What we've learned so far

Editor’s note: Scott Young is president of Perception Research Services, Fort Lee, N.J.

Over the past five years, my company has been fortunate to work with many marketers and design professionals as they implement their Web marketing strategies. In some cases, we’ve used traditional methodologies (such as focus groups, surveys and in-depth personal interviews) to better understand visitors’ needs and expectations from Web sites. On other studies, we’ve used relatively new approaches (such as usability testing, eye-tracking of screen viewing patterns, and Web-based interviewing) to document and measure visitors’ satisfaction with their Web experiences.

On one level, our experience has certainly confirmed that all of us (marketers, designers and researchers) have a lot more to learn about creating excellent Web experiences for our customers/visitors. However, our research studies have revealed several fundamental insights that appear applicable to most, if not all, Web marketing and design efforts. In this article, I’ll share several of these insights and suggest their potential implications for effective Web design.

Starting at the home page

As you might expect, a good deal of our research has focused on home page content and design. In fact, as an adjunct to “traditional” usability testing, we’ve used a new version of our PRS Eye-Tracking technology to document viewing patterns and to uncover exactly what people see (and ignore) as they visit home pages.

The most important learning we’ve found is intuitive, but often ignored by Web marketers: People nearly always visit Web sites for a specific purpose. They are not inclined to browse until after they have completed their intended task (checking a stock quote, finding a product, getting a recipe, etc.). Therefore, efforts to market to people on the home page (via links and banners) are likely to be ignored. In fact, we’ve seen that most visitors spend under 10 seconds viewing home pages, typically starting in the middle of the screen (with the dominant graphic or visual element) and working their way left towards the search engine and/or navigation bar. Unfortunately, the overwhelming majority of visitors never return to the home screen, once they have left it to complete their intended task.

This finding draws into question the design strategy of most Web marketers, which is to cram as many links/messages as possible onto the home page, since this screen typically receives the most visits. When you consider the sheer number of elements on most home screens - and the limited time that people spend there - it is obvious that most links are never seen/considered. Our research suggests that Web marketers and developers would be best served by:

  • Clearly emphasizing three to five links on the home page

People typically will not take the time to read through a laundry list of 10-12 different site features. A cluttered screen is likely to discourage involvement and hinder usability. On a home page (as on package labels and print ads), less is more in terms of engaging visitors and getting your key messages across.

  • Using visual icons to draw attention to these primary links

In our studies, we’ve repeatedly found that links with accompanying visuals (such as logos/icons) are far more likely to be considered than those without a visual treatment. Predictably, however, if each link on the screen is treated this way, the effect/impact is lost.

  • Positioning key messages above the fold, but below the top banner

Our research confirmed the well-publicized fact that many people never scroll down below the initial screen viewing area. However, we’ve also found that many people have trained themselves to ignore the top banner of Web screens, on the assumption that it will contain advertising. In most cases, people begin their viewing below this top banner (in the content area) and never bother to move upward.

  • Positioning the navigation bar on the left side

Our studies suggest that the left-column navigation bar has become the standard with which most visitors are most familiar and comfortable. It is where people now naturally gravitate, rather than the top banner. Interestingly, this gravitation to the left has made visitors prone to ignoring the items in the far right-hand column of the screen.

Figure 1

The current IBM home page (Figure 1) applies many of these design principles, most notably its relative simplicity and use of a dominant visual to lead people to three to five key links. A more typical home page, such as that of Compaq (Figure 2), fails to set clear priorities and requires involvement far beyond the willingness of most visitors.

Figure 2

The moment of opportunity

In our studies, we’ve repeatedly seen that the moment of opportunity comes immediately after the person has satisfied his/her primary reason for visiting. At that moment, once the visitor has collected a recipe, placed a trade, or gathered product information, he can decide to leave or take a few minutes to explore other features of the site. Assuming that the site has served his objective, it’s also likely that he is somewhat favorably disposed towards the site or company. Unfortunately, it is at this point - occasionally referred to as the seductive moment - that most Web sites fail to seduce. Rather than communicating a compelling reason to stay, many sites leave visitors with only a navigation bar and/or a text link back to the home page. By failing to send a call to action (“Did You Know That We Also Offer…”), Betty Crocker (Figure 3) has thrown away an excellent opportunity to make its site “stickier” and move visitors beyond short, single-function visits.

