and
online observation sessions in which respondents were
introduced, online, to a range of ideas and executions for
visualised metaphors. A number of requirements and
guidelines emerged from this phase that inspired and
informed the development of the overall project.
1. The metaphors must be intuitively understood
Cognitive linguistic theory would suggest that if we
are to use metaphors for measuring emotions, they
should be intuitively understandable. If the metaphor
needs to be ‘processed’ (i.e. thought about) it defeats
the object of encouraging a spontaneous and, above
all, non-thought through response. Unsurprisingly, it
emerged from development work that visualisations
based on primary metaphors proved most effective in
this respect.
Thus, in developing Metaphorix™, considerable emphasis
was given to selecting metaphors that are both intuitive
and have potentially universal application. Clearly those
(primary) metaphors linked with subjective emotional
states – such as love, affection and intimacy – have
greatest relevance to the measurement of emotion. In
each case a particular (sensory) experience correlates
with the emotional judgement to create the metaphor.
(See figure 1.)
Thus if we consider the construction of the primary
metaphors illustrated in figure 1, it can be seen that
each represents an emotional state, in terms of a
specific sensory experience. For example, affection is
represented in terms of warmth, intimacy in terms of
physical proximity, happiness/sadness in terms of up/
down and so on. Each one finds also expression in
familiar linguistic expressions, for example:
• ‘‘I feel really cold towards him…’’
• ‘‘ I have a very close relationship with her …’’
• ‘‘ I feel very down today…’’
Ta ble 1
Problem Hypothesised Solution
When we think and consider we get further away from our
emotions
Encourage intuitive, spontaneous and non–thought through
response
When we talk about emotion we use metaphorical rather than
literal language
Allow respondents to express their emotions via metaphor
We think mostly in images not words, but conventional MR is
text based
Use measures that are visual and non-verbal, rather than text
based
Conventional online research doesn’t engage with respondents
as well as it should (hence compromising data quality)
Use a seamless visual interface and level of interactivity that
engages the Web.2.0 generation.
Use of language-based verbal scales (and in particular use of
top boxes) differs across cultural boundaries, making crosscultural
brand evaluation difficult
Visual animations based on primary metaphors transcend
language barriers and aid our interpretation
Copyright © ESOMAR 2008
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
2. The visual interface must fit seamlessly with
the medium
In the course of the development work, it became clear
that respondents engaged most rapidly and intuitively
with well-executed and well-finished graphics. Having
a high quality respondent interface is not therefore
merely a question of aesthetics, but a crucial means of
gaining and maintaining respondent engagement with
the interview process. Furthermore it has a functional
benefit, because the quality of the online interface
is crucial to how easily respondents actually use it.
Online users are now accustomed to an extremely high
quality of interface and react negatively towards poorly
executed visual graphics, dismissing them as ‘badly
drawn cartoons’ or ‘amateurish’. Poor graphics and
‘clunky’ interactivity actually get in the way, because they
encourage respondents to evaluate the medium rather
than the message, and defeat the object of what we
are trying to achieve – an instantaneous and intuitive
response.
We need to recognise that, from a consumer perspective,
online is not a ‘means of data collection’ but a medium.
What we present via that medium needs to be consistent
with the other experiences that the respondent may
have with, for example, Facebook, You Tube or with
any number of other high quality sites that they may
visit in a Web 2.0 environment. Just as poorly executed
TV commercials stand out (for all the wrong reasons),
poorly executed online studies impair the quality of
interaction/response – taking people further away
from their emotions and into unhelpful (cognitive)
judgement of the medium itself.
For these reasons, considerable investment was put in
to creating high quality animations for Metaphorix™, that
deliver an online experience equal to the best that can be
found in the Web 2.0 environment. A creative company
in London’s Soho was commissioned to produce the
designs and a specialist animation house appointed to
execute them.
