Seductive Design for Web Sites

We used to think it was impossible to design a web site that successfully supported both information retrieval and browsing. We now believe a site can do both—but only when designers know what their audience is interested in.

Users Have a Goal

When users go looking for information, they have a goal. They want to find out what time “Saving Private Ryan” is playing at the local theater, or whether a Palm Pilot III will meet their needs.

A successful seduction depends on both the context and the strength of the goal: users won’t be deterred until they’ve accomplished some or all of their goal. For example, someone visiting a mall to buy new clothes would shop differently if she were just browsing than if she’d arrived in a distant city to address the board of directors and discovered that the airline had lost her luggage.

In the first case, the shopper might be lured into a bookstore to look at the latest paperbacks, or into the food court for a latté. But in the second case, she probably won’t do anything else until she’s found suitable attire for the meeting. Attaining the goal is paramount.

There’s a Seducible Moment

Education has a concept called “the teachable moment,” the point when a learner is ready to learn, willing to change, and can act. For web sites, the parallel is something we call “the seducible moment.” This is the point at which designers can entice users off the path to their original goal with the lure of something else.

The seducible moment can happen only when users have completed at least part of their original quest. It’s difficult to lure users away until they’ve reached this (self-defined) point; before that, they will simply ignore distractions. We’ve seen this frequently in our web-site usability testing.

Given this finding, the concept of the seducible moment argues against placing web ads in their traditional place at the top of a page. Users seem more willing to be seduced if an ad appears after they’ve found the information they’re most interested in.

Designers of The Mining Company’s pregnancy site (www.pregnancy.miningco.com) seem to understand this. The site features a week-by-week chronicle of pregnancy and the unborn baby’s development. Each page ends with a “Suggested Reading” section that links to the Amazon.com site. One pregnant user reported that she always ignored the top-of-page ads in her eagerness to learn this week’s fetal development, but sometimes clicked on these recommended books after she’d read this material.

WhoWhere takes a similar approach. If the site can’t find the person you’re looking for, it displays an ad for a people-finding service. This timing is important: users said they wouldn’t consider paying for this service until they knew they couldn’t find the information for free.

Placing the Lure

The seducible moment doesn’t necessarily happen only at the bottom of a page. Edmund’s, which offers a wealth of information for car buyers, makes a significant portion of its profits through referrals to Auto-By-Tel. But you won’t find an Auto-By-Tel banner ad emblazoned across the home page (although there are ads below the fold on the home page). Instead, the site strategically positions most of its ads within the content pages.

For example, on Edmund’s 13-screen-long Ford Windstar page, most of the ads appear in the middle of information about the vehicle. Similarly, ads for GEICO insurance appear immediately after information about insuring this vehicle. Edmund’s apparently is betting that users who get this far are strongly interested in the vehicle and have received enough content so that they’ll allow their attention to be shifted—to the site’s sponsors.

On Amazon.com, the lures appear higher in the page than on Edmund’s, but not at the top, where a summary of a book appears. After users have had a chance to view this information, the site invites them to check out other titles buyers of this book have purchased. One user was annoyed by the placement of these recommendations—but did go back and examine the suggestions after deciding about the book he’d originally looked up.

We’ve started using the seducible moment on our own web site, where the content includes articles from previous issues of Eye For Design. Following each article is a section titled “For more usability information,” which has links to pages on the newsletter, courses, and other material. Users seem to respond positively to this strategy: they click on these links far more than the equivalent links in the left panel.

Identifying the Lure

The first challenge in creating a seducible moment is to discover what information would interest your users. It was probably easy for the developers of the pregnant mom’s site to guess; some of them may know about pregnancy firsthand.

Other cases can be trickier. For example, how do you create a seducible moment for someone who has just gotten the correct answer to his tech support question?

Sometimes it works just to ask. By getting users to volunteer what information they’d like right now we can see if our guesses are correct.

The Strength of the Goal

But even knowing the users’ interests can’t pinpoint the seducible moment; that depends partly on how strongly users feel about their original goals. The stronger the goals, the more content they require before they’ll allow themselves to be pulled away. This can vary widely among users; some may be less goal-focused and more open to other activities.

We recently saw this at an antique fair where we offered attendees $20 and a cold soda to let us watch them search for antiques on the web. After asking them what antiques they collect and how much they know about antiques in general, we asked them to decide on and search for a specific item that interested them.

People with the most experience in antiques were the least likely to stray from their original search; they were searching for something that was important to them. Interestingly, people with the least amount of antique experience were also less inclined to stray, but people with moderate experience were more easily seduced away from their goals. These users may have had less-specific goals in mind and were willing to browse other types of antiques that appeared on web pages.

It’s hard to test for seducible moments unless users have some interest in the subject area; otherwise, they’ll simply do as they’re asked. In typical user tests, where we set the tasks, users may not be consumed by their goals (in this case, they’re our goals) and may be distracted more easily. By letting users define their own tasks in these antique-fair tests, we learned something about the strength of their goals.

Seduction by Design?

There’s still a lot we don’t know about creating seductive designs, such as how to measure the strength of a user’s goal or determine the optimal placement of the lure. But we believe that seducible moments succeed best when they complement a search for information and don’t attempt to interfere with it.