Syntagm > Design for Usability > Resources > Articles

Applying Research to Design:
Bridging a Widening Gap

(interactions magazine March/April 2004)

The world of e-commerce is a competitive one, with unrelenting pressure to attract users to your site and improve conversion rates (the number of users who become purchasers). Imagine then your delight at the prospect of increasing sales through the addition of a few carefully chosen images. A short paper entitled Do pretty women inspire men to discount the future? appears to offer great promise. It was originally published in the Royal Society's Biology Letters in December 2003 and was quickly reported worldwide in both the science and national press . The authors reported that male participants were more inclined to accept smaller rewards now rather than larger rewards later, after being shown images of attractive women. Could this mean that men might be more inclined to buy from a web site showing such images? Would they be more inclined to pay for quicker delivery? So what to do now?

  1. Engage a team of researchers to investigate the likely impact of the
    "pretty women" effect on your male-oriented site
  2. Wait for a usability evangelist to either praise or condemn the practice
  3. Plaster your site with gratuitous images of attractive women (figure 1)

Photo of an attractive female at a call centre

Figure 1: Can this attractive image improve your sales?

Sadly, in most cases the realistic answer is the last: add the images and see whether there is any improvement. The first option, research, is expensive and time-consuming, plus many e-commerce companies would not know where to start. The second alternative is too hit-and-miss. It could be years before anyone in the HCI or usability communities decides to provide a practical test of this effect for e-commerce sites.

This is a fairly extreme example of the research-to-design gulf since the original paper was oriented towards human behavior rather than HCI (although the testing was actually done on computers). However, there are many other examples. In a study of menu design reported by Wichita State University, the researchers show that “index” menus (similar to those shown in figure 2, but occupying most of the page) were both more efficient and better-liked by participants than either horizontal or vertical “cascading” menus. Designers and usability specialists might be tempted to take this as proof that cascading menus should be avoided. But before embracing this conclusion we should consider more carefully what was being compared. The “index” menus tested occupied most of the web page – a design that is rarely seen in practice these days since it virtually precludes any substantial content on the page. This approach also presents significant problems for navigational consistency, since the menus disappear once a new page is loaded.

Indexed menu (subject headings with topics under each)

Figure 2, "Indexed" menu similar to that used by Bernard and Hamblin

A more realistic menu design to test against cascading menus can be seen in figure 3. This shows a typically lengthy, cramped, left navigation menu that usually has to be scrolled to view all items. This is a much more common design than that shown in figure 2 and of greater relevance to new web sites. So we still have a research-to-design gulf, albeit somewhat smaller than the "pretty women" case.

Long menu in single column with many items

Figure 3, A more typical menu design found in current e-commerce sites

Naturally there will always be a gulf between research and design. But it seems to me that the gulf between HCI research and web design is increasing rather than diminishing. HCI as a field is growing in popularity, but a great deal of the resulting research is aimed squarely at the HCI community. Even design-oriented research, such as the index versus cascading menu example, is of limited practical use because of its narrow focus. At the design and end of the gulf, it is becoming increasingly difficult to find designers or usability specialists with a good grounding in HCI, since many have come from other backgrounds (good from a multidisciplinary viewpoint, but bad for the sensible interpretation of research.)

Longer-term we may need a more design-oriented discipline to fill this expanding void. As much as I hate to concoct new roles, I can see the need for an "interaction science" that would bridge the gulf between psychology, HCI research and interaction design in much the same way as materials science does between physics, chemistry and civil engineering. In the short term we can try to narrow the gap by...

  • doing more design-oriented HCI research;
  • supporting collaboration between researchers and designers in order to make
    more research relevant to commercial needs;
  • trying to ensure that designers and usability specialists are motivated to study HCI;
  • promoting the idea that HCI is an important area of experience for practitioners.

And if in the meantime you start to notice images of attractive women appearing on web sites even more frequently than they do at the moment, don’t say I didn’t warn you.


Wilson, M., Daly, M. 2003, Do pretty women inspire men to discount the future?,
Royal Society Biology Letters, DOI 10.1098/rsbl.2003.013 (discussed on the New Scientist web site at http://www.newscientist.com/news/news.jsp?id=ns99994469)

Bernard, M., Hamblin, C., 2003, Cascading versus Indexed Menu Design, Usability News,
Vol 5, Issue 1, http://psychology.wichita.edu/surl/usabilitynews/51/menu.htm

The Author

William Hudson is principal consultant for Syntagm Ltd, based near Oxford in the UK. His experience ranges from firmware to desktop applications, but he started by writing interactive software in the early 1970's. For the past ten years his focus has been user interface design, object-oriented design and HCI.

Other free articles on user-centred design: www.syntagm.co.uk/design/articles.htm

© 2001-2005 ACM. This is the author's version of the work. It is posted here by permission of ACM for your personal use. Not for redistribution. The definitive version was published in interactions, {Volume 11, Issue 2, March + April 2004} http://doi.acm.org/10.1145/971258.971290

© 1995-2013 Syntagm Ltd | Contact Us