3. Interactivity turns respondents into participants
Metaphorix™ is, essentially, a Web 2.0 product and the
essence of Web 2.0 is user interactivity. It is about
opening up access and allowing users to co-create. Tim
O’Reilly (who invented the term), describes it as ‘‘going
beyond the page metaphor of Web 1.0 to deliver rich
user experiences.” User engagement and involvement
are, therefore, not benefits of Web 2.0, they are the very
essence of it.
One of the guiding principles of the Metaphorix™
approach is that respondents should be enabled to
interact and co-create. It became evident that most
respondents share this aspiration, but many want more
than this – they actually want to enter the process by
projecting themselves into it. Encouragingly, this desire
is entirely consistent with some of the Metaphorix™
measures that ask the respondent to express his feelings
towards a brand through movement/ spatial relationship.
figure 1
some primar y me taphors
Copyright © ESOMAR 2008
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
For example, the visualisation of intimacy is proximity
invites a respondent to demonstrate how close they
feel to a brand by moving themselves towards or away
from it.
4. The importance of avatars
Whilst relatively few people actually visit Second Life
there is quite widespread familiarity with the idea
behind it: that a participant can represent himself by
means of an avatar and project himself into a scenario.
Early testing of Metaphorix™ indicated that this was a
highly engaging and effective idea, which has particular
relevance to and significance for the measurement of
emotion. After all, qualitative research practitioners
have, for some decades now, been using projective
techniques to help penetrate the unconscious mind by
eliciting feelings below the surface of conscious, rational
response.
Thus, in the first stage of the Metaphorix™ interview,
respondents are asked to choose an avatar to represent
themselves within each animation. The choice of avatars
has been specifically designed to represent gender
and racial diversity so that respondents can choose a
representation of themselves with which they feel both
empathetic and comfortable. The avatars validated for
use in Metaphorix™ are illustrated in figure 2.
International testing indicates that these avatars are
widely applicable across different markets, including
China. Table 2 demonstrates that respondents – interviewed
across four countries and three continents –
were comfortably able to pick an avatar from the set
of choices.
METAPHORIX™ IN ACTION
How to bring visual metaphors to life
Metaphorix™ currently comprises a range of five
metaphoric visualisations, all of which are presented
as interactive online animations that respondents can
use intuitively, without having to think or consider their
response.
figure 2
the me taphorix TM ava tars
© Copyright 2008 Conquest Research & Consultancy Ltd. All rights reserved.
The MetaphorixTM tool and brand name MetaphorixTM are the exclusive property of Conquest Research & Consultancy Ltd.
Copyright © ESOMAR 2008
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
Within each animation, respondents can move their selfrepresenting
avatar to show how they feel about a brand,
a piece of advertising, or any other marketing stimulus.
Each animated visual scenario represents a metaphor
associated with an emotion: for example, proximity/
closeness as a metaphor for intimacy (see figure 3).
The emotion is represented by a visual space in which
the respondent places his avatar in relation to another
person, an object, or value, such as a brand. The
respondent is given a non-leading instruction such as:
‘‘move your avatar to show how you feel about Brand
X’’. A key element of the respondent interface is that the
animations are free-flowing, allowing the avatar to move
seamlessly across the entire animation, without the
respondent being aware of any scaling or calibration.
The Metaphorix™ tool uses visualised metaphors
(of sensory experience) to represent five emotions/
feelings. These are summarised in table 3.
Q12. And which of these phrases best describes how you found picking a character to represent you in the survey?
Base: All respondents Language Gender
Total UK US Italy China Male Female
Unweighted Base 1228 306 306 308 307 612 616
Weighted Base 1228 306 306 308 307 614 614
% % % % % % %
It was fine - I had no problems choosing an avatar
that I wanted to represent me
77 74 76 85 71 80 74
It was ok - none of the avatars represented me
well, but I could choose one
22 24 23 14 28 19 26
Not good - none of the avatars were suitable to
represent me
1 1 1 1 * 1 *
Ta ble 2
The Ava tars Work Interna tional y
figure 3
the proximi ty me taphor in ac tion
© Copyright 2008 Conquest Research & Consultancy Ltd. All rights reserved.
