Project Management

Working the guidelines

Web shui guru Mary Deaton offers advice on researching usability designs and what to avoid when building your site.


By Mary Deaton

Your homework assignment in our first Web Shui class was to locate Web sites similar to the one you are designing and identify common design features. Design conventions represent the dominant and successful methods of Web site planning and creation and give you a clear set of dos and don'ts for your own site.

Now you can begin interpreting the information that you've collected. For example, if 90 percent or more of the sites use a vertical navigation bar on the left side of the page, that convention suggests that you should do the same. You can learn more about Web conventions by reading "When Bad Design Elements Become the Standard" (Alertbox, November 1999).

But, you say, doing what 90 percent of your competitors have done doesn't make you very original. True, but the feeling of being instantly at ease with a familiar design will score more points with users than deciphering totally original design tactics.

Reviewing the research of others
Copycat design makes the assumption that the methods of successful sites work. However, we don't want to copy other sites unless we feel confident that what they do works well. Even Amazon.com, one of the best known and well-regarded e-commerce sites on the Web, scored only 72 percent compliance against a set of design guidelines in a large usability test conducted by the Nielsen Norman Group.

Of the sites that you've uncovered in your research, think about the conventions and what may or may not work for your site and your customers. If you don't think the conventions make sense for you, you have two choices: develop and conduct a series of usability tests that compare the convention with any alternative you want to use or look for research that validates or refutes the convention.

You'll save money and time by reviewing the research of others rather than conducting your own. Let's explore the published standards, guidelines, and pundit edicts.

Gaining insight
Design is not a hard science based on a body of facts and formulas. Testing the usability of a design is very qualitative, even if some test results get published as a set of quantitative data.

You can't generalize to all Web sites the results of one study on one site. Most Web usability studies are limited in scope, looking at very narrow issues related to a specific design, and they may be based on inappropriate test methods, depending on what the researcher wants to uncover.

It's not about numbers
I spoke with Jakob Nielsen, who is generally considered the god of Web usability, about Web standards. The usability of a Web site, he told me, cannot be determined by any one element.

In design, many different solutions can address the same problem. And most quantitative studies on Web usability are not useful to generalize. "I put more value in qualitative studies any day," said Nielsen.

It's about insight
"The purpose of usability is insight, not numbers," Nielsen said. "You need to have multiple levels of insight into usability. "

The levels of insight Nielsen would want you to consider include:
  • Minimum heuristics of human behavior
  • Principles of Web usability—very general ones that apply to all Web sites
  • Best practices for specific elements—press pages, shopping carts
  • A style guide, different for every company and very specific

Looking for heuristics
Nielsen references the minimum heuristics of human behavior, based on research results over many years and by many researchers. These results are generalized into guidelines for the design of interfaces, whether on the Web or elsewhere.

"Effective interfaces are visually apparent and forgiving, instilling in their users a sense of control. Users quickly see the breadth of their options, grasp how to achieve their goals, and do their work," says Bruce Tognazzini, who founded the Apple User Interface Group.

Checking a site's heuristics
"Tog," as Tognazinni is known, has a list of interface design heuristics at AskTog called First Principles. Nielsen also has an heuristics list, Ten Usability Heuristics.

You can test the sites you study or your own site by comparing it to the heuristics checklists and gauging how the site measures up. Does the design of the site let people behave in the way they want to behave? If not, you'll want to find a design principle to solve that problem.

Applying design principles
Web design principles can be applied to all Web sites and to all Web users. You'll encounter some that are drawn from general design principles, such as the use of white space or the optimum line length for highest readability. Other standards came about specifically for the Web environment, such as scrolling and the use of color to denote links.

We have one little bitty problem, however. Not everyone agrees on the principles of Web design. Don't panic. First, remember that good Web design is not based on personal opinion. You must read sources that form conclusions based on studying real Web users dealing with real Web sites.

Since Web Shui isn't afraid to ask for a little homework from its readers, let me suggest that you read these three books (if you haven't already): Designing Web Usability: The Practice of Simplicity by Jakob Nielsen; Web Style Guide by Patrick Lynch and Sarah Horton; and Web Site Usability: A Designer's Guide from Jared Spool and his cohorts. You'll come away with a solid grounding in basic Web design principles. Later in this column, I'll discuss the Research-Based Web Usability Guidelines, a project of the National Cancer Institute's Office of Communication. While not the definitive word on the subject, this interesting approach to design guidelines will get you thinking, and it contains citations for dozens of research studies.

