predicting what users will want

Whether you're launching a new site or evolving an old one, your goal is to predict what users will want.

2 ways to learn what your users want:

• Watch what they do. Your traffic logs can tell you how people use your site: Which features are used and which are ignored?

• Ask what they think. Using surveys, focus groups, or customer interviews, you can ask users how they'd respond to a new feature or what they'd want in your site.

watching what users do The most reliable way to learn what your users want is to watch how they currently behave on your site: What features do they use? What products do they look at? What articles do they read? How long do they stay?

This is the great advantage of web sites: You can see with great precision how your site is used. (See monitoring & evolving your site, p. 240.) Few industries enjoy this luxury (which is why they rely more heavily on asking users what they want).

You can also look outside your own site for insight. Using market research tools like those from Media Metrix or Nielsen NetRatings you can study traffic patterns for other, similar sites.

asking users what they think If you're launching a new site or adding a feature unrelated to your current site your traffic logs won't tell you if it's a good idea. So in these cases, you'll want to ask users what they want. Using focus groups, user interviews, or surveys, you can ask users (or people who fit their description) what they think of your proposed idea. Do they need it? Will they use it?

Now, if you've already decided to build a particular site (a grocery-delivery service, say), it's essential to talk to users about what they need from it. (See understanding user needs, p. 60.) But if you're still deciding what to do, consumers are a little less helpful.

tools for learning what users want

• Traffic analysis tells you which parts of your site are the most popular.

• Market analysis tells you which other sites on the web are popular.

• Task analysis helps you understand how users currently accomplish specific tasks.

• Focus groups tell you how customers might respond to a particular feature or site.

• Preference rating tells you how customers would prioritize features on your site.

• Surveys let you collect user opinions.

  See tools for user research, p. 48

Focus groups and surveys can help you predict consumer response. But bear in mind: Users don't always do what they say they'll do. They may express interest in sites they'll never use or feign disinterest in those they use often.

A real-world example: When you ask city-dwellers why they like urban life, most will mention cultural outlets, like opera or ballet. But let's face it: Most of them have never stepped foot inside an opera house.

Users don't always do what they say they'll do. They may profess a deep interest in features they'll never use.

Same goes for web users, who may profess a deep interest in features they'll never use. Maybe they like to think they would use the feature. Maybe they think it's a neat idea for someone to use. Or maybe they're just being polite. Whatever the reason, users are reliably unreliable in predicting their own behavior.

lesson from the trenches: your site is not the center of the user's universe

"Do not punish people for leading their lives while they're using your site."

Peter Merholz

Your web site may mean the world to you, but it's only one small part of your users' lives. And the sooner you recognize this humbling fact, the more effective your site will be.

"It's natural to overestimate the importance of your product," says online marketing expert Hunter Madsen. "But it's not helpful. You have to remember that unless your product is a heart-lung machine which they need and for which there is no substitute your product's not as important to them as you'd like it to be. And it never will be."

In order to effectively build or promote a site, you have to "understand the role of your brand in the universe of the consumer," Madsen says. And for most sites, it's a cameo appearance.

It's helpful, then, to get the bigger picture. "The most important consideration for someone building their site is the context in which the customer will use it," said Peter Merholz, a partner with consulting firm Adaptive Path.

And the big picture is often a humbling one. You can't assume that the user understands your site or even that you have their full attention. "It's very unlikely that someone is using your system so much that they develop the understanding of it that you have," Merholz explained.

And if you're trying to teach them about your site or encourage them to 'hang out' there you probably need to adjust your thinking.

"You have to look at the web site as a part of someone's life," says Mike Kuniavsky, author of Observing the User Experience. "And you have to look at the other parts of their life in order to understand how they're going to use this one little, tiny, itsy-bitsy part, which if you're doing a really good job they'll use for an incredibly short period of time."

This is difficult to remember, because it goes against our own experience. We spend so much time thinking about our own web sites that it's inconceivable to us that users wouldn't do the same.

But they won't. And you have to accept this if you're going to effectively serve them. "Their goal is not to hang out at the web site," Kuniavsky explains. "Their goal is somewhere else. Their goal is to have a couch or buy insurance or know what's happening in Pakistan."

So you need to account for their lack of time, knowledge, and, well, interest when you design your site. "You need to make sure that interactions are obvious, that you're not using jargon, that you respect the user's time, that you allow them to do things that they need to do in five or ten minutes," Merholz said.

Similarly, you have to remember that people get distracted while they work, and it may take them longer than expected to complete tasks.

"The number of sites that have session timeouts after 20 minutes causing all the work that person has done to disappear that, I've never understood," Merholz said. "That person on the other end of the screen might have had to walk their dog, deal with the baby, answer the phone, whatever it is."

"Do not punish people for leading their lives while they're using your site."