Designers often use flat design when they want their interfaces to be perceived as trendy or modern. There’s an underlying assumption that flat UIs are targeted at Millennials, who are expected to appreciate this design style more than their parents and grandparents. So it’s ok if older users don’t quite ‘get’ how to use flat UIs, these designers argue. They aren’t the target group anyway.

In our qualitative research, young adults do seem to be slightly more effective at navigating flat UIs. However, they are still susceptible to click uncertainty — an uncomfortable emotional state that occurs when users aren’t sure where to click next.

So we wondered, is the click uncertainty generated by flat interfaces worth the perceived payoff of aesthetic appeal? How do young and older adults feel about this design style? Are flat designs actually considered more attractive by young adult users than by older adult users?

Methodology

To explore these questions, we conducted an online survey. We showed five different screenshots of websites to 229 young adults (18–25 years old) and 228 older adults (35 years or older). We asked them to rate the attractiveness of each screenshot on a seven-point semantic differential scale (1 = unattractive, 7 = attractive). We also asked them describe each screenshot by choosing 5 words from a given word list (a modified Microsoft Desirability Toolkit test).

To avoid potential bias caused by brand recognition (for example, BMW aficionados rating the BMW website more positively), we selected four fake-content fake-branding screenshots of theme demos from two website builders (Squarespace and Wix). An additional screenshot was real but safely unrecognizable: a marketing site for an unreleased application.

The four flat-design screenshots used in our study were a variety of different styles of completely flat designs.
Bryant Hill
Flat 1 ("Bryant Hill")
lifeaweso.me
Flat 2 ("Lifeaweso.me")
Rochette
Flat 3 "Rochette"
Van Mar
Flat 4 ("Van Mar")

Four of the five screenshots were a variety of different styles of completely flat designs. The fifth screenshot was a traditional nonflat design with some elements of skeuomorphism. We included this as a control. Without it, we couldn’t be sure if any differences in the ratings were due to some idiosyncrasy of the young-adult population (for instance, a tendency to give higher ratings in general, regardless of the design).

Steakhouse
‘Steak House’, the traditional design included in the survey, acted as a control. It was rated identically by young and old users.

Results

A bar chart showing the attractiveness ratings for each design.
The average attractiveness rating for each screenshot for both young adults and older adults ( 1 = unattractive, 7 = attractive)

A multiple regression analysis with age and flat designs as independent variables shows that age is a statistically significant predictor of the rating for flat UIs (p<0.00001), even when we control for the different flat designs. Thus we can say that young adults do rate flat designs higher than older adults do. On average for all four flat designs, young adults rated the flat designs more attractive by 0.53 rating points, compared to older adults. The 95% confidence interval for the difference between the two age groups is 0.4 to 0.7.

With respect to individual designs, we found statistically significant differences in the attractiveness ratings of three out of the four flat designs. There was no statistically significant difference between the two groups in the attractiveness rating of the control.

lifeaweso.me design
Flat 2 (lifeaweso.me) in our survey generated the largest difference in attractiveness ratings between the two groups in our study.

The largest difference in rating averages was for Flat 2: the extremely flat and minimalist site. Young adults gave this screenshot an average attractiveness rating of 4.11 out of 7. Older adults rated the screenshot 1.26 points lower at an average of 2.85.

When we look at the descriptive words that respondents chose for Flat 2, we can get some insight as to why older adults might have rated this design lower. 71% of older adults chose the word ‘boring’ to describe the lifeaweso.me design — 29% more than the proportion of young adults who thought ‘boring’ described the design. Conversely, 59% of young adults described the design as ‘professional,’ compared to 29% of older adults.

Interestingly, the single nonflat design included in the survey (‘Steak House’) received nearly identical attractiveness ratings and descriptive words from both groups. Young adults gave it an average rating of 4.96, while older adults gave it 4.97. 71% of young adults and 75% of older adults described the design as ‘professional’. The closeness in both attractiveness ratings and descriptive words is remarkable — no similar effect was seen in any of the flat designs. This result suggests that flat designs may be more polarizing than traditional or skeuomorphic designs, perhaps as a consequence of their novelty.

This table contains the top five most-selected words for young adults and older adults for the control design. The selections are nearly identical — a phenomenon that did not occur for any of the other designs included in the survey.

Young adults’ top 5 words used to describe the control

Older adults’ top 5 words used to describe the control

% of young respondents selecting that word

% of older respondents selecting that word

‘professional’

71%

‘professional’

75%

‘familiar’

60%

‘familiar’

67%

‘trustworthy

44%

‘trustworthy’

47%

fresh

43%

‘fresh’

46%

‘impressive

35%

‘impressive’

41%

Discussion

Before we dive into a discussion of what these results might mean, we have to acknowledge the limitations of this study.

  • We only measured users’ perceptions of the attractiveness of the web designs; we didn’t study actual task performance. There is a positive relationship between subjective impressions and objective performance, but the correlation is not perfect even when users rate the site after interacting with it. In other words, there’s much more to usability than how attractive things look at the first glance. Additionally, flat design can reduce user efficiency, with the associated drop in conversion rates and long-term use and satisfaction.
  • There is a wide range of possible ways to design a flat UI, even a totally flat UI. We only picked 4 out of the universe of possible flat designs. We must acknowledge that we can’t generalize our findings here to all flat designs.
  • The photographs and subject matter inside the screenshots likely influenced ratings and word selection. For example, Flat 3 features a big image of a fancy watch. That screenshot was the only flat design that did not have a statistically significant difference in attractiveness ratings. Additionally, about half of both young and older adults described the design as ‘expensive.’ We suspect that the big image of the expensive-looking watch affected users more than the stylistic design of the website. We know from previous research that users care more about content than about style, and that the content will dominate any user feedback. This doesn’t mean that style is irrelevant, just that it’s harder to study — which is why we tested 4 designs with a range of topics.
  • To make firm conclusions from this type of study, we would need to conduct a much larger survey, with dozens if not hundreds of screenshots, to help dilute the effects of individual designs and to better represent the diverse body of flat UI design.

That said, the results of this study do imply that young adults on average interpret some flat designs as more attractive than older adults do. That difference is highly statistically significant. However, we have to consider whether the amount of that difference is also practically significant — that is, whether the magnitude of the difference justifies the potential risks.

We need more research to determine if young adults prefer all — or even most — flat designs over more traditional ones. But given the current findings from our sample of four flat designs, designers must ask themselves whether a very slight increase in aesthetic appeal is worth the potential cost in lowered interaction efficiency, click uncertainty, and alienating older user groups.