Why Pretty Things Seem to Work Better
Good-looking things are judged easier to use, and the judgement is worth less than it looks. The effect is real, small, conditional, and it thins out as soon as time or money enters the room.
Answer to the question
Does aesthetic usability make things work better?
A little, on average. The best current estimate, a preregistered 2026 meta-analysis of 31 studies and 18,794 participants, found a small to medium effect of visual attractiveness on measured performance, more than twice the figure from a 2019 pooling. The judgement half of the effect fades with two weeks of use and disappears when people bid money.
Key takeaways
- A preregistered 2026 meta-analysis of 31 studies, 234 effect sizes and 18,794 participants found a small to medium effect of visual attractiveness on measured performance, with no sign of publication bias.
- The earlier 2019 meta-analysis of 25 studies and 3,025 participants put the effect at less than half that size, and the newer analysis is stricter, with 21 of its 31 studies new to it.
- In a two-week field experiment with 60 mobile phone users, the lift attractiveness gave to perceived usability waned as exposure time grew.
- When Tamar Ben-Bassat and colleagues replaced questionnaires with auction bids, bids tracked measured usability alone and ignored how the system looked.
You are three screens into a new banking app and it is beautiful. The type is set properly, the spacing is generous, somebody drew those icons by hand. You have not yet moved any money, and you have already decided the thing is well made.
Here is the part worth having before any of the evidence. The aesthetic usability effect, the tendency for a better-looking interface to be judged easier to use and sometimes to be used better, is real, small and conditional. If you are running a usability test, it will inflate what people say about ease of use on first contact. It does not reliably survive a fortnight of use, and it thins out when the person has to put money behind the judgement. So time a task rather than ask about one, ask again two weeks later, and where you can, make somebody bid.
The study everybody cites for this is a two-page conference note from 1995, in which Masaaki Kurosu and Kaori Kashimura at the Hitachi Design Center had students rate 26 cash dispenser layouts and found that beauty ratings tracked judged ease of use. We have already written about a halo running from looks to judged ease of use. The thirty years since are the interesting part, and they are a ledger rather than a verdict.
The judgement that moves first
Noam Tractinsky and colleagues gave the aesthetic usability effect the form everyone now cites, in a paper published in 2000 under the title "What is beautiful is usable". The sample was 132 third-year industrial engineering students at Ben Gurion University of the Negev, chosen because their curriculum contains no design or aesthetics teaching at all. Each used a computerised stand-in for a cash machine, with layouts pre-rated and sorted into high, medium and low attractiveness, crossed with how usable the layouts actually were. Everybody got the same four-digit code and worked through eleven tasks.
Before use, how attractive people found a layout tracked how usable they found it. After use, it still did. The paper's own finding is that the attractiveness of the system changed post-use judgements of both attractiveness and usability, "whereas the degree of actual usability had no such effect".
That is a strong result and a narrow one.
Every outcome measured was a perception. Nobody's performance is reported in the paper, and the authors say plainly that "the relative homogeneous nature of the participants in this study (all were engineering students) restricts the generalizability of the experimental results".
Performance, when actually measured
Andreas Sonderegger and Juergen Sauer put the aesthetic usability effect to a harder test in 2010 at the University of Fribourg, with two simulated mobile phones that were functionally identical and differed only in how they looked. Sixty adolescents aged 14 to 17 worked through ordinary phone tasks. The appealing handset was rated more usable, and it also produced shorter task completion times.
That is the strongest single-study evidence in this whole file that looks moved behaviour rather than opinion, and it rests on sixty teenagers and a simulation.
A year earlier, Morten Moshagen and colleagues had run 257 volunteers through a series of search tasks on a health information website, in four versions crossing good and poor visual design with good and poor usability, measuring completion times and error rates. Usability moved both. Visual design moved performance too, but only in one place: the pretty version helped when the site was hard to use. Their phrase for it is that attractiveness "compensated for poor usability by speeding up task completion".
