upa - home page JUS - Journal of usability studies
An international peer-reviewed journal

When Left Might Not Be Right

Xristine Faulkner and Clive Hayton

Journal of Usability Studies, Volume 6, Issue 4, August 2011, pp. 245 - 256

Article Contents


Web site layout was once an area of bitter dispute but over the last few years accepted best practice design principles seem to have emerged, resulting in a consensus about what makes a good Web page. Web sites today have remarkable conformity when one considers the size of the Web. How much this is attributed to Web designers copying what they see as successful designs, or as a response to real usability issues, is unclear.

Web site layout dictates how easy or difficult the navigation of a site will be, so getting it right is important. Also research shows that users can form an impression of a Web site in as little as 50 milliseconds (Lindgaard et al, 2006), and this might cause them to decide whether to stay on the site or to click to the next one. A Web site needs to grab the user’s attention and keep it for the duration of a sale, if the site is a commercial one. Keeping users on a Web page therefore becomes of paramount importance to the designer of a Web site and particularly for a commercial one.

The Web is huge and users are fickle so the problem for the Web designer is whether to stand out from the crowd and produce something different or to capitalize on the user’s Web skills and use a design s/he already knows. The thinking is that if a site follows the conventions of every other site then it is easier to use because the user knows exactly what to expect. Many usability experts would give this advice: Stick to what users know, and don’t try anything different. Jakob Nielsen, for example, has been vociferous in persuading Web designers to follow the trends because, he argues, it makes sense from a user perspective (1999a; 2004). However, there is an alternative view that if a Web page design echoes every other design on the Internet how can a Web site look special to the user and cause them not only to stay on the Web site but to return to it in the future?

A conventional Internet Web site displays menus on the left-hand side, often with the navigation panels (or menus) repeated at the bottom and/or top of the screen. Indeed, the left-hand justified menu accounts for the majority of Web interface designs, followed by navigational menus at the top (Mardiros, 2006). This consensus design and the reasoning behind it are probably best summed up by Mardiros Internet Marketing who in their guidelines for good Web page design say confidently: “Left Web site navigation is the most common type of navigation” (2006, Design good primary web site navigation section, para 1, bullet point 1).They go on to suggest that “For English language based Web sites people read from left to right. Thus, a menu situated on the right-hand side would be difficult to use” (2006, Design good primary web site navigation section, para 1, bullet point 3).

Certainly this view has common sense on its side. For example, English is read from left to right so perhaps it could be deduced from that fact that for a left to right language, menus would be better placed on the left in order to conform to expectations of a book or newspaper. In other words, the Web page should act like the pages of a book and treat everything as if it is a progression from left to right and top to bottom. Conforming to expectations is one of the principles of usability. A predictable model that fits in with the rest of the world is easier for people to use.

With common sense in mind, it is therefore not surprising that the popular view is that the main navigation elements on a Web site should be placed on the left-hand side of the screen. Usually, the left justified menu is accompanied by a section at the top of the screen that consists of a site’s identifying elements (for example, banners, site names, logo) and perhaps additional navigation components. This layout is referred to as an "inverted-L navigation" or "L-shaped navigation" (Kalbach & Bosenick, 2003), and this form of navigation is very popular on the Internet today.

Some usability experts believe that users categorize what they see on the screen into genres. The argument is that they do this very quickly and that the categorization helps them to navigate the Web site much faster than they would be able to if they didn’t have a fixed idea of a genre in their head. The conventions of a particular Web genre are the visual elements and components that the user expects and recognizes. The user will recognize and categorise those conventional elements before they examine the content, but the recognition will make it easier for them to navigate the Web site (Santa Maria & Dyson, 2008).

However, although research does suggest that violating expectations and conventions of the genre does have a negative effect on user performance, it also indicates that this negative effect does not last for long and that users rapidly recover from any lost performance levels. In other words, Internet users very quickly accommodate to a different design and its expectations (Santa Maria & Dyson, 2008). Research suggests that the second visit to the site sees none of the loss of performance time that the first visit might have possibly produced for the user. Novel design should not, therefore, be avoided because users can clearly adapt to the unconventional. However, this is a double-edged sword in that users might be put off by an unconventional design that looks complicated, and thus never persevere long enough to recover lost performance time.

Putting aside what common sense might tell us, there have been some small studies of left-hand versus right-hand menus that show conclusively that the left-hand menu is both faster and more popular with users than the right-hand menu. For example, Kingsburg and Andre carried out two studies with 16 users and found in both of their studies that selection from a left-hand menu was faster than from a right-hand menu (2004). However, their research also showed that selections were best done from the same panel, whether that was on the right or left. Thus it is better to have a single design, either on the left or the right, rather than a mixed navigational method that requires the user to select from both left and right panels (Kingsburg & Andre, 2004). This is hardly surprising and is both predicted and supported by Fitts’ Law. (1954).

