How did you get interested in computers?

In my senior year of high school, I took a course called something like Introduction to Computers and Programming. I really, really liked it, and I was also pretty good at it. I was good at the programming of BASIC and read the textbook on computer architecture from cover to cover. The computers we used in the course were Radio Shack TRS-80s—a really old machine. The storage medium was a cassette tape.

So is what you learned in college almost 20 years ago different from what computer science majors learn today?

I don’t think it’s terribly different because the program I was in was fairly theoretical, and a lot of the foundations of computer science are still very applicable today. Things like the basics of data structures and algorithms, the underlying constructs of programming language, and computer architecture and operating systems—those are all core courses.

What has changed is the surface layer of the technology—instead of Pascal and C, people use Perl and JavaScript. I haven’t spent a lot of time keeping up-to-date on the technical side of the latest programming languages or technologies, but I’m familiar with what they can do because I’m familiar with the underlying theory behind them.

What did you do after you finished college?

After graduating, I took an entry-level technical support job for Santa Cruz Operation, the company that owns the UNIX operating system. I was helping customers solve problems, either using or developing for the UNIX/Xenix operating system. Because I was trying to help people, I started to get interested how they learn and why it’s difficult to teach them technical subjects. That led me—after five years of consulting and working—to a doctoral program at Stanford in the School of Education, which was a cognitive science program focusing on how people think, learn, and know.

Going to Stanford was a great experience because it led me to the field of human-computer interaction. A professor in the computer science department, Terry Winograd, showed me that there’s a whole field concerned with how people interact with computers. He also showed me how design as a discipline can be applied to help solve problems that people experience in dealing with complex technologies.

I had thought that the way to solve people’s problems with technology was through education. But people don’t want to learn technology! What I learned in graduate school was that technology can serve us well without requiring us to be knowledgeable about its inner workings.

When you finished your Ph.D., you went to work in the corporate world. What was that transition like?

It was wonderful because I love to apply my work in a way that has an impact on the world, and I think my last two jobs have given me that. I feel that a lot of academic endeavors tend to be motivated by things that aren’t necessarily going to have an impact on most people’s lives. I think they do help in an indirect way—doing basic research is fundamental to advance the science in various disciplines—but I personally wanted to have a direct impact.

You recently moved from Yahoo! to eBay. Why?

I was very happy at Yahoo! I loved the company, and I loved working with all the people there. It came down to impact. eBay is probably the most customer-centered company I know of. It succeeds only when its customers succeed—in the selling process or in finding what they’re looking for. This linkage between user experience and revenue puts me in a very powerful position. My team’s job is to understand what the user’s needs are, articulate them clearly, identify some of the usability problems, and make sure that the company addresses them. The company deeply wants to address these things because they know it’s related to their bottom line.

Yahoo!’s business model doesn’t work that way. You go to the Web, type in a search, and then you get these sponsored links, which are advertisements for related products that you may or may not want. If you’re on Yahoo! Mail, you might get an advertisement for a diet program, and you may not be interested in that. You can see how those revenue-generating things might be very different from your goals. Your goal might be to check your e-mail, but their goal is to tell you about a service or product. The revenue model is advertising, which is not always about meeting users’ goals. That’s why I can make a bigger impact at eBay.

How do you conduct user-experience research studies?

We conduct two different kinds of studies: lab-based usability studies, where people interact with a piece of technology that we’re studying in the lab, and field studies. In field studies we look at how people interact with technology in their environments—in their homes, in their offices—and we just go in to observe. Researchers can conduct these studies on any type of technology, from the eBay Web site to desktop computers to iPods.

In both types of studies, especially the field study, the data are a combination of self-reported information and observed behavior. Behavior and self-reported data are often very different. People often do things that they’re not really aware of. So it’s important to look at both behavioral data and self-reported data and treat them as independent views into the same experience.

What kinds of behaviors are you looking for in your studies?

We’re usually looking for things like: How long is it taking people to complete a task that we’ve provided them, such as buying or selling something at an eBay auction? What kind of emotional states are they going through? Facial expressions are useful to look at, and it’s useful to listen to their tone of voice. We look for errors that they may make, even ones that they’re not aware of. And we look for any kind of deviations from what we would consider the optimal way of doing things. For example, a buyer might bookmark an interesting listing as a way of keeping track of it, but a better way would be through the use of the “watch this item” feature, which lets them track its progress against other similarly tracked listings.

