Aesthetic appeal versus usability Implications for user satisfaction

Aesthetic appeal versus usability: Implications for user satisfaction Gitte Lindgaard & Cathy Dudek Carleton HOTLab Ottawa, Canada

2

Satisfaction • …is the poor cousin of usability • Satisfaction defined as attitudinal – Avoid negative feelings – Measured in rating scales – Outcomes, summaries • We are interested in the experiential – Process – Construct 3

• One site tested was very high in appeal and very low in usability • Appeal = reliable ‘wow’ effect • Usability: • Heuristic evaluation found 157 unique problems • 121 of these were severe • A subset of these were exposed in the 8 usability tasks 4

Research question • Does the first impression persist after completing usability tasks? • Or do users change their mind after encountering serious usability problems? 5

Measures • Satisfaction: proportion of positive statements in – – – Perceived usability Perceived aesthetics emotion likeability expectation (Lindgaard & Dudek, 2001) 6

Issues raised here • Task demands – Do users who anticipate a usability test pay more attention to usability problems when first browsing a site than users who do not expect a test? – If so, perceived usability scores will be lower for the former than for the latter 7

Experimental design 8

Findings • Subjects completed, on average, 3. 8 of 8 tasks successfully • no subject completed all the tasks • no task was completed by all subjects So, it is safe to conclude that usability levels were very low 9

Findings: perceived usability 33% 31% 13% Site was not perceived as usable by either group Site was seen to be less usable after than before the test 10

Findings: perceived aesthetics 91% 87% 91% Site was perceived as beautiful by both groups Site remained beautiful after the usability test 11

Findings: emotion 87% 66% 20% The ratio of positive: negative emotion statements was lower for the test group both before and after the test. 12

Findings: likeability 79% 49% 25% The ratio of positive: negative likeability statements was Lower for the test group both before and after the test 13

Findoings: expectation 64% 11% 5% The ratio of positive: negative expectation statements was Very much lower for the test group before and after The test 14

Findings: satisfaction 66% 51% 25% Satisfaction appears to be determined by several factors 15

First Impressions • The first impression apparently rests on aesthetics • The perception of beauty persists, but • Perceived usability, likeability and expectation change after facing serious usability problems 16

First impression • Formed in an instant (3 -5 msec) • Based on changes in arousal levels (Berlyne, 1971; 1972) • Evoked via the amygdala, not via the hypothalamus (Damasio, 2000; Le. Doux, 1994; 1996; Goleman, 1996) • Can be overridden by pre-exposure decision (Epstein, 1997) • Is difficult to change – confirmation bias (Doherty, Mynatt & Tweney, 1977) 17

One question is… • Does emotion precede cognition… (Zajonc, 1980; Bornstein, 1992) • …or is it the other way around? (Epstein, 1997) • I. e. are first impressions ‘what my body tells me to feel’, or are they ‘what my brain tells me to think’? 18

Issues raised here • Confirmation bias – If the first impression drives satisfaction, it should not change after usability test – If usability drives satisfaction, it should change after usability test – and it should vary between high- and lowusability sites 19

Experimental design • Group 1 (n = 40): Browse interview/ratings • Group 2 (n = 40): Browse interview/ratings usability test interview/ratings • 2 e-commerce sites tested • Scores: • (a) proportion of positive statements • (b) WAMMI (Kirakowski et al. 1998) 20

Results: Satisfaction Mean satisfaction scores before & after test Main effect, before/after (p <. 001); main effect for web site (p , <. 05) 21

Perceived usability Mean perceived usability before/after test Main effect before/after (p <. 001); main effect for web site (p < 22

Aesthetics Mean aesthetics score before/after test No significant effects 23

Satisfaction, before test only Mean satisfaction scores, first interview Main effect for web site (p <. 05); main effect for subject-group (p < 24

Perceived usability before test Mean perceived usability scores before test only Main effect for web site (p <. 001) 25

Conclusion • Confirmation bias – Aesthetics scores taken on their own did not differ before the test –. . and they did not change after test – Confirmation bias on the aesthetics dimension • But – Satisfaction scores decreased after the test – Perceived usability scores decreased – No confirmation bias on overall satisfaction or on usability 26

Conclusion • Task demands – Lower satisfaction scores for subjects expecting a usability test than for browsing-only subjects suggest that task demands do affect attention to usability – Subjects are sensitive to actual usability levels – As evidenced both in satisfaction scores and in perceived usability scores 27

Conclusion • So, satisfaction appears to be driven partly by actual usability • Aesthetics judgments appear to be independent of perceived usability 28

Next steps • Currently developing satisfaction scales that enable developers to pinpoint where to improve their sites to increase user satisfaction 29

So, now to aesthetics • Gary Fernandes MA thesis: – 125 sites collected, all of unknown companies – Preliminary study, n = 22 – Selected 25 best and 25 worst sites – N = 30 – Viewed sites for 500 msec, then rated visual appeal in two rounds 30

Measurement scale Very Unattractive 31 Very Attractive

33

34

35

36

37

Appeal ratings, study 1 38

Appeal ratings, study 2 39

Visual appeal ratings, study 1 vs study 2 40

Conclusion • Aesthetics judgments are made very quickly • They are highly robust • New results show that they persist even when subjects are able to inspect the home page for an unlimited period of time 41

Next steps • Expose stimuli for 40 msec • Collect genres of sites • Evolve tool enabling companies to test their own web site against others 42

43