So the performance half of the aesthetic usability effect is conditional, and the condition is that the thing is difficult. On a well-built interface there was nothing for looks to rescue.
Two weeks later, again
The group that produced the clearest performance effect also produced the finding that bounds it.
In 2012 Sonderegger and colleagues ran a two-week field experiment with 60 mobile phone users, varying both appearance and usability between people and measuring at three points across the fortnight.
Their major finding was that the positive effect of an attractive product on perceived usability "began to wane with increasing exposure time". The practitioner version, in their own words: product aesthetics "influences perceived usability considerably in one-off usability tests but this influence wanes over time".
Marc Hassenzahl had reached a complementary result in 2004 with two studies on four different skins for an MP3 player. Perceived usability moved with experience, and so did overall satisfaction with the product. Beauty did not budge. He also found that beauty depended largely on identification, the product's ability to say something about its owner to other people, rather than on anything to do with getting a task done.
Read together, the two papers say the aesthetic usability effect is a first-impression effect. The looks stay. The borrowed credit for ease of use drains away as the person finds out for themselves.
Money riding on the answer
The sharpest test in the literature asked people to bid. Tamar Ben-Bassat and colleagues, in 2006, built four versions of a data-entry application with attractiveness and usability varied independently, then had participants evaluate them in five different ways: by questionnaire alone, or by auction as well, with and without cash incentives, with and without prior experience of the system.
The questionnaires behaved as the aesthetic usability effect predicts, with each attribute colouring ratings of the other. The auction bids did not, and that is the single most awkward result for anyone selling the aesthetic usability effect as a commercial argument. "Auction bids were only influenced by the system's usability: bids corresponded to the objective performance levels that could be attained with the different systems." Tractinsky, who established the effect six years earlier, is the third author on the paper that bounds it this way.
Ask people what they think and beauty counts. Ask them to bid, and only usability does.
Then there is the result that reverses the arrow. Alexandre Tuch and colleagues, in 2012, had 80 participants each use one of four versions of the same online shop, rating it before and after use on four validated instruments. "Results show that aesthetics does not affect perceived usability. In contrast, usability has an effect on post-use perceived aesthetics." Frustration, they suggest, lowers how good a thing looks. Their own hedge is that the reversal holds "under certain conditions", namely a strong usability manipulation alongside a medium to large aesthetics one.
The pooled estimate, finally
| Study | What was changed | What was measured | What happened |
|---|---|---|---|
| Tractinsky and colleagues, 2000 | Look of a cash machine screen | Judgements of 132 students | Looks moved judged ease of use, real usability did not |
| Moshagen and colleagues, 2009 | Look of a health website | Times and errors, 257 people | Looks helped only where the site was hard to use |
| Sonderegger and Sauer, 2010 | Look of a simulated phone | Task times, 60 adolescents | Appealing phone was faster and judged easier |
| Sonderegger and colleagues, 2012 | Look and usability of real phones | Two weeks of use, 60 owners | The judgement effect faded with exposure |
| Ben-Bassat and colleagues, 2006 | Look of a data-entry tool | Questionnaires against auction bids | Bids tracked usability alone |
| Tuch and colleagues, 2012 | Look and usability of an online shop | Ratings before and after use, 80 people | Usability changed judged looks, not the reverse |
Two meta-analyses have tried to pool the performance half of the aesthetic usability effect, and they disagree by more than a factor of two.
The first, led by Meinald Thielsch in 2019, gathered 25 studies, 101 effects and 3,025 participants, and asked only about objective performance. It found a small significant overall effect, with wide variation between studies that none of the moderators it could test would explain, and it closed with a call to action to its own field.
