The role of emotions
How do consumers feel about a brand? And how can we assess emotions towards a particular brand? The role of emotions in consumer behaviour has received great interest lately. Both practitioners and researchers agree upon the fact of emotions being very relevant and evident in human consumption. Several studies confirm this by showing that emotions, affect, feelings, and moods play a critical role in determining consumers’ behaviours (Bagozzi, Gopinath, Nyer, 1998; Huang, 2001; Thomson et al., 2005; Carlson et al., 2007).
Even though emotions profoundly influence consumer decisions there is no tool to date that has been consistently used across companies to draw up an emotional profile of a brand.
What is even more surprising, only 47% of marketing leaders are aware of what their brand feels like for consumers (YouGov Deutschland AG & Supersieben GmbH & Co.KG, 2015). The underlying cause seems to be a lack of standardised procedures to be able to determine an emotional profile of a brand. In fact only 24% reported having applied a tool specifically designed to measure emotive characteristics of a brand.
Finding a relevant scale for assessing emotions
Within the psychology and consumer behaviour literature, as well as in the marketing literature, there is a considerable number of scales for measuring emotions (Hansen & Lundsteen, 2006). However, each tends to have problems in application to advertising research. For example, one of the most influential emotional theories (Plutchik’s circular model of emotion, 1980) shows that one can combine and mix basic emotions to create new ones. That is, love can be understood as a combination of joy and trust. Nevertheless, Plutchik’s classification has been frequently criticized for its limited number of underlying dimensions of emotions.
As the existing approaches do not provide much insight into the full range of emotions a marketing representative may want to study, Aaker and colleagues (1988) generated a list that represents the full spectrum of emotions likely to be generated by advertisements / brands. In order to do this Aaker et al. first compiled a list relying on validated theoretical frameworks of emotions, including Plutchik’s (1980) 40 emotions, Smith and Elssworth’s (1985) 15 emotions, and Davitz’s (1970) 80 emotions. Additionally, lists from audience perception studies were adopted such as Aaker and Bruzzone’s (1981) list of 20 adjectives, Schlinger’s (1979) 32 response statements, and 600 adjectives from Wells, Leavitt, and McConville (1971). Emotions most relevant to the marketing / consumer literature were derived from this long list and divided into sets of positive and negative feelings. Within each set (positive | negative) similar emotions were grouped together. Each feeling cluster is headed by a label that attempts to capture the essence of the cluster. For example, “fascinated”, “impressed”, and “attentive” forms the feeling cluster “interested”. In the end, Aaker et al. came up with a list of 180 items, clustered into 31 feelings (16 positive & 15 negative) (for a more detailed explanation of the procedure refer to Aaker et al., 1988).
One hundred and eighty items may cover a thorough range of emotions relevant to the marketing context. However, this list is not suited for practical use. This is why we, firstly shortened Aaker and colleague’s list from 180 to 50 emotions to overcome response rate failure (Galesic & Bosnjak, 2009), Secondly, we use a list of emotions and values since these are interlinked. “Emotions are judgments of value” as stated by Todd (2014). Yet values are fact based and hence add a new quality to the panel. Values can be incorporated depending on the brand we are looking at (i.e. security for airlines). Please note that emotions and values may be used interchangeably from now on. Lastly, it is important to mention that whereas the original study contained equal amounts of positive and negative emotions we have focused on the positive with only a few negative emotions. This is due to the fact that a brand positioning will always be formulated in a positive way.
Having developed a list of 50 emotions, we thought about a way of how the contained emotions should best be measured in order to learn about consumers’ emotional attachment towards a particular brand. Rossiter and Bellman (2005) defined two types of emotions and subsequently described how they should be measured. This measurement has important implications for our tool and its application.
Measuring Emotions – Type 1 vs. Type 2 Emotions
Rossiter and Bellman (2005) differentiate between Type 1 Emotions (E1) and Type 2 Emotions (E2). Type 1 Emotions are instinctive reactions and comprised of pleasure, arousal, and dominance (P-A-D) (Russel & Feldman Barrett, 1999). During a typical day, an individual will experience a range of Type 1 Emotions, which may at times trigger physiological reactions. Thus, they can be measured by psychophysical recording such as galvanometric skin response (GSR). Type 2 Emotions, on the other hand, are extreme states of Type 1 Emotions that individuals have learned to associate with different cognitive appraisals. These cognitive appraisals answer the question: “What emotion am I feeling?” The answer may be for instance: “joy”, “fear”, or “love”. Logically, Type 2 Emotions cannot be measured by psychophysiological recording. Instead, researchers have to rely on self-report ratings. Rossiter and Bellman (2007) further argue that Type 2 Emotions, as opposed to Type 1 Emotions, are not continuous but rather are discrete (all-or-none; “yes” or “no”, respectively) and hence must be measured binary. Capturing consumers’ Type 2 Emotional attachment to brands with a binary rating is highly accurate as it reflects an individual’s own interpretation of their emotional state, rather than an average of aggregated responses (Bellman, 2007).
