Seducing consumers, evaluating emotions Claire Dormann Vrije University, IMSE De Boelelaan 1081 1081 NC Amsterdam, The Netherlands
[email protected] Telephone: (31) 20 444 7823 Fax : (31) 444 7728
SUMMARY
This article is part of an investigation in the role and value of emotion in electronic commerce. As marketing studies have shown, emotion plays a critical role in shopping. Although emotional research is emerging in human-computer interaction, practical measures of affective responses are few. SAM (Self-Assessment Manikin) originally developed by Lang, is proposed as such a tool. This study examines the suitability of using SAM for the emotional evaluation of web pages. Moreover by carrying out such an evaluation it is hoped to gain some insights in emotional design. Results suggest that SAM can be used to find the emotional value of home pages. With SAM, designers can compare solutions to find the most appropriate to a given situation. SAM could also serve to assess the suitability of designs for different cultures. KEYWORDS : Electronic commerce, Emotion, design,
Self-Assessment Manikin, evaluation. INTRODUCTION
This research is part of an empirical study into the role of emotion in electronic (e) commerce. Our objectives are to establish guidelines for emotional design, and to develop assessment tools integrating usability and emotional evaluation. To do so, we advocate a tighter integration between marketing and HCI. In this paper, we will present a study conducted for the emotional evaluation of home pages using SAM (Self-Assessment Manikin), and successfully used in marketing. With the SelfAssessment Manikin (SAM), developed by Lang [1] emotion is measured by three scales: pleasuredispleasure, degree of arousal and dominancesubmissiveness. We will first put the present study into context by discussing findings relating emotion to shopping. In order to find out if sites and web pages are emotionally successful, we need tools to assess emotions. An evaluation carried out with SAM will be presented. This evaluation
should enable us to: assess the suitability of the tool in this context, to contrast the emotional value of home pages, and to gain a deeper knowledge of emotional design. SAM has been successfully used in the context of home page evaluation. As could be expected, strong imagery and colour influence the emotional reaction of participants. To conclude, future research questions are outlined. EMOTION AND SHOPPING
For 20 years, research in marketing has followed “information-processing theory”, viewing the consumer as a decision maker who searches for, attends to, perceives and evaluates data to make brand choice[2]. Then evidence emerged suggesting that choices can be based on affective-emotional factors. The traditional utilitarian view in particular seems inappropriate for products whose selections and uses are based upon satisfying emotional want rather than fulfilling utilitarian functions. Emotional desire can dominate utilitarian motives in the choice of products. Besides, according to the Elaboration Likelihood Model, for low-information products (e.g. sweets are low-information while cars are highinformation), processing of the ad content is performed using peripheral cues, such as message source and multisensory cues [3]. This implies that especially for some types of products and services, decisions are not based on searching and finding information but on desire, fantasy and reactions to an emotional climax. Climax, for example provoked by multi-sensory cues and emotional information. It is believed that emotion plays a key role in e-commerce. In some purchasing situations, web pages need to be emotionally successful to fulfil their role. EMOTION EVALUATION
The objective of the present study is, to propose and examine the validity and suitability of using SAM for assessing the emotional impact of home pages. With re-
spect to other emotional questionnaires, SAM offers two advantages, a fast and easy response mode and its culture free, visual aspect. The later could prove an asset for the web. To gather more information on emotion and design, a brief open-ended questionnaire was added to a small sample of home pages. We will briefly discuss previous evaluation studies, then introduce SAM and present our evaluation. Affective evaluation
Few affective evaluations have been carried out in ecommerce. Kim [4] while interested in measuring trustworthiness of cyberbanking systems reminds us of three types of emotional scale: the differential scale, the semantic differential scale and free labelling. However, a large numbers of scales have been developed mostly in marketing studies, for the measure of affective responses. These studies addressing different aspects of emotion, such as mood [5], affective cognition [6], or appeal [7] vary considerably in the scope and range of emotion assessed. An alternative to this sometimes cumbersome verbal self-report is SAM. Self-Assessment Manikin
The Self-Assessment Manikin (SAM), visually represents three dimensions of emotion by using three axes: pleasure-displeasure, degree of arousal and dominancesubmissiveness. SAM depicts each dimension with a graphic character arrayed along a continuous nine point scale. For pleasure, SAM ranges from a smiling happy figure, to an unhappy and frowning figure. For arousal SAM ranges from to an excited figure to a sleepy figure (see Figure 1). Last, the dominance scale goes form a very small figure to a very big forceful half figurehead.
