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

Marketers are spending millions to develop and promote their Web sites and marketing programs, as they face their greatest opportunity (and challenge) in recent memory. In the midst of rapidly changing technologies and competitive threats, however, there is one constant: The Web puts the customer firmly in control. With this point in mind, I would like to offer several guidelines for speaking with your customers. I’ll also share insights gathered from recent Web marketing studies.

Don’t rely on secondary or syndicated research.

Research regarding Web trends (“45 percent of Web users are expected to make an online purchase this holiday season…”) can be valuable in generating ideas and guiding your company’s Web strategy. However, to successfully execute against any strategy, you simply have to speak directly with your target customers. Specifically, we recommend involving customers at three stages of the site development process (upfront, pre-launch and post-launch), as shown below.

Exhibit 1

Involve customers early in the development process.

Many marketers wait until a Web site is nearly finalized (in beta form) to speak with target visitors. This often results in wasted resources, because site features have been developed that are not valued by customers (“Bad ideas, well-executed.”). Research conducted up-front nearly always pays for itself by focusing Web development efforts. For example, a recent study with Jeep owners uncovered that they were very interested in online vehicle maintenance and servicing information -- and had limited interest in several other proposed site features. These insights helped focus resources on the Web site’s primary source of value, and away from features that were unlikely to be used.

Start with focus groups and the offline experience.

Early in the Web development process, focus group discussions are often the best forum for identifying opportunities, generating ideas, and gathering initial reactions to new concepts. To guide development, it is best to start with a discussion of the offline process that will be replicated via the Web - and to focus on identifying problems/limitations (i.e., waiting for customer service, dealing with aggressive salespeople, etc.) that can be addressed via Web technology. Share your ideas a bit later in the discussion, because once you do, customers will focus on that idea - and they will be less likely to help you identify larger issues and opportunities. While focus groups can certainly be conducted online, we have typically found that the face-to-face conversations provide more learning. Inevitably, some insight is sacrificed when you can’t see facial expressions and body language - and all communication must occur via keyboard.

Present new concepts via Web screens, advertisements and mailings.

When gathering reactions to new ideas, it is important to introduce them as they would actually appear to customers, rather than through elaborate concept statements. This realism helps keep focus group participants in a “customer’s mindset” and provides the most accurate insights as to how they will interpret and react to a new idea. For example, in a recent study for a Web start-up, we showed target customers a prototype home page, and found that they widely misinterpreted the company’s sales proposition. This finding led to changes in both the Web site and the offline marketing campaign to promote this new company. Presenting concepts this way also provides a good opportunity to assess and refine marketing materials. For example, in a recent study for The Prudential, we found that proposed direct mail pieces failed to describe a new Web site in necessary detail. As a result, people tended to form unrealistic expectations, which would have resulted in disappointment.

In usability testing, gather insights (not numbers).

Nearly all companies do some form of usability research to test the navigation and functionality of their Web sites prior to introduction. However, beta testing can easily turn into a mechanical exercise, in which large numbers of people complete narrowly defined tasks, and technicians tally up the mistakes. We have found that a more in-depth, qualitative approach (in which people discuss their expectations and reactions as they navigate) provides a better understanding of a site’s ability to meet visitors’ expectations. We have also found that observing and speaking with 30-40 target visitors is generally sufficient to uncover navigation patterns, diagnose major limitations and problems and identify potential solutions. For example, 30 in-depth interviews with doctors and patients suggested that most people came to Medical Economics’ PDR.Net site with questions regarding medications and medical conditions. These insights led to an overhaul of the site’s search engine - and the development of the “Getting Well Network” feature organized by health condition.

Understand how your site is viewed.