Figure 3

Improving usability

Every Web researcher has his or her Top 10 list of common usability problems, and I’m no exception. However, I’ll avoid dwelling on the very obvious (such as long download times and browser incompatibility) and point out three issues that may be a bit less intuitive:

  • Link descriptions

In our studies, we’ve found that unclear or misleading link descriptions are a very consistent source of navigational problems and/or unmet expectations. Several factors appear to drive this problem, most notably the mismatch between marketer language and customer language. However, Web designers sometimes compound this problem by pushing for short (one- to two-word) descriptions or trying to be too cute/clever (i.e., “Turbo News”). In our experience, we’ve found that a variation of Murphy’s Law is in effect: Any link that can be misinterpreted will be misinterpreted, resulting in a confused, frustrated, or disappointed visitor. Using rollovers to provide more detail is valuable, but not a perfect solution, since we’ve seen that many people never bother to explore a link that does not immediately convey a clear message. Better to spend a few more words getting the terminology exactly right.

  • Search engines

The disconnect between marketer language and visitor language also dooms many visitors’ search efforts. Often, they type in “their” term, only to find no matches, because the company calls the same product something else. Typos and misspellings also sabotage many searches, which is why some sites have wisely moved to a search-by-letter approach (e.g., Type in “A” and see all the options starting with “A”). While offering a search engine is a must, we feel that the best way to enhance usability is to focus on minimizing search engine use via effective site navigation and design.

  • Sorting processes

Sites offering thousands of products or services face a difficult challenge in helping people sort and consider different options. Here, we’ve also found that less is more: Trying to match people to the ideal product in one step rarely works, because it requires the person to enter too much information. Asking a person to enter seven pieces of information gives him or her seven opportunities to make a mistake, which often results in zero matches. A better strategy is to tackle the issue in two steps: First, ask for two or three critical pieces of information (desired destination, desired activities, etc.) and provide a broad range of matches (i.e., 100 trips that match those basic criteria). Then, allow people to quickly sort those options by several key dimensions (i.e., by price, by length of vacation, etc.). Figures 4 and 5 illustrate the “Trip Finder” process on the Web site Gorptravel.com, both before and after usability research.

Figure 4


Figure 5

Promoting visitor registration

Acquiring registered users (and their e-mail addresses) is the Holy Grail for many marketing Web sites, but we’ve seen many sites that do a very poor job of encouraging people to register. Specifically, they fail to:

  • Communicate a compelling reason to register

Keep in mind that “You will not have to provide this information again” is not an end-benefit for a visitor. Coupons, savings and advance notification on new products are compelling benefits and reasons for registering. In our experience, we’ve found that news (via e-mail) can also be compelling, but only if it is tied specifically to the person’s interest area. For example, an e-mail regarding the latest news in arthritis treatment will obviously be meaningful/valuable to someone who has indicated that he has this condition. An e-mail with the day’s general health headlines (which forces the person to scan through to find relevant articles) is usually not worth sending.

  • Provide reassurances about the use of personal information

Obviously, people need to know that their e-mail addresses will not be sold and that they will not be bombarded with sales calls. However, it is best to keep these reassurances short and simple. Interestingly, we’ve consistently found that “shouting” security reassurances (“You’re safe with us”) tends to raise more concerns than it solves.