All home pages belong to the low-information class or are for hedonic products, products for which the information component is as important as the emotional component. For example, evaluated pages are for perfumes, chocolates or travel. In order to compare them, pages were arranged by groups of two: sites selling the same product, or two versions of the same site. In this study stronger and more neutral pages were alternated. Some pages were chosen for their strong and dense visual content (including imagery and symbolic associations, colour, or characters). These pages only use text for titles, slogans or navigation. Quieter or more neutral pages completed the selection. These pages use more neutral colour, text description, and smaller graphics using most often standard product presentation. RESULTS AND DISCUSSION
In this paragraph, preliminary results from the evaluation are presented. Using SAM, each page has been characterised by distinct emotional values. There are marked differences between pages. Interesting differences were found between pairs of sites (quieter versus stronger) related to the same products or between pairs related to the same sites. Here we will focus more specifically on pages belonging to the same trademark. Furthermore some emotional design factors are also discussed. Using SAM
Relevant differences were found between different versions of Kilroy (two design variants) and for La Redoute between two different countries. Results are given in details for these two cases. The table includes means and standard deviations, in bracket. Graphs show the mean obtained for the emotional dimensions: pleasure, arousal, and dominance. The scales were rated as followed: 1 for the smiling figure to 9 for the unhappy figure on the pleasure scale, 1 for the excited figure to 9 for the sleepy figure on the arousal scale, and 1 for the small figure to 9 big figurehead on the dominance scale.
Figure 1: SAM the arousal scale Pleasure The present evaluation
The evaluation which was carried out with a class of 67 students following a course on Text and Image Representation, part of a computer science curriculum. Thus, the majority of the students were male. Home pages were chosen for the evaluation as they play a critical role in ecommerce [4]. Twelve home pages were presented, after each one students had to rate the page using SAM. Due to the course set up, it was not been possible to vary the order of presentation. Additionally, after the first four pages, students had to fill a short open-ended questionnaire to gather their impressions and feelings, or to state ideas evoked by these pages.
arousal dominant
Kilroy1 Before33
Kilroy2 Next
Redoute1 British
Redoute2 Belgian
4.26 (2.06) 4.85 (2.11) 5.95 (1.87)
6.02 (1.66) 6.59 (1.83) 5.64 (2.41)
2.68 (1.79) 3.19 (2.11) 5.25 (2.32)
6.02 (1.79) 7.59 (1.39) 4.95 (2.63)
Table 1: SAM, mean and standard deviation, in bracket.
Consumers can buy clothes on-line from La Redoute. Customers access the site through their national home pages. The designs of most home pages are different. The British home page picturing a very sexy model in
underwear, as could be predicted, was judged strongly pleasant with a high arousal level. The Belgium home page contains a smaller picture of a woman dressed elegantly in brown colour. This page obtains a different rating. By contrast to the British page, this version is liked less with bigger scores on the arousal and pleasure scales, thus judged less pleasing and less exciting (see Figure 2). No difference was found for dominance between these two pages.
8 7 6 5 4 3
R-british R-belgian
2 1 0 pleasure
arousal
dominance
Figure 3: Kilroy home pages, Before 33 and next version.
It has to be noted that not all pages with a high score on the pleasure scale have a high arousal level. Lancôme is such a case. Lancôme is a very calm page with a white rose on a white background accompanied by the slogan “believe in beauty”. The only other item on the page is a small menu. SAM could prove an important assessment tool for the emotional evaluation of web pages. Especially for shopping, it might be crucial to test pages for which the emotional user’s reaction is particularly important. Such pages are home pages that should persuade consumers to visit or promotional pages that should generate impulse purchase. Furthermore, during prototyping stage, SAM could be used to compare alternative versions and assess the appropriateness of affective responses. Moreover, with SAM, as indicated by Morris [8], we could study cultural and national differences, or evaluate the suitability of designs for each culture. Design and emotion
Figure 2: La Redoute, British and Belgian home pages
Kilroy, a student travel agency, has a black and yellow colour scheme. Beside the company name and logo the page consists mainly of, a slogan “Go before its too late, for young people under 26 and students under 33” illustrated by a picture (e.g. K-before33). The picture is a humorous representation of a male character going from active adulthood to a decrepit character. In the next version (e.g. K-next), the colour scheme remains, but now the page has a more traditional format, with a search form, options, information, and a list of destinations. The emotional rating of these two pages is strikingly different, the first one being more positive (see Figure 3). The first page rates lower on the pleasure scale but more importantly perhaps, on the arousal scale. Again no differences was found for dominance. For choosing travels and vacations, consumers might have to feel excited and anticipate future pleasure. Consequently, designers should make changes very carefully and test the emotional value of new designs. 7 6 5 4
K-before33
3
K-next
2 1 0 pleasure
arousal
dominance
Answers from questionnaires provided insights into the design of emotional experience. These answers gave some indications of which elements influence users’ reactions. Colour has often been associated with subjective reactions [9]. Thus in web design, colour might have a strong emotional value. It seems that colour, especially the background colour has indeed a high emotional value. Participants often commented on the background colour and its suitability for the products or services proposed. For Kilroy, the dark background colour was in many instances judged too strong, too heavy (as an index for death), and thus not appropriate for a travel site. The yellow was liked and deemed attractive. For Chokolic, a site selling chocolates, the brown background was mostly interpreted positively in relation with chocolates, but also in smaller measure negatively with brown as a colour associated with old people. Although Lancôme rates high on the pleasure scale, generating feelings of quietness, white was sometimes associated with sadness, or coldness. It seems that colours mostly generated positive feelings. However negative reactions can be produced by a mismatch between colour expectations and products. Thus when a colour scheme is chosen, designers should assess the fitness of the proposed scheme, taking care of minimising negative effects. As with colour, participants did show a high sensitivity to representation in particular if a representation matched or evoked a quality that is desirable in the product or service. Chokolic, (rated very pleasant) was liked (or disliked) by participants according to the product representation. Some participants noted that chocolates were
shown in a very nice way, so nice that it makes a participant mouth water. When disliked, it was also due to the product representation, because the design did not transmit the creamy filling that chocolates should have. Purchase decisions can be based on symbolic elements of products or services, as conveyed in pictures and processed as imagery, rather than by information [10]. Thus, paying more attention to picture effectiveness and emotions could be a key factor to the success of electronic shopping.
