Catching brand butterflies with reactiontimes A new method to measure brand appeal
BR-ND August 2010
Celeste Miller Kim Cramer Alexander Koene
6
Catching brand butterflies with reaction times A brand new method to measure Brand appeal Brand appeal is the attraction one feels for a brand. It is comparable to the butterflies dancing around in your stomach when in love! One of the most fantastic, crazy, but reassuring feelings - so very present, but so extremely difficult to put in to words. It is obviously simply impossible for someone to provide the precise amount of butterflies there were flying around in your tummy when last in love, and the same thing applies for Brand appeal. This since our introspective capacity is limited and our best attempts are influenced by our (unconscious) desire to conform to social norms. So how do we measure Brand appeal, if people cannot explain it well enough themselves? BR-ND’s Brand Lab has spent the last 2 months developing a state-of-the-art method to measure Brand appeal in a new way.
Brand appeal Brand Appeal stands for the positive feelings we have towards a brand. Brands with high brand appeal cause an unconscious ‘underbelly’ feeling which excites us and provides a fantastic feeling. This fantastic feeling works as a heuristic, which in busy and time-limited situations make our brand choices easier and better 1. Brand appeal is the result of three factors – brand awareness (BAW), brand expectations (BEX) and its 23plusone identity (24 unconscious motives and aspirations) – see figure 1.. Brand awareness stands for the degree to which the consumer spontaneously knows the brand. Brand expectations is the degree in which a person knows what to expect from a brand and the 23plusone identity is based on the 24 fundamental human drives
Figure 1.
1
Article ‘A relationship between brand appeal and the similarity of 23plusone profiles between brand and consumer‘ (BR-ND, August 2010), shows that Brands which have higher brand appeal, have a 23plusone identity which is very similar to the consumer’s. Thereby being more relevant and better to the consumer. The brand becomes a lengthening-piece of the consumer.
6
Old (explicit) versus new (implicit) measuring method Currently the brand appeal score is made up from the average score of two extensively validated questions (picture 1.). This approach has been validated extensively and it seems to work quiet well. It is also very quick and easy, but lacks construct validity as it was an explicit method attempting to measure something implicit (unconscious). As the unconscious aspect of Brand appeal is exceptionally fundamental, we made the important decision to diverge from our formal explicit measuring method.
Picture 1.
In the formal explicit fashion scores are influenced by biases as the lack of introspective capability, social desirability and self-perception. Introspective capability refers to the fact that subjects are often unaware of their true attitudes. Social desirability describes the tendency of respondents to adjust to the social norms made by our peers/family/loved ones. This will generally take the form of overreporting good behavior or underreporting bad behavior (e.g. indicating that you don’t love Pepsi, because this is in agreement with your friend’s norm, whilst you actually love it and hate Coca-Cola). Self-perception is people’s tendency to infer their attitudes from observing their own behavior and circumstances in which it occurs. In regard to brand attitudes tests, people are hesitant to have more positive attitude about brands which they own. Research has discovered that these three factors influence explicit methods, but not implicit methods. Thus in addition to explicit methods, there are also implicit methods. Implicit relays to the fact that subjects are not aware of what the test measures, thereby reducing biases, such as social desirability which influence scores on explicit methods. Implicit measures focus on participant’s attention on performing some task that can indirectly reveal the underlying construct such that inaccessible and closely held attitudes can be measured. Implicit measures can be defined as ‘introspectively unidentified’ (or inaccurately identified) traces of past experiences that mediate favorable or unfavorable feelings, though or action towards social objects. More specifically – a 6
consumer’s past purchases do not influences scores, seeing that the consumer does not know what is being measured. Since Brand appeal stands for an implicit attitude towards brands, sensitive to answer biases, it was very important that the method to measure Brand appeal transformed into an implicit one.