The second is the current best estimate, and it is where this article's title comes from. Marleen Schlamann and colleagues published a preregistered meta-analysis in 2026, preregistered meaning the analysis plan was filed before the results were seen so it could not be chosen afterwards. It covered 31 studies, 234 effect sizes and 18,794 participants, admitting a study only where the authors had shown their aesthetic manipulation actually worked. The pooled effect of attractiveness on performance was small to medium, more than twice the 2019 figure, with no sign of publication bias.
The two numbers differ because the newer analysis is stricter and mostly newer, with 21 of its 31 studies appearing in no earlier pooling, and because it excluded much of the work that mixed product attractiveness with interface attractiveness. The right move is to report both and the reason, not to average them.
The part still unsettled
Richard Shotton's The Illusion of Choice, the 2023 book that seeded this piece, puts the caution well on page 160: "while the halo effect reveals that many metrics correlate, it doesn't show they move in lockstep". The primary literature then vindicates that sentence more thoroughly than the chapter claims. Its own apparatus is weaker: chapter 14 names the 1995 cash dispenser study on page 164 and the chapter's reference list has no entry for it.
What the pooled evidence on the aesthetic usability effect cannot establish is longer than what it can. Variation between studies remained large in 2026 and unexplained by anything the authors could code. Two moderators reached significance, the device used and whether typography was manipulated, and neither survived being tested together. The authors grade the literature they pooled as mediocre on a standard quality instrument, with poor reporting of methods and outcomes the main fault.
Nothing in this file tested whether attractive interfaces make people forgive errors, which is the claim most often repeated in design writing. Nothing measured a shipped product over a year. The samples are 132 engineering students, 60 adolescents, 80 shoppers and 257 website visitors, which is the usual position a design decision has to be made from.
Which leaves the question a designer most wants answered, and it is open. The 2026 analysis raises the possibility that the effect is largest in prototypes rather than in finished software and websites, then takes it back, because that moderator lost significance under a stricter method of estimating the variation.
If it is true, thirty years of evidence collected on mock-ups says less than it looks about anything that has shipped. If it is false, the effect is small, real and worth a design budget. Those are different worlds, one where a good-looking product genuinely helps a person finish the job and one where it only buys the benefit of the doubt, and the studies as they are currently reported cannot tell you which of them you are selling into.
Common questions
Does aesthetic usability make things work better?
A little, on average. The best current estimate, a preregistered 2026 meta-analysis led by Marleen Schlamann covering 31 studies and 18,794 participants, found a small to medium effect of visual attractiveness on measured performance, more than double the estimate from a 2019 pooling of 25 studies. Variation between studies was large and the authors could not explain it.
What is the aesthetic usability effect?
It is the tendency for people to judge a better-looking thing as easier to use. It was named from a 1995 study of cash dispenser layouts and established experimentally by Noam Tractinsky and colleagues in 2000, who found that how attractive a system looked moved judgements of its usability while its actual usability did not.
Does the aesthetic usability effect wear off?
The judgement part does. A two-week field experiment with 60 mobile phone users, published in 2012 by Andreas Sonderegger and colleagues, found the lift that attractiveness gave to perceived usability waned as exposure time grew. Marc Hassenzahl had already found that perceived usability moves with experience while judged beauty stays put.
Should I trust a usability test on a good-looking prototype?
Not on its own. Ask people to do timed tasks rather than rate ease of use, repeat the test after a fortnight, and if you can, attach a cost to the answer. When Tamar Ben-Bassat and colleagues had people bid at auction rather than fill in a questionnaire, the bids tracked real usability and ignored how the system looked.