When testing usability, it is important to understand which features people choose to investigate and which they regularly bypass or ignore. Our firm uses PRS Eye-Tracking to document exactly how people view Web screens, to measure the attention given to specific links, and to assess alternative screen layouts. We have found that Web screens are rarely viewed thoroughly - and that the more cluttered the screen, the less likely that any individual link will be considered. We have also seen that page layout, particularly the placement of the main visual, has an enormous impact on which links and messages are considered. Our eye-tracking insights led IBM to dramatically re-design its site’s home page to ensure that it featured important products and led visitors directly to the site’s most frequently utilized features.

Speak with experienced Web users.

When considering usability testing, it may seem intuitive to focus on inexperienced Web users, on the assumption that if these “newbies” can navigate the site, it must be easy/acceptable to sophisticated visitors. However, we have repeatedly found that people new to the Web lack the context to provide meaningful feedback - and tend to blame themselves for navigational difficulties. Experienced Web users generally provide more blunt and valuable feedback, as they are less tolerant of poor Web design. They can also often provide examples and suggestions for fixing problems. In fact, several companies have recognized this fact and recruited experienced Web users to serve on customer advisory panels.

Exhibit 2

Focus on links, search engines and the registration process.

Across numerous usability studies, we have found that many navigational problems are rooted in visitors’ misinterpretation of link descriptions, which are often somewhat cryptic and/or misleading. In other words, people click expecting one thing and they find themselves somewhere else. Longer, more descriptive link descriptions nearly always have an immediate positive impact. Many companies also tend to over-engineer their search engines, and confuse or frustrate visitors with too many options. Finally, and perhaps most importantly, many sites do not give visitors a compelling reason (or the reassurances necessary) to provide their e-mail addresses and/or personal information. The chart above offers several guidelines for ensuring Web usability, based on findings from recent studies.

Regularly survey your site’s visitors…

Once a Web site is up and running, visitor surveys can be a valuable way of measuring return and guiding enhancements. Typically, these surveys involve the intercept of every 20th or 50th site visitor, with the request for their feedback in exchange for an incentive and/or entry in a sweepstakes for a larger reward. Because they generate structured feedback from a representative sample of hundreds or thousands of site visitors, these surveys are very different from “leave us your comments” site links, which typically generate comments from a very small sample of enthusiastic (or angry) visitors. Listed below are several guidelines for conducting Web-based surveys effectively.

Exhibit 3

Site-based surveys are most commonly used to gather feedback regarding the site and to develop more detailed profiles of site visitors (i.e., Are they the same people who get our catalog?). Recently, however, direct marketers have also begun to use these surveys to quantify the value of their Web sites, in terms of contributing to sales that ultimately take place offline and/or at other sites. Through site-based surveys, Random House found that nearly 60 percent of visitors to its Books@Random Web site had discovered new books there that they eventually purchased. This meant that the site was facilitating millions of dollars in sales. Further questioning also guided investment and promotional decisions, by identifying the specific site features (such as chapter excerpts and author interviews) that were helping drive sales.

…but recognize the limitations of intercept surveys, and respect your customers.

While Web-based surveys are valuable tools, they also have their limitations. Most importantly, they reflect only the attitudes and opinions of people visiting your site. They can’t tell you why other people are not choosing to visit your site

- or why some other people may have visited and never returned. To understand the wider universe of potential visitors, and how they differ from current site visitors, focus groups and e-mail surveys are recommended. Finally, it is important to remember that Web-based surveys represent an extension of your relationship with your customers. To maintain their trust, surveys should be conducted and used properly. The chart below lists several guidelines for doing so, all of which involve treating site visitors with the respect and consideration they deserve. Following these principles is a matter of common courtesy and common sense.

Exhibit 4

Planning ahead is critical

While all Web marketers acknowledge the value of speaking with their customers, few companies have made a full commitment to conducting an ongoing dialogue with their site visitors. In our experience, we have found that the major barrier to conducting research properly is frequently timing, rather than budget constraints. The exceptional companies build research into their project planning, while others always scramble to find time and resources at the last minute - and inevitably compromise the value of the research.