The creation & validation of a measuring instrument There a several implicit measuring methods which can be implemented to measure Brand Appeal. One is sequential priming. In this task, participants classify positive or negative words into categories and the effect of the preceding (prime) stimulus on classification speed is observed. A typical priming effect is that the participant’s speed of classifying the words is greater when the prime is associated with the words than when it is not. Thus if someone has a positive attitude towards BMW and a negative attitude towards Skoda, priming with the BMWlogo will facilitate the categorization of the positive words, but will impede the categorization of the negative words. On the other hand; priming with Skoda will facilitate the categorization of the negative words. Priming of the brand activates the associated attitude – negative vs. positive, which will facilitate categorization of the respective words. Priming researches however have established limitations for the measure. In particular; effects can be detected even with weak primes, and the recency of prior access to the attitude toward the prime might affect priming results by increasing the activation level of the association. Also, priming measures also have relatively low internal consistency (between brands) and test-retest reliability (time interference) – thereby being insensitive to individual differences. As result to the limitations of the sequential priming task, BR-ND chose a variant to the IAT (Implicit Association Task 2), the SC-IAT (Single Category-Implicit Association Task) to measure Brand appeal. The SC-IAT works by measuring the valence and the strength of the association between the mental representations (association network) of the category and the self (‘ME’). Ever since our first encounter with a category our memory has been creating huge networks of associations, based on all of our experiences with the category. The strengths of these associations can vary – if someone has had more positive encounters with BMW than with Skoda, he or she will have a stronger positive association to BMW. Stronger associations are represented by thicker ‘positive’ links between the categories (nodes) ‘BMW’ and ‘Me’ in the brain (picture 2.)
BMW +B
ME
M W
B B M Picture 2. M W W 2 Since the introduction in 1998, the IAT has been the most popular to measure the unconsciousness. It has been designed by researchers at Harvard University, University of Virginia and the University of Washington. The IAT is designed to measure unconscious attitudes in regard to two categories (e.g. young vs. old; fat vs. thing; republicans vs. democrats) in contrary to the SC-IAT which measures the attitude to one category. .
+ + M E
6
+
The power of the SC-IAT lies in the fact that it can measure the strength of the association brand – consumer and furthermore does not limit itself to the conscious associations. Thereby it can provide a complete picture of the degree of (unconscious) preference to socially weighted concepts as brand preference. Measuring the strength and the valence of the association is done by registering the time it takes a participant to categorize words that appear one-by-one in the middle of a computer screen. These words can be related to one of three categories – love (‘liefde’ in Dutch), hate (‘haat’ in Dutch) (mentioned respectively left and right at the top of the screen) or a brand (mentioned just below positive or negative). The test has two trials. During one the participant has to press the same key (‘E’) when the brand or love related words appear and another key (‘I’) when hate related words appear. This difference in categorization is also made clear on the screen – the brand is mentioned below the category ‘love’ (picture 3a.). During the other trial the participant has to press the same key when the brand or hate related words appear (‘I’) and another key (‘E’) when love related words appear – on the screen this categorization is shown by means of the brand being mentioned just below the category ‘hate’ (picture 3b.). If the participant categorizes the words faster when the brand is mentioned under the category ‘love’ and therefore has to press the same key when love related words or the brand appear, then when the brand is mentioned under the category ‘hate’, he or she has a positive (implicit/unconscious) attitude towards the brand. This difference in categorization speed is related to the valence of the association, causing the participant to see ‘love’ and the brand as an unit (it went against his or hers nature to categorize hate related words and the brand as an unit). The strength of this positive attitude is scored by the average time it takes to categorize the brand and love related words.
E
I
Picture 3a.
Picture 3b.
6
The (SC-) IAT has several key facets ensuring the task to indirectly reveal (implicit) brand appeal. One is highlighting the fact that the test is a simple categorization task (participants get to see the appearing words before the test as well, so that recognize them and only have to categorize them accordingly). Additionally a red cross appears in the middle of the screen when a categorization error is made, and the subsequent word only appears as soon as the correct button (‘E’ or ‘I’) has been pressed - ensuring the idea of a categorization task and minimizes the idea of an evaluation task. Another important facet is that participants have to work as intuitive as possible, ensuring minimal to nihil rationality. Speed is also important – it is essential that the participants react as fast as possible, whilst staying as accurate as possible. Categorizations longer than 2000ms cause a warning to appear in the middle of the screen requesting the participant to work faster. A few errors due to speed however, is not a problem and this is mentioned as well. Furthermore it is stressed that the participant has to keep his or her full attention to the task.
