Jakob Nielsen’s reputation as a Web
usability expert makes his writings required reading for Web builders. You may
not agree with all of his conclusions and recommendations, but you should be
aware of them nonetheless. In the Sept. 13, 2004
edition of his Alertbox
, he makes a case for standardizing numerous details
of Web page design.

As usual, he bases his usability
recommendations on solid research and observations. However, I believe his
conclusions, and especially his recommendation to standardize design, are
slightly off base. I suggest a somewhat different interpretation of his basic

The case for standardization

One of his key points is that Web
visitors view the Web as one large, integrated resource; therefore, they expect
a degree of consistency in the common user interface elements that appear on
most Web pages. He also reminds Web builders that visitors spend most of their
time on other Web sites, and they arrive at your site with expectations based
on experiences at those other sites.

He goes on to analyze the treatment
of several design elements on more than 50 different Web sites, and rates the
degree of consistency using the following criteria:

  • Standard:
    Elements that receive the same treatment on more than 80 percent of the Web
    sites. According to him, visitors expect these elements to always work the
  • Convention:
    Elements that receive the same treatment 50-79 percent of the time. Visitors
    expect these elements to usually work the same.
  • Confusion:
    Elements that receive the same treatment 49 percent of the time or less.
    Visitors don’t know what to expect when they encounter these elements on a new
    Web site.

According to the article, here’s the
breakdown of the standardization of the design elements on the surveyed Web

  • Conforming
    to standard:
    37 percent (examples: a logo in the upper left corner, a
    search box on the home page)
  • Conforming
    to convention:
    40 percent (examples: distinctive color for visited links,
    shopping cart links in the upper right corner of the page)
  • Confusing
    23 percent (examples: location of the main navigation scheme,
    search box, and help)

His interpretation of this data is
that inconsistent treatment of nearly a quarter of the design elements commonly
found on Web pages causes usability problems for visitors, and that Web
builders should address this deficiency by adopting design standards for those

Suggesting standards with a spin

I believe he’s right on target when
he stresses that usability is a matter of meeting user expectations. However, I
differ with his conclusion that greater standardization of Web page design is
the best (or only) way to achieve better usability.

First, let’s consider what visitors
really expect. I think he goes a little too far with his assumption that
visitors perceive the Web as a “single, integrated resource.” The
question is: How thoroughly integrated do visitors expect various parts of the
Web to be? To draw a shopping analogy, he seems to think visitors expect the
Web to be like one big department store, where the same rules apply throughout
the store even though different departments may sell different products. I
suggest that user expectations may be more like a mall or shopping center, where
several stores share a common location and may keep the same hours, but each
store has its own distinct identity. Shoppers tolerate, and even expect, some
differences from store to store. After all, it’s those differences that give
each store its unique appeal. I think the same applies to Web sites.

If you accept the premise that
visitors will tolerate a degree of inconsistency between Web sites in order for
those sites to have individual identities, then the fact that some elements
don’t receive the same treatment on most Web sites doesn’t necessarily make
those elements confusing. In fact, it’s quite possible that multiple treatments
exist for a given design element (e.g., the location of the main navigation
scheme), and visitors can recognize and use any one of those treatments without
difficulty, even though none of the solutions meet his statistical test for a
design standard or convention.

The real test of Web site usability
is how well visitors can find what they’re looking for on the site. A site design
might conform to a set of design standards and conventions derived from common
practices on other sites and still not create a good user experience because
the site organization isn’t obvious to the visitor, or because it just isn’t
visually appealing. Another site might break from the standards in many areas
yet still be highly usable if it presents the visitor with clear and obvious
organization and choices.

Be aware of the emerging and
evolving standard practices for treatment of common design elements, and how
those standards influence user expectations. However, I don’t believe that
effective Web site design is likely to be the result of following a codified
set of design standards. On the contrary, great design intelligently and
selectively diverges from the standards in a continuing effort to find even
better solutions to design challenges. That’s a process that is likely to
continue to defy statistical analysis. But then, that’s why design is an art
rather than a science.