The MetaphorixTM tool and brand name MetaphorixTM are the exclusive property of Conquest Research & Consultancy Ltd.
Copyright © ESOMAR 2008 10
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
For example, the metaphor affection is warmth is represented
visually by temperature (from hot to cold), since
the sensory experience that underpins the metaphor is
warmth. In each case there is no need for the animation
to have numbers or words, but behind the visual animation,
response can be scaled and calibrated quantitatively.
An example of output for the Proximity metaphor is
shown in figure 4, which shows the presentation and
calibration of data (using a 10 point scale), for three
different brands.
Figure 4 indicates that respondents are quite likely to
use the extremes of the scale – if they feel strongly
about the brand(s) in question – uninhibited by having
to give/choose a semantic response. Indeed, one of
the key findings to emerge from the validation exercise
is that metaphoric measures enable consumers to
express themselves more fully, which has the effect
of ‘pulling apart’ responses where standard measures
tend to produce a flat-line.
Ta ble 3
Emotion Visual metaphor used
Affection Temperature
Intimacy Proximity
Excitement Being off the ground
Happiness Facial expression
Empathy Approaching/Coming towards you
figure 4
proximi ty to bran d
Copyright © ESOMAR 2008 11
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
VALIDATION OF METAPHORIX™
The testing programme
Metaphorix™ was launched internationally, in May 2008,
after completion of an extensive (iterative/multi-country)
validation programme. It comprised over 5,000 interviews,
focussed on major international brands, and was
designed to test the following hypotheses:
1. That animated visual metaphoric scales better
capture consumers’ emotional response to brands
and advertising than conventional measures.
2. That this approach can be employed universally to
measure engagement with brands and advertising.
3. That these new metaphorical measures work effectively
alongside conventional verbal scales to enhance
insight into the contribution of emotion to building brand
share.
By exploring response to a wide range of brands and
advertising – some emotional, others more rationally
based – a balanced test design highlights where
conventional research falls short of capturing emotion,
and also shows how visual metaphoric measurement
can fill the gap left by conventional measures.
The two key phases of the programme were as follows:
Test 1 assessed 18 TV commercials, and comprised
3681 respondents, interviewed online in the United
Kingdom during March/April 2008.
The research approach and rationale is illustrated in
figure 5. Three matched cells were needed, for each
brand and advertisement, to establish the benefit of
using visual metaphors over a conventional approach
(to measuring emotion). The test design thus allowed
robust analysis of:
• The impact of exposure to advertising versus a
non-advertised cell (who saw only a brand logo).
• The effect of combining Metaphorix™ measures
with verbal scales versus a matched cell, which
was exposed to the same stimulus, but given only
conventional verbal scales.
figure 5
tes t i: overview of me tho dolog y
Copyright © ESOMAR 2008 12
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
Test II comprised 1,228 respondents interviewed online,
during May 2008, in China, USA, Italy and the United
Kingdom. It focussed on response to major international
brands across four categories within each country,
using both conventional and Metaphorix™ measures.
Each respondent assessed only one brand from each
category with order of exposure rotated using a balanced
incomplete block design. The test was designed to
explore consumer relationships with pairs of high equity
brands – for example, Coke and Pepsi, Nike and Adidas.
The methodology is illustrated in figure 6.
WHAT ARE THE BENEFITS OF METAPHORIX™
OVER CONVENTIONAL MEASURES?
The programme of testing revealed a number of
benefits of incorporating the Metaphorix™ approach
into brand and advertising evaluation. These benefits
are described below:
1. Enabling consumers to express themselves fully
The results of the validation programme indicate that
visualised animations allow respondents to express
their feelings – about brands or advertising – more
freely than do conventional verbal/semantic scales. In
particular, it appears that respondents are more likely
to use the top end of scales and less likely to use the
middle responses.