BIBLIOGRAPHY
1.
Lang P (1985) The cognitive Psychophysiology of emotion: anxiety and the anxiety disorders. Hillsdale NJ: Lawrance Erlbaum.
2.
Holbrock and O’Shaughnessy (1984) The role of emotion and advertising. Psychology and marketing Vol. 1, No. 2, 45-64.
3.
Singh S. (1999) Web homepages as advertisements. Communication of the ACM Vol. 42, No. 8,91-98.
4.
Kim J. and JY. Moon (1998) Designing towards emotional usability in customer interfacestrustworthiness of cyber-banking system interfaces. Interacting with computers No. 10,1-29.
5.
Rusting C. and Larsen R. (1995) Moods as sources of stimulation: relationships between personality and desired mood states. Personal individual differences Vol. 18, No. 3, 321-329.
6.
Chaudhuri, A. and R. Buck (1995) Affect, Reason, and Persuasion. Human communications research Vol.21, No. 3, 422- 440.
7.
Hassenzahl, M. Platz A., Burmester M. and K. Lehmer (2000) Hedonic and Ergonomic quality aspects determine a software appeal. Proceedings of conference on human factors in computing system CHI’200, DenHag, Vol. 2, 201-208.
8.
Morris J. (1995) Observations: SAM: the Self – assessment Manikin, an efficient cross-cultural measurement of emotional response. Journal of advertising research, November / December, 1-5.
9.
Saint-Hilaire L. (1997) Comment faire des images qui parlent. Les edtions transcontinental: Quebec.
CONCLUSION AND FUTURE RESEARCH
Marketing studies have shown that emotions are more important in some types of purchase or situations. Indeed, emotions can be the key to purchasing decisions. Thus in order to see if pages were “ emotionally” successful, SAM was proposed as a method to evaluate emotions. Positive results seem to indicate that SAM is appropriate for the emotional evaluation of home pages both during development and the site life cycle. With SAM, it is possible to situate each individual page within a specific emotional space, the pleasure by arousal space, which can be divided into four quadrants for further analysis. Moreover by correlating SAM with Mehrabian and Russell [11] emotion adjectives list, it should be possible to characterise each page by specific emotions. It remains to be seen if SAM, or an equivalent interactive version developed by Desmet and al [12], could be used to evaluate the overall emotional value of sites. Further studies should also be conducted to integrate SAM with traditional usability evaluation. With this study, we were also able to gather preliminary findings on the relation between emotions and design factors. It seems that colour and product representation influence the user’s reaction. Further study is needed in this area in particular concerning the role of visual representation, as it is still not well understood. Moreover, we should also focus on the interaction between usability and emotions. It is possible that emotions and usability are correlated. If a consumer is not emotionally satisfied with a page or a site then the site might be perceived as less usable. The reverse might also be true. Lastly, emotion probably has a strong influence on people’s behaviour. Thus we should measure consumer’s behaviour for different emotional values, both positive and negative.
10. Mackay K. and D. Fesenmaier (1997) Pictorial element of destination in image formation. Annals of Tourism research Vol. 24, No. 3, 537-565. 11. Mehrabian A. and J. Russel (1977) Evidence for a Three-Factor Theory of emotions. Journal of Research in Personality No. 11, 273-294. 12. Desmet P. Hekkert P. and J. Jacobs (2000) When a car makes you smile. Advances in consumer research No. 27, 111-117.