To sum up, the best measure of Type 2 Emotions is a simple binary question: “Are you feeling this emotion? – Yes or No”, as an individual’s labelling is precious when predicting that person’s behaviour. Thus, in order to detect whether a certain emotion is linked to a brand we follow the principle of how to measure Type 2 emotions.
Degrees of emotions
Bellman (2007) puts forth that there are degrees of Type 2 Emotions. Have you ever wondered what differentiates “angry” from “really angry”? That is, anger may conflict with another emotion, such as love. One might feel really angry, because a beloved person was acting inappropriately.
This implies that the addition of two emotions results in a completely new emotion. When determining what kind of love a participant feels towards a brand it is important to look at the emotion closest to love in his or her response: is it more of a compassionate love or rather an open-hearted friendship?
It is the combination of emotions that creates new dimensions – by taking this into account the list of 50 emotions & values adds up to 1032 possible combinations of emotions.
This is a huge amount of data points needing to be analysed. To explain how this is done practically, we first describe how a typical study design looks like, highlighting its advantages. Then we show how the obtained data points are being analysed and finally how they can be interpreted.
Study design / Methodology
1.000 participants are recruited to participate in a web-based self-report study. YouGov, specialised in market research worldwide, helps gathering study participants. Participants answer a simple binary question (“are you feeling this emotion when thinking of e.g. xx? Yes or no”), being shown 50 emotions and values one at a time. From this we conceive around 50.000 data points. These data points are then translated by an algorithm into 3 to 6 emotional territories.
This study design bears several advantages. First, due to a large sample results are applicable to the general population (Cochran, 1977). Second, the questionnaire is short enough to overcome response rate failure (Galesic & Bosnjak, 2009). Third, capturing consumers’ attachment to brands with a binary rating is highly accurate as it reflects an individual’s own interpretation of their emotional state, rather than an average of aggregated responses (Rossiter & Bellman, 2007).
Analysing data points – The Algorithm
The obtained data points are being analysed by an algorithm developed by Dr. T. Meyhöfer (2015).
To understand how the algorithm works it is helpful to think about a situation which most of us have probably encountered in everyday life: if you buy a product from Amazon you subsequently receive an e-mail which states: “Customers who bought the “Philips Hue LED Lamp Starter Set” also bought these products” – Philips Hue LightStrips, Reality Light, Wireless Dimming Kit, Bulb holder etc.
A similar method is applied by our algorithm: i.e. those participants linking love with a particular brand may also connect joy, friendship, curiosity, and creativity with it. Emotions that are intertwined result in a so-called emotional territory.
How to interpret an emotional territory
Having first applied quantitative measures we now apply qualitative evaluations to interpret an emotional territory unveiled by the algorithm. Using different research methods can provide new information that quantitative cannot address (Gaur, Herjanto, Makkar, 2014). Basically, each emotional territory is headed by a label that attempts to capture the essence of the territory. Emotions are always grouped in pairs which in turn are part of an overarching cluster; this is continued until all emotions within that territory have been covered.
Deep and Broad Exploration
The territories can be explored in different ways – either by looking at the range of territories overall (broad) or exploring the different angles within a territory (deep) (Woods, 2001).
One example: Many whisky brands occupy the territory of maturity. Some associate maturity with particular qualities such as discernment, wisdom or friendship; extra social and sexual confidence; passing of experience and wisdom from father to son.
When applying the emotional territory tool there is no guessing about which maturity is most fitting for the brand – it is clear from the outcome.
The difference between success and failure
As Woods (2001) has stated “brand marketing is in the business of selling emotional connections rather than product benefits”. In defining these connections lies a huge potential for brands: to differentiate from others, to change or stretch an existing positioning. Finding a particular nuance within an emotional territory can be a powerful and lasting bond to the consumer. This is why the qualitative element is key here – the type of emotion makes all the difference.
We believe that a company does not have to be born with the emotional DNA of Apple or McDonalds to succeed. Even a vacuum cleaner or moist food for cats can form powerful connections. You simply have to find out about your company’s’ / brand’s DNA to enhance the nuance that is important to your target group and that no competitor can claim.