Figure 7 shows the comparison of results (from Test 1)
of a Metaphorix™ Brand Warmth scale (10 point) with
a comparable verbal scale (where 10 = I feel extremely
warm towards brand / 1 = I feel extremely cold towards
brand). Test 1 comprised matched independent samples
and, in each cell, respondents were asked to express
their feelings towards a brand, after seeing some TV
advertising, one cell using a metaphoric Brand Warmth
scale and the other the comparable verbal scale. Each
cell was matched to the other in terms of demographics
and brand usership, and was shown exactly the same
advertising stimulus.
The results indicate that respondents are significantly
more likely (when evaluating advertising metaphorically)
to use the top end of the scale and significantly less likely
to use the middle (4 - 7) responses.
This pattern of response is also evident internationally:
results from Test II (conducted across four countries)
indicate that the same pattern of response is evident
figure 6
tes t ii : overview of me tho dolog y
Copyright © ESOMAR 2008 13
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
for measures of both Empathy and Proximity (see
figure 8) when respondents are asked to assess major
international brands, such as Subway and McDonald’s,
Apple and Microsoft.
Why is “freedom of expression” important when
measuring emotional reactions to brands or to marketing
communication? The results of the IPA study (outlined
earlier in this paper) highlighted the difficulty of picking
‘winners’ using conventional advertising pre-testing,
particularly for advertising that speaks mainly to the
emotions. Conventional pre-testing typically measures
emotional engagement via scalar responses (on
dimensions such as likeability, enjoyment or involvement),
and this can produce an undifferentiated response,
making it difficult to pick out winners. This effect
was apparent in the test of 18 TV commercials in the
United Kingdom (Test 1) described earlier. Each of the
commercials submitted to this test focussed on nonrational
(emotional) communication.
The results are shown in figure 9, which compares
response to the 18 commercials using a conventional
5-point enjoyment scale with a Metaphorix™ Happiness
measure. Overall levels of response were, on average,
higher using the metaphorical measure, but, most
importantly, there are a number of cases where
response is significantly higher than one might expect.
Of most interest is the ‘winning’ commercial (the award
winning ‘Gorilla’ commercial for Cadbury’s Dairy Milk,
a major UK confectionery brand) which is significantly
ahead of the next best commercial on the metaphorical
measure, whereas on the conventional measure it
scores at a similar level to three other commercials.
Moreover, the percentage gap between this and the
bottom commercial is 42 percentage points on the
figure 7
bran d warm th
conven tional vs . me taphorix TM scale (uk )
Base: 3681 respondents in UK
Copyright © ESOMAR 2008 14
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
conventional measure, and 63 percentage points on
the metaphorical scale.
These results support the hypothesis that the visual/
metaphorical approach allows respondents to express
themselves in a more unfettered, spontaneous manner,
which helps to discriminate between different advertising
executions, particularly those – like Cadbury’s Gorilla –
that appeal directly to the emotions. Response to the
commercial for Brand Z is also particularly interesting
in this respect. Brand Z is a leading UK beer brand,
aimed at young men and using the creative theme of
male bonding in its advertising. Again, the intent of the
advertising is to appeal to the emotions rather than
communicate a rational benefit. Respondents of this age
and gender are notoriously difficult to interview about
their emotions, and it seems likely that the conventional
verbal scales constrained their response in this instance,
whilst the visualised metaphoric scale seemed to
encourage a ‘freer’ and less constrained response.
(See figure 9.)
It can be also be seen in figure 9 that, in a number of
cases, there was no significant difference between
the scores on the metaphoric and conventional scales,
indicating that where difference does arise, it is because
the metaphorical approach is better able to capture
emotional response.
2. More sensitive measure of brand engagement
shifts
The ‘test and control’ methodology of Test I also enables
an assessment of the ability of metaphoric measures
to detect underlying movements in brand engagement.