We also use what we call a think-aloud protocol: We ask them to describe what they’re thinking as they’re doing whatever we’ve asked them to do. So if we’ve asked them to, say, find a product on eBay they’re interested in and place a bid on it, we’ll ask them to tell us what they’re thinking. They might say something like, “OK, well, I’m going to go ahead and search for an MP3 player. I’d like to get an Apple iPod, but I seem to be getting so many results I’m not sure what to do now. I want to see all of the total prices at once, with shipping included, but I don’t see how to do this.” Having them describe their thinking helps us to understand their cognitive state. We also ask them about their attitudes about their experiences.

What is an example of a problem that your research has identified?

On eBay, the standard way to make a purchase is to bid on something in an auction and wait for the result. But you can also use a feature called “Buy It Now.” The way it works is that you can click on “Buy It Now,” and it will ask you, “Are you sure you want to buy it now?” And you just need to confirm; you just say, “Yes.” And then, it says, “Congratulations. You have won this auction,” because it’s an instant end to the auction. Then customers are asked to provide a form of payment and say where they want it shipped.

Our recent studies have shown that this is actually a deviation from most people’s mental model of how you purchase things on the Web: You find what you’re looking for, you specify that you want it—you add it to your cart or “buy now”—then you provide your credit card and shipping information. Finally, you have a confirmation. You’ve put all your information in and then you say, “Yes, I’m committed. I’m going to do it.” The act of pulling out your wallet and typing in your financial information is a powerful indicator to the average person who uses the Web that they’re making a commitment, and they really are buying something.

Because eBay’s “Buy It Now” feature differed from this mental model, some people who weren’t familiar with eBay were agreeing to buy things without knowing that they were really committing to buy them. They weren’t whipping out their wallet and typing in their financial information, so that’s where they got caught up. We didn’t know about this problem going into the field studies, where we observed them, but it turns out to explain a good number of “unpaid item” incidents from new buyers.

We are now working on finding and implementing a solution to this problem.

What is the biggest online usability challenge?

Passwords. There’s a tension between the need for security and the need for usability. Each site that uses passwords has different rules for what those passwords can be. Those rules have been applied because Web sites want their end-users to use passwords that will be hard for other people to guess and hard to break. It’s really a problem for people because they can’t always use the same password.

I had this one program recently that required me to use at least one upper-case character, at least one number, and at least one special character, like an exclamation point or a star. This became a new password for me, and now I’ve got to remember yet another. That causes me to have to write it down, and then it becomes insecure. So this is just a real usability nightmare.

Have you come across any solution for the password problem?

I haven’t actually focused on how to solve that problem—but it would be hard. I think it would come down to having password standards that were consistent across all the different sites. But that would make it easier for hackers to know all the basic possibilities, so I don’t know if there’s an easy solution to this.

What advice would you have for someone building a new Web site?

Put the site—an early version of it, even just print-outs of what you think the site is going to be—in front of some members of your target audience. Put them on the most important page, give them a realistic scenario that they might be in, and watch what they do. You’ll be amazed. When you watch them try to figure it out, you’ll learn all kinds of interesting things that you can directly apply to your next version. It doesn’t cost anything to do. You don’t even have to do it in a usability lab; you can do it in a conference room. It doesn’t need to be recorded; you can just have a few people watching. And you’ll come away with some insights that you didn’t have before.

I used this method for a small legal firm’s Web site, and they were shocked to learn that typical users didn’t have a clue about what they were offering. They were trying to use their Web site to describe and sell unique legal services, but the message just wasn’t getting through to their target audience. The initial site just looked too much like a typical legal firm’s site. So they had to take a more direct approach in their site design in order to meet their goals.

For someone just starting off today, what are your recommendations for how to get into the career?

There are three routes into the profession: computer science, design, and research. But I feel that the research side should come in only after people have a solid grounding in one of the other fields, maybe in graduate school or in the third or fourth year of college. Having a psychology degree, for example, could work well as long as you combine that with some kind of understanding of how the technology is built or how to apply design to meet users’ needs.

What do you think the future of this career is? Will all of the usability problems be solved in ten years?

No. The ratio of engineers to user-experience researchers is rather small. The minimum requirement to build anything is an engineer. On the other hand, some places still think of a user-interface specialist as a luxury. But as customers demand easy-to-use products, people specializing in this area will become in greater demand because they’ll be linked to better business performance. What drives everything, really, is money. I think those who aspire to influence and improve the user experience will have a great career and great prospects for the future.

Further Reading