Finding best practices
Best practices are applications of heuristics and principles to a specific design problem, such as shopping carts, product pages, home pages, press information, and so on. The best practices for any design problem are discovered, typically, through rigorous usability analysis and testing of a large number of sites.

E-commerce best practices
The Nielsen Norman Group recently conducted a set of studies that looked at e-commerce usability issues. (You can read summaries of the studies at the Nielsen Norman Group site, where you can also purchase the full reports.)

The study looked at catalog pages, product information pages, shopping carts, product support, and so on. A large population of Web shoppers and potential shoppers were observed using multiple sites. The study was even conducted in multiple countries and languages.

I certainly cannot spend the $750,000 that it took to conduct the research in order to challenge the results, but because I trust the researchers, I am going to put great weight on their results.

Creative Good, a Web design and consulting firm, has also developed a set of best practices for selling goods on the Web, and the company has published them as "The Dotcom Survival Guide" (you must provide your name and e-mail address to view the guide). The principal of the firm, Mark Hurst, also writes a best practices column for ZDNet.

On its site, Dack.com includes an excellent description of best practices for shopping carts. When researching best practices for other types of design problems, I suggest that you begin with Usable Web, a detailed list of Web usability resources.

Looking for style
To determine your site's design, you need to utilize Web heuristics, design principles, and best practices. Based upon all of the information that you collect by exploring these insights, you can put together a style guide for your site. The style guide defines how you will implement the page layout, the navigation, the use of color, the use and placement of graphics, and so on.

Writing a style guide
I start a project style guide by defining how the site should reflect basic design principles and best practices. Then I look for site elements that do not easily seem to fit any of the known principles and practices; I may want to test these elements on the site's target users.

The entire site team should see the first draft of the style guide. Team members won't feel invested in or trust a style guide unless it reflects their input and expertise. Coming to a team consensus can be difficult, as opinions about Web design are like navels—everyone has one. Make sure the discussion does not focus on personal preferences and opinions; instead, ensure that each point can be backed by evidence.

You'll find good resources for helping you develop a style guide in the newsletter of the Usability Special Interest Group of the Society for Technical Communication. I found two articles, "Using a Style Guide to Build Consensus" and "Guidance on Style Guides," to be particularly helpful.

Assessing usability research
The "Research-Based Web Design and Usability Guidelines" of the National Cancer Institute began as an internal project. The Institute hoped to help employees maintain consistency in the publication of hundreds of documents and multiple sites.

The team that worked on the guidelines site, however, decided that the guidelines had potential for general use and decided to make the information public.

The guidelines cover many, many issues. A unique aspect of the guidelines is the use of a rating scale for each single point. The rating scale is based upon evidence, largely published research studies that support the validity of the guideline.

Though useful in determining if there has been research on any one guideline, the scale may not help you decide which guidelines to adhere to in your particular style guide.

You need to look at the source of the research, too, and find out if the audience tested matches yours, if the goals of the tested sites are similar to yours, and if the test itself seems extensive enough or appropriate enough for you to feel confident with the results.

You also need to set priorities. The guidelines do not suggest that any point is more important than another. My experience suggests that we never have enough time or resources to follow every published guideline worth following.

Every good Web designer must learn the value and importance of compromise and prioritization. Choosing which guidelines to follow and which ones to ignore—setting priorities that fit your audience, your site goals, and your time frame—has to be based on a strong understanding of usability, not simply on the weight of evidence any one guideline has over another.

The NIC Guidelines team wants to add priority guidelines in the next version, due in 2002.

"The next set will be much more comprehensive and will include several cycles of review by 10 usability and Web design specialists to determine the relative importance of each guideline to Web design," said Sanjay Koyani of the NIC Office of Communications usability group, in response to my e-mail query to him.

"This will help designers and managers pinpoint what will have the deepest impact. How we'll use this information to lay out the next set of guidelines is being planned."

On the whole, the site is a useful tool, but I would caution you not to take it as the last word on any Web design issue.

Resolving design disputes
Web teams have several options when it comes to resolving style guide or design disputes; reaching consensus through discussion, appealing to a higher authority (read manager, not god), mud wrestling, or usability testing. I usually vote for usability testing.

Observing real users is the only way I know of to determine how they respond to a particular design. In next month's column, we will look at how to design cost-effective and easy-to-implement usability tests.

Editor's Picks

Free Newsletters, In your Inbox