In the control cells, respondents saw no advertising
and were asked to rate the brands (represented by
their logos) using both Metaphorix™ measures and
figure 8
conven tional vs . me taphorix tm scales - in terna tional com parison
Base: 1227 in UK, USA, China and Italy (approx 300 per country)
Copyright © ESOMAR 2008 15
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
conventional brand measures, such as brand disposition
and purchase propensity. In the parallel test cells,
respondents were shown TV commercials (for the same
brands) and asked to rate the brands using the same
mixture of metaphorical measures and conventional
brand measures. Comparison of movements between
the two cells can thus tell us how well each approach
identifies brand engagement shifts engendered by
exposure to advertising.
The results show (see figure 10) that conventional
persuasion measures produced only two significant
shifts between the test and control cells – for brands
B and M – one moving positively and the other moving
negatively. The metaphorical measures were generally
more sensitive to movements in brand engagement
engendered by exposure to the commercials. Six of the
commercials (33% of the total) produced a significant
positive shift on at least two of the three Metaphorix™
measures (Empathy, Warmth and Proximity) versus the
control cell, with no corresponding movement on the
conventional measures.
3. Making the interview more engaging
In previous sections, we discussed the importance of
engaging respondents with the interview process. Thus,
at the end of each interview, respondents were asked
to say how much they enjoyed taking part in it, and to
give their comments. Comparison of response in cells
where Metaphorix™ measures were used versus those
where only conventional measures were employed,
allows examination of the effect of including such scales
on respondent engagement and survey enjoyment.
Figure 11 shows that, in the United Kingdom, levels of
survey enjoyment were significantly enhanced (over a
parallel conventional study), with approximately 60%
saying “I really enjoyed it”. A similar outcome was
evident in the United States, and in both China and Italy,
levels reached c. 70%. The levels of survey enjoyment
observed across different countries suggests therefore
that the Metaphorix™ approach has wide cross-cultural
acceptance and applicability.
figure 9
adver tising engaemen t: conven tional vs . me taphorix tm measures
Base: 3681 respondents in UK
Copyright © ESOMAR 2008 16
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
figure 10
com parison of bran d engagemen t shif ts, pos t adver tising :
persuasion shif ts vs . me taphorical scales
figure 11
surve y en joymen t: conven tional vs . me taphorix tm
Copyright © ESOMAR 2008 17
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
4. Predicting and diagnosing brand success
Can metaphoric scales predict the success of a brand?
For two years prior to launching Metaphorix™(in May
2008), Conquest Research incorporated experimental
visual metaphoric measures (alongside more conventional
questions) into its online brand health and advertising
studies in the United Kingdom and elsewhere. Over 120
brands have now been assessed across more than
20 categories.
These tracking data allow us to test the hypothesis that
metaphoric scales are predictive of brand success by
observing the relationship between a dependent and
independent variable. The dependent variable is brand
share/first choice brand, and the independent variable is
the proportion that placed the same brand at the top
end (top three deciles) of a visual brand warmth scale.
This relationship (illustrated in figure 12) shows that brand
warmth is a statistically significant predictor of brand
share/first choice. The high correlation (The Pearson
Correlation Coefficient is 0.76 and p>0.01) provides strong
evidence for the effectiveness of metaphoric scales
as both a measure of emotional engagement and as a
predictor of success in the market.
Figure 12 also indicates that successful brands tend to
enjoy high levels of emotional engagement. Generally, the
higher the level of engagement, the higher the market
share, which is perhaps unsurprising, but figure 12 also
indicates that high engagement does not always create
successful brands. Thus there are outlying brands that
have a high level of warmth, but low brand choice; and
those with low warmth and high choice.
figure 12
me taphors can pre dic t bran d suces
Copyright © ESOMAR 2008 18
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
What does this tell us? Where engagement is high relative
to brand share, it may be because recent experience
or knowledge of the brand is too patchy or too weak to
forge a strong link between the brand and what it offers.