Sources
Tractinsky, N., Katz, A. S., & Ikar, D. (2000). 'What is beautiful is usable.' Interacting with Computers, 13(2), 127-145. doi:10.1016/S0953-5438(00)00031-X. [Full text read, author's copy at the Ben-Gurion University faculty site; 132 third-year industrial engineering students, eleven tasks, perception measures only]
Hassenzahl, M. (2004). 'The Interplay of Beauty, Goodness, and Usability in Interactive Products.' Human-Computer Interaction, 19(4), 319-349. doi:10.1207/s15327051hci1904_2. [Abstract read, OpenAlex record; two studies, four MP3-player skins; sample sizes not stated in the abstract]
Moshagen, M., Musch, J., & Goeritz, A. S. (2009). 'A blessing, not a curse: Experimental evidence for beneficial effects of visual aesthetics on performance.' Ergonomics, 52(10), 1311-1320. doi:10.1080/00140130903061717. [Abstract read, OpenAlex record; 257 volunteers, health information website, completion times and error rates]
Sonderegger, A., & Sauer, J. (2010). 'The influence of design aesthetics in usability testing: effects on user performance and perceived usability.' Applied Ergonomics, 41(3), 403-410. doi:10.1016/j.apergo.2009.09.002. PMID 19892317. [Abstract read, Europe PMC record; 60 adolescents aged 14 to 17, simulated mobile phone]
Sonderegger, A., Zbinden, G., Uebelbacher, A., & Sauer, J. (2012). 'The influence of product aesthetics and usability over the course of time: a longitudinal field experiment.' Ergonomics, 55(7), 713-730. doi:10.1080/00140139.2012.672658. [Abstract read, OpenAlex record; two-week field experiment, 60 mobile phone users, exposure time as a repeated measure]
Tuch, A. N., Roth, S. P., Hornbaek, K., Opwis, K., & Bargas-Avila, J. A. (2012). 'Is beautiful really usable? Toward understanding the relation between usability, aesthetics, and affect in HCI.' Computers in Human Behavior, 28(5), 1596-1607. doi:10.1016/j.chb.2012.03.024. [Abstract read, University of Basel institutional repository record; 80 participants, four versions of one online shop]
Ben-Bassat, T., Meyer, J., & Tractinsky, N. (2006). 'Economic and subjective measures of the perceived value of aesthetics and usability.' ACM Transactions on Computer-Human Interaction, 13(2), 210-234. doi:10.1145/1165734.1165737. [Abstract read, Crossref deposited abstract; auction bids tracked usability only; sample size not stated in the abstract]
Thueuring, M., & Mahlke, S. (2007). 'Usability, aesthetics and emotions in human-technology interaction.' International Journal of Psychology, 42(4), 253-264. doi:10.1080/00207590701396674. [Abstract read, Crossref deposited abstract, truncated; used for the two-factor framing only, not for a figure]
Thielsch, M. T., Scharfen, J., Masoudi, E., & Reuter, M. (2019). 'Visual Aesthetics and Performance: A First Meta-Analysis.' Proceedings of Mensch und Computer 2019, 199-210. doi:10.1145/3340764.3340794. [Full text read, author's copy at thielsch.org; 25 studies, 101 effects, 3,025 participants, overall g = 0.12, no significant moderators]
Schlamann, M., Nestler, S., & Thielsch, M. T. (2026). 'Attractive Things Do Work Better: A Meta-Analysis on Visual Aesthetics and User Performance.' International Journal of Human-Computer Interaction, advance article, 1-43. doi:10.1080/10447318.2026.2664081. [Full text read, Taylor and Francis, open access; preregistered, 31 studies, 234 effect sizes, 18,794 participants, overall g = 0.29, no publication bias detected; supplement at doi:10.5281/zenodo.18777943]
Kurosu, M., & Kashimura, K. (1995). 'Apparent usability vs. inherent usability: experimental analysis on the determinants of the apparent usability.' Conference Companion on Human Factors in Computing Systems (CHI '95), 292-293. doi:10.1145/223355.223680. [Full text read 2026-10-16, on the vault record; cited in full in article 99 and only mentioned in passing here]
Shotton, R. (2023). The Illusion of Choice: 16 1/2 Psychological Biases That Influence What We Buy. Harriman House. Chapter 14, The Halo Effect, pages 160 and 164. [Read in full; vault reliability audit at Books/The Illusion of Choice/The Illusion of Choice.md]