Not just for kids

If you have been anywhere near the popular media in the last few weeks, you have heard about of the studies by Moss, Gunn and Kubacki (in press, 2005). This is the University of Glamorgan trio that has garnered attention for having "proved" that what catches men's eyes on the Internet is different from what catches women's eyes. Shocked? Didn't think so.

Actually, to be fair to the authors, the media (in search of a sound byte) has reported both more and less than what the research really says. What Moss and colleagues have really reported is that:

1. most of the Web sites they examined are designed by men
2. most Web sites designed by men reflect a male aesthetic
3. men's ratings for sites with a male aesthetic are higher than women's ratings for the same sites

Still waiting for the shocking part, aren't you?

Actually, the research is not quite as simplistic as that. Earlier research by members of this team demonstrated that male and female graphic and product design preferences tend to be distinguishable (Moss, 1999; Moss and Coleman, 2001.) Further, they demonstrated that designers tend to prefer designs by same-gender designers.

Would the same thing, they wondered, hold for Web sites? After all, Moss and colleagues observed, males comprise the majority of industry leaders and IT workers today. Most Web sites are designed by males. Yet there are many industries where women are (e.g., cosmetics) or are becoming (e.g., higher education in the UK) the majority consumer group.

Do Web sites have genders? And, if they do, what implications would this have for the effectiveness of the brands?

Which sex is your Web site?

To identify the sex characteristics of Web sites, Moss, Gunn and Heller (in press) evaluated 60 Web sites, half designed by women and half by men across 24 characteristics. They identified 12 characteristics that robustly differentiated those designed by males from those designed by females. Most saliently, they reported that:

based on their statistical analysis, female Web sites:

• contain links to fewer sites
• use more informal language
• use more abbreviations
• show a greater tendency toward self-denigration
• tend to use more non-expert language

with respect to visual presentation, female sites:

• tend to use rounded rather than straight shapes
• tend to avoid horizontal layouts
• use more informal typography
• contain more colors (particularly white, yellow, pink and mauve)
• tend to contain more pictures of women

In contrast, male Web sites use more crests and contain more pictures of men.

Based on these findings, they concluded that there is a continuum of the male/female elements of aesthetics in design. While designers use a broad array of these elements, they tend to produce sites using elements consistent with their own gender. (See examples of male and female designs at the Moss and Gunn blog.)

But does it matter?

When opposites don't attract

Moss and Gunn (2005) extend their work to higher education sites to explore whether the male/female "gender bias" of a site correlates with its desirability by visitors of either gender.

In order to test this, they randomly selected 32 of the 121 Higher Education Institutions in the UK and evaluated their Web site pages based on the sex parameters identified in the previous study. Each site was assigned a "gender bias coefficient" based on the predominance of male or female characteristics on the home page.

Of the 32 sites selected, 30 presented with a male orientation. The predominance of male-oriented sites is not surprising. A post hoc telephone survey of 28 of the site designers revealed that 93% of the sites were developed by male only or mixed gender teams. They do not, however, report whether the two sites reflecting a female "gender orientation bias" were designed by women.

To test the impact of a site's gender bias on its perception by users, Moss and colleagues presented 64 students to rate seven home pages on a score of 1-20. The to-be rated sites included three designed by women, three by men and one commercial Web site. No gender bias values are reported for the sites. Participants were not given any particular guidance how to rate the sites.

Moss and colleagues used the rating scores to derive a "liking" factor which measured a given participant's relative preference for a site across the seven. Analysis of variance showed that men showed a significant preference for sites designed by men, whereas women showed a preference for sites designed by women.

Words to live by

Women and men have different aesthetics. People who have been married already knew this.

This work subtly reinforces some things that usability people have known for quite some time as well:

• know who you are designing for
• understand their conceptual model... or in this case their aesthetic...
• validate graphic choices and task-flow designs through field usability testing.
• test users that are representative of the priority customer groups

Moss and colleagues' work does not only demonstrate the obvious, though. It subtly emphasizes the importance of two parameters of design that usability practitioners evaluate less systematically: brand effectiveness evaluations and message/content testing.

On the Web, brand effectiveness reflects an intersection of site usability and brand presentation for target user groups. It makes sense then, that as the Web becomes more integrated with other corporate communications channels, usability testing will begin to enfold brand effectiveness evaluations more routinely. Further, the branding and branding effectiveness of the Web will be integrated into corporate communication strategies.

Message effectiveness testing is similarly up and coming. Usability professionals have spent the last decade establishing how to create usable navigation. Now that we can get people to the content, we need to focus on creating the parameters of persuasive message design.

Moss and colleagues, like Tannen before (2001), observe that there are gender differences in the way we use language – women are more conversational and informal whereas men are more self-promoting and use more "expert" language. This observed disparity may have implications for designing content to establish and build trust between an organization and its target consumer group. In short, to be effective, the language of the message must reflect the aesthetic of the target consumer group.

If you know what the target consumer group is and you know what their basic aesthetic is, you can design for that group independent of your gender.

Reader comments