Pilot test Before introducing the method we held a pilot tests. Eight car brands were used; BMW, Audi, Alfa Romeo, Mini, Volkswagen, Fiat500, Fiat and Skoda (n = 55). This list includes four premium brands – BMW, Audi, Alfa Romeo and Mini, two mid-range brands – Volkswagen and Fiat500 and two budget brands – Fiat and Skoda. Although Fiat500 is model of Fiat we decided to add it, because its image differs a lot from Fiat’s image and it has practically become a brand on its own. Participants were recruited via social media (Facebook, Linkedin, Blogs and Twitter) and could participate in a Dutch test online. The software used was Inquisit by Millisecond. To maximalize the ammount of participants per car brand – participants were requested to participate in atleast two links (of which each link excited of two randomized brands), to maximalize the ammount of participants per car brand. Before the actual tests, the participants saw a clear briefing about the task after which they could get familiar with the style of the test during two practice trials. The practice trials as well as the test trials were divided into the two (randomized) trials. After the test the participant was thanked and briefed with their Brand appeal scores of the two car brands and the meaning. The inquisit software used the newest method to calculate the score, known as the ‘new scoring algorithm’. In general this renewed method takes too fast (< 300 ms) and too slow (> 3000 ms) responses into account, aswell as differences in cognitive fluency (individual differences in reaction time) due to age and experience.
6
Hypothesis & results Since explicit and implicit methods measure two different aspects, we hypothesized that the ranking of the implicit car brand appeal scores measured by the SC-IAT would be similar but have small differences to the ranking explicit car brand appeal scores, measured in the 23plusone study. Therefore participants might receive brand appeal scores from this SC-IAT which vary (slightly) to their expectations. Since the SC-IAT is a well-validated method, different scores do not indicate that something is wrong with the construction of the test. Different scores simple indicate that new method measures something different than conscious opinions â&#x20AC;&#x201C; exactly what we are aiming for. Below the 23plusone ranking and the new, SC-IAT ranking. 23plusone (explicit) ranking of car brands. (Fiat500, Fiat & Mini was not tested) 1. 2. 3. 4. 5.
Audi BMW Volkswagen Alfa Romeo Skoda
SC-IAT (implicit) ranking of car brands. 1. Alfa Romeo 2. Mini 3. Fiat500 4. Audi 5. Volkswagen 6. BMW 7. Fiat 8. Skoda. Looking at the car brands which appear in both rankings, there are some differences and some similarities. One obvious one is that Alfa Romeo is the brand with the most brand appeal in the new test, whilst Audi was that in the (explicit) former test. Volkswagen stays in the middle in both test, in the neighborhood of BMW. Interesting is that Skoda is mentioned as the brand with the least amount of brand appeal in both tests.
6
Implications This SC-IAT method is a very advanced and innovative method to measure unconscious attitudes. Previous research at universities used the technique to measure attitudes towards divergent subjects from presidential candidates to brands. Research into brand attitude also showed that scores have a high predictive validity of brand preference and consumer behavior. The applicability and high validity of the method caught BR-ND’s attention. The pilot test for eight car brands indicates that we judged its potential correctly. Feedback of the participants provides some interesting insights. First is that the participants experienced the test as more appealing, attractive and motivating than more regular ‘questionnaires’. They appreciated the different approach, the short duration (5 minutes per brand) and regarded their scores as educative and fascinating. The comment below is a fine representation of the positive feedback we received. Celeste, I must say - the test in developed ingeniously. Beforehand I thought that I would have the highest score for Fiat (in comparison to Mini), because I have a weakness for their refined Italian design en not the least for their technology. On top of this I have zero connection to English cars! What drives me to bias the Mini then? I have a small suspicion, and to not bore you with a long-winded story – here a sneak preview. Very long ago (before my wifes time) I had a romantic adventure with a fine, dark-skinned lady in Portugal. We drove in a original Mini, that master product of designer Alec Issigonis. In this week I didn’t only get more admiration for the Mini, but naturally also for the dark lady who looked like a copy of the Egyption Nefertiti! I think it is clever, en actually completely insane, that such a sentiment of 35 years ago, made sure that the test showed a different outcome than expected. I am more than convinced! Thank you very much for reminding me on the impact this lady (and the Mini) had on me!
To conclude, the first baby steps in launching this new method to measure brand appeal have been very successful. Naturally BR-ND is looking forward to integrate this state of the art method in upcoming in- & outsiders. Please keep an eye out for our next publication, which provides another mind blowing insight – ‘Is there a match between what people find particularly important in life and the drives triggered by brands they find appealing’. Please think about this question in the mean time and find out if your thoughts match our results.
6