There have been some experiments on right-hand menus, but these have tended to assume that left-hand menus would always perform much faster. There have been relatively few studies on where a navigation menu might be best placed on the screen. Kalbach and Bosenick mentioned this lack of research when they carried out their work on menu placement. They identified just a few studies prior to their comparison between left and right justified menus (2003). Spool’s study suggested that top and bottom placed menus were more successful than left justified ones (1997).  However, the National Cancer Institute’s usability researchers suggested a right-hand navigation system is actually more effective. They showed that users click on menu items in the right-hand margin more efficiently than on menu items placed on the left. They concluded that this efficiency was because right-handed items are closer to the scroll bar thus allowing users to move the pointer quickly between the scroll bar and the navigational item (Kalbach & Bosenick, 2003).

Nielsen suggested that left-hand justified menus were becoming de facto in 1999, and that if “80% or more of the big sites do things in a single way… you have to comply” (1999b, para 2, bullet point 1). Although he conjectures that in theory right-hand navigational panels ought to be faster because they are closer to the scroll bar, he does not go on to suggest them as a better means of designing a Web page (Nielsen, 1999b).  He was reluctant to suggest a method that was different from the norm because he believed that standards inevitably make for faster performance (Nielsen, 1999a, 1999b, 2004).  Nielsen thought that sites that don’t stick to the norm were too difficult for users to use and he described the standard as when “80% or more of Web sites use the same approach” (2004, para 3, bullet point 1). Thus, by 1999, the left-hand justified menu was already entrenched as good design and therefore in Nielsen’s view it ought to be emulated by would-be Web site builders (1999a, 2004).

A study in 2008 on browsing behaviour by sighted users undertaken by Michailaidou and her colleagues suggested that common gaze patterns begin at the salient parts of the screen, move to the main content, the header, the right column, the left column, and then finish at the footer.  In theory then the right column is scanned before the left one and ought to be marginally faster although in reality the rate of scan by the human eye is so fast that the difference in most cases would be negligible. However, given that psychologically the right column is given priority—in that it is scanned sooner—it follows that the right-hand area of the screen would be worth investigating as a possibility for the placement of the main navigational panel.

Kalbach and Bosenick’s study that compared right-hand and left-hand navigational menus found no evidence that left-hand menus were significantly faster though they did find a small difference between the performances of the two placements. They concluded that top-aligned menus performed the best, but found no evidence for right-hand menus being less attractive to users (2003). Their study is one of the few studies that set out with a hypothesis that there would not be a significant difference between the performance of the two navigational menu placements—left and right.

One of the reasons given for sticking to convention and following expectations is that users become disorientated when they are given layouts that don’t conform to their expectations. This is more of a problem with older users who prefer more predictable and unchanging systems (Arch, 2009). Disorientated users refer to themselves as “feeling lost.” However, a study by Santa Maria and Dyson found that although users are disorientated when faced with a layout they aren’t expecting, they do recover rapidly and search time reverts to what would be expected from using a conventional, anticipated layout (2008). This suggests that novel design has only a short term effect on user performance.

Finally, the reasons given for left-hand menus being faster seem vague and unclear and not well supported by any convincing psychological explanations or research findings. Similar arguments can be found over which side of the road it is best to drive on. There are equally valid arguments to support both left- and right-hand driving, and the fact that both solutions have been adopted probably indicates that neither one is better than the other. Although, one could argue that in this case standardisation is desirable and that a novel—drive which side you like—would not be suitable.

Therefore the question remains: If left-hand menus are indeed faster than right-hand ones why might that be? If Internet users look for content on the left side because they read from left to right (Mardiros, 2006), then what about Internet users who read from right to left? Do they look for content on the right? Have Internet users been conditioned to expect particular layouts by designers copying each other’s designs and therefore expect navigational tools to be on the left? In other words, have Internet users developed an expectation of a particular type of “Web genre”? Given these questions, it was decided to carry out an experiment to see if factors influencing the time to purchase differences between left- and right-hand justified menus could be isolated, and thus the left-hand preference be explained in terms of concrete human factors rather than conventional wisdom.

The demographic factors included in the sample were direction of writing of the initial language acquired by the user, gender, handedness, Internet experience, computer experience, time spent online each day, and age.


The following hypotheses were formulated:


Previous | Next