Alternatively, it may be that the brand is not marketing
itself strongly enough; some brands have a huge
emotional potential, which is not fully exploited by its
marketers. Engagement is not enough to drive sales –it
has to be supported by a good price, product offer and
widespread availability.
Where there is low brand engagement, but high brand
share, it may be because the brand is offering an
attractive set of product features, price, deals, etc., but
is not creating enough engagement with its customers.
Engagement cannot usually be bought through price,
deals or promotion – it usually requires sustained, brandled,
communications activity.
It is also possible to combine metaphorical measures
of engagement with more conventional ones (such as
brand consideration) to segment consumers in terms
of their rational and emotional commitment to different
brands. Table 4 shows an analysis of two brands
from the UK cell phone network market. Brand A has
converted a high percentage of those who are warm
toward the brand, but has a high percentage of Cool
Considerers, who are not particularly positive. There is
probably a need to forge a stronger emotional connection
with the brand before these consumers would choose
it. For Brand B, on the other hand, the opportunity may
be to build a relationship with the Warm Prospects, who
already view the brand positively, but may not yet have
been given a reason to buy. A campaign focussing on
rational benefits might well resonate with this group.
5. A cross cultural measure of engagement with
brands and advertising?
Encouragingly, it appears that most the benefits of
metaphoric measurement, described above, are available
to researchers worldwide. We have seen that levels of
respondent engagement – across markets as diverse
as Italy and China – are consistently high when visual
metaphoric scales are used. The hypothesis that visual
metaphorical measures work across cultures is further
supported by very positive open-ended comments,
which suggest that respondents in different cultures
enjoy, understand and engage with the interview
process. Moreover, the high correlations observed
internationally (see figure 13) between Metaphorix™
measures (of Proximity, Warmth and Empathy) and
conventional brand choice/consideration scales suggest
that visual metaphorical measures are predictive of
brand success in a range of international markets.
Ta ble 4
Bran d Warm th vs . Rational Consi dera tion : Drive Matrix
Brand Engagement
BRAND A BRAND B
DRIVE MATRIX COLD LUKEWARM WARM COLD LUKEWARM WARM
1st Choice 0% 4% Committed
10%
0% 0% Committed
4%
Would Consider 1% Cool
considerers
17%
Warm
Prospects
8%
1% Cool
considerers
8%
Warm
Prospects
12 %
Not in
Consideration
Set
22% 33% 6% 32% 42% 2%
Consideration
Copyright © ESOMAR 2008 19
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
Clearly a lot more work remains to be done, but the
results of the Metaphorix™ project thus far provide
strong initial evidence both for the effectiveness of visual
metaphor as a means of measuring emotion, and for
employing online metaphoric scales universally, across
cultures.
CONCLUSION
Let’s get animated about emotion
It is high time we stopped talking about measuring
emotion and started doing something about it. The
Metaphorix™ approach – based on the applying the
cognitive linguistic theory of metaphor to the world of
Web 2.0 – suggests that animated visual metaphors can
work where most conventional approaches fail. Given
the burgeoning understanding of the emotional unconscious,
it is astonishing that, even now, in 2008, most
conventional market research is text-based and literal,
not visual or metaphorical. Most conventional research
actually divorces respondents from their emotions – the
emotion may still be present, but it becomes more prone
to rationalisation as they consider their response.
Whilst neurological/biological measures seem to
provide an exciting alternative, even their strongest
proponents accept that they are at best a complement to
conventional approaches, not a replacement for them.
There is strong evidence both for the effectiveness of
visual metaphor as a means of measuring emotion,
and for employing online metaphoric scales universally,
across cultures. These measures seem to go beyond
purely rational response, freeing respondents to express
feelings and emotions which conventional (verbal)
approaches miss or suppress. They can thus better
capture emotional response and, when employed
alongside rational measures, provide a commercial
insight into brand health denied to conventional
approaches alone.