  • Ask for only the most relevant information

The more information you request, the more likely a person is to refuse to register. Most marketers only need two or three pieces of information (name, e-mail address and area of interest) to begin a meaningful dialogue with a potential customer (by sending a relevant e-mail), but they inevitably ask for 10-15 pieces of personal information and often don’t allow people to skip certain questions and answer others. As a result, people simply decide not to register. From a design standpoint, it is important to keep these benefits and reassurances in view as the person provides the information, rather than just putting them on an introductory screen. It’s also best to let people know exactly how many questions they will be asked (i.e., “question #2 of five questions”), so that they know the registration process will be relatively short. Figure 6 contains a registration screen that fails to provide visitors with the reassurance that they need.

Figure 6

Guiding content decisions

When it comes to site content, I’m skeptical of broad generalizations regarding the value of different features and applications. However, we have learned one thing (about site content) with significant implications for Web design:

  • For nearly every Web site, there are three or four main reasons why most people visit.

For example, on medical sites, we’ve seen that most people visit to educate themselves about an illness or a specific medication that their doctor has prescribed. On automotive Web sites, most people visit to find vehicle specifications (for a potential purchase) or to check maintenance and warranty information (for their current vehicle).

In terms of Web development and design, this fact suggests that if you focus on identifying, highlighting and executing these key applications, you will consistently meet the needs of 80-90 percent of site visitors. Unfortunately, many companies are in such a hurry to launch (or re-launch) their sites that they don’t take the time to speak with visitors and identify the most important site benefits. As a result, they assign equal importance (and visual prominence) to eight to 10 features and inevitably waste time and energy developing features that people don’t want or never see. This failure to prioritize is nearly always reflected in busy home pages that sacrifice usability and fail to insure consideration of the most important links.

Understanding the visitor

The need to prioritize underlies perhaps the most important truth we’ve found regarding Web marketing and design:

The “rules” of communication that apply to packaging and print advertising are also in effect on the Web: Effective communication is about making choices, setting priorities and focusing on key messages.

To make intelligent decisions, Web marketers must start with an understanding of their visitors’ expectations and objectives. With this in mind, I’d like to conclude by offering these two suggestions for using visitor research in the Web development and design process:

  • Think beyond usability testing

To some, Web site research has become synonymous with usability testing. While usability testing (observation of visitors as they navigate the site and use various features) is an important piece of the puzzle, it has two important limitations:

- It only provides information after a Web site is nearly finalized (in beta form). At this point, a tremendous amount of time/money has already been invested, timelines are inevitably tight, and marketers are generally unwilling to truly re-think their site. In short, they are approaching the research as a disaster check and looking for quick answers/adjustments, rather than ideas and insights.

- By its nature, usability testing focuses the visitor on the micro-level of site functionality (Does it work?), rather than the macro-level of the site’s value to them (Is this feature worth using?). By showing a person a functional prototype, you are automatically narrowing his/her field of vision in thinking about how a Web site can be of personal benefit.

Taken together, this means that usability research can often result in the refinement of features/functions that are not valued by customers (Bad ideas, well-executed.).

This is not an argument against doing usability testing, but rather an argument for involving the visitor earlier in the design process. In our experience, we have consistently found that upfront research (concept testing) nearly always pays for itself by focusing development efforts on the site’s primary sources of value, and away from features that are unlikely to be used.

Build research into project planning

If research is not incorporated into timelines from the beginning, it is likely to end up rushed (and compromised) at the end of a project. Ideally, we recommend speaking with visitors at three points in the Web development process:

1) concept testing early in the development process, to challenge/confirm assumptions and guide decisions regarding content and design (overall look/feel and organization);

2) usability testing at the working prototype (beta) stage, to confirm functionality and guide refinements prior to going live;

3) site-based surveys after the new (or re-launched) site has been introduced, to measure visitor satisfaction, to document the site’s impact on visitors’ attitudes/behavior and to guide further site refinements.

Too late

Following this process will insure that Web designers hear from end-users as key decisions are being made, before it is too late to change direction. And, it will document the value of effective Web design in terms of satisfying and influencing site visitors, rather than often-misleading measures of site traffic. In short, designers who involve the visitor throughout Web development are likely to be rewarded with a streamlined process, a more effective final product, and more compelling evidence of the impact/effectiveness of their work.