Marketing is all about imagination and metaphor, yet
we so easily default to prosaic and literal means of
measuring it. By sticking with the conventional paradigm,
online research has thus far failed to exploit the most
exciting opportunity that Web 2.0 offers – to turn
respondents into participants by fully engaging them with
the interview process. It is time to seize the opportunities
that the medium can offer and get truly animated
about emotion!
figure 13
in terna tional me taphorix tm corela tions wi th bran d choice /consi dera tion
Copyright © ESOMAR 2008 20
CONGRESS 2008
par t 12 / FRONTIERS IN RESEARCH: UNSPOKEN
References
Binet, Les and Field, Peter (2007). Marketing in the Era of
Accountability. WARC, Henley on Thames, Oxon.
Damasio, Antonio (1994). Descartes’ Error. GP Putnam’s Sons, New
York
Damasio, Antonio (1999). The Feeling of What Happens. A Harvest
Book, Harcourt Inc, San Diego, CA,
Franzen, Giep and Bauwman, Margot (2001). The Mental World of
Brands. WARC, Henley on Thames, Oxon.
Grady, J. (1997). Foundations of Meaning: Primary Metaphors
and Primary Scenes. Ph.D. dissertation, University of California,
Berkeley.
Johnson, Christopher (1997). Metaphor vs. Conflation in the
Acquisition of Polysemy. In M.K. Hiraga, C. Sonha, and S. Wilcox,
eds. Current Issues in Linguistic Theory 152. John Benjamins,
Amsterdam.
Johnson, Mark (1987). The Body in the Mind: The Bodily Basis of
Meaning, Imagination and Reason. University of Chicago Press,
Chicago.
Johnson, Christopher (1997). Learnability in the Acquisition of
Multiple Senses: SOURCE Reconsidered. In J. Moxley, J. Juge, and
M. Juge, eds., Proceedings of the Twenty-Second annual Meeting
of The Berkeley Linguistic Society. Berkeley Linguistics Society,
Berkeley.
Kövesces, Zoltán (2000). Metaphor and Emotion: Language, Culture
and Body in Human Feeling. Cambridge University Press.
Kövesces, Zoltán (2002). Metaphor in Culture (2005). Cambridge
University Press.
Lanham, Richard. A (2006) The Economics of Attention: Style and
Substance in the Age of Information. University of Chicago Press.
©2006
Lakoff, George and Mark Johnson (1999). Philosophy In The Flesh.
Basic Books, New York, NY.
Le Doux, Joseph (2002). Synaptic Self. Penguin Books, London.
Le Doux, Joseph (1996). The Emotional Brain. Touchstone, Simon
and Schuster, New York.
Mast, Fred and Zaltman, Gerald (2006). Anatomy of Engagement
in Definitions about the Anatomy of Engagement, The 52nd ARF
Convention, 2006.
Mc Clure, Samuel M; Li, Jian; Tomlin, Damon; Cypert, Kim S;
Montague, Latane M and Montague, P.Read. (2004). Neural
Correlates of Behavioural Preference for Culturally Familiar Drinks.
Neuron, Vol. 44, 379-387. Cell Press.
Narayanan, Srini (1997). Embodiment in Language Understanding:
Sensory Motor Representations for Metaphoric Reasoning About
event Descriptions. Ph.D. Dissertation, Dept of Computer Science,
University of California, Berkeley.
Pinker, Steven (2008). The Stuff of Thought. Allen Lane, London,
England.
Singerland, Edward. G. (2005). Conceptual Blending, Somatic
Marking and Normativity. Cognitive Linguistics 16-3. Pages 557-564
Zaltman, Gerald (2003). How Customers Think. Harvard Business
Press, Boston.
Zaltman, Gerald (20008). Marketing Metaphoria. Harvard University
Press
The Author
David Penn is Co-founder and Managing Director, Conquest
Research, United Kingdom.
No comments:
Post a Comment