Interview with Web Usability Guru,
Jakob Nielsen

In this article, we’ll be focusing on web usability and more specifically, on the views of world renowned usability expert, Jakob Nielsen.

He’s been called “the guru of Web page usability” by the New York Times and “the king of usability” by Internet Magazine.

Through his Alertbox newsletter and useit.com website, he has been educating hundreds of thousands of web designers around the world since 1995.

While his views can be controversial, especially for web designers, he remains the top leader in the usability field.

I recently interviewed Jakob Nielsen exclusively for WDD and asked him a few questions that should be relevant to all web designers interested in creating user friendly websites.

Can you please tell us a bit more about yourself and how you got started in this field?

I have worked in the usability field since 1983: my first projects were with text-only UIs on mainframe computers.

I then proceeded to mainly work on graphical user interfaces. For example my students and I did a lot of studies of early Macintosh software which was not always as great as people have idolized it to be.

This early experience came in handy later, because the first ten years of Web applications were remarkably similar to the old IBM 3270 mainframe applications in their interaction style.

In general, it’s very useful for a usability specialist to have experience with multiple generations of computers, because that allows you to identify bigger trends in human behavior and not be seduced by the latest fads.

The first decade of my career was focused on two problems: how to get usability methods more widely used, using “discount usability”, and how to improve the usability of online information.

As a result, I wrote one of the first books on hypertext in 1989 (published 1990), and a widely adopted textbook on usability engineering in software projects.

In 1994 I started doing Web usability projects which happily fused these two interests into one topic. I find it quite amusing that in the early days of Web usability, critics complained that you couldn’t apply usability methods to websites because they only work for software applications.

In contrast, in recent years, the enemies of usability have started to claim that usability is so focused on websites that the findings don’t transfer to applications, AJAX, and such. Some people will take any excuse to ignore their customers.

Of course, the reality is that usability applies to anything that has a user interface, whether website, application, mobile phone, camcorder, or anything else. The specific guidelines will differ, but the broad principles are all dictated by the psychology of the human mind, which has been steady for 10,000 years.

With the widespread use of broadband these days, do we still need to consider page weight and loading speed?

Yes, but the restrictions are certainly not as tight as they were in the days of 28.8 kbps dial-up.

The response time guidelines remain the same as always, because they are set by the way people are wired, not the way the Internet is wired. So the findings from, say, testing pilots in World War II are still valid.

One of the main guidelines is to show the next state (e.g., the next page) with one second of the user’s action (e.g., click) in order for users to experience the feeling of a freely-flowing interaction, as opposed to a sensation of delays. In one second, you can download about a megabyte over a typical American broadband connection (and much more in Asia) if you have full throughput.

The main problem for response times today is not download delays, but rather server delays, as people stick too many widgets and dynamic objects on their pages.

Remember: 1.0 sec. response time, or users won’t feel that they’re navigating freely. Also remember that direct-manipulation options, such as within-page AJAX controls require 0.1 sec. response times to avoid feeling sluggish.

In your opinion, what is the best way to test the usability of a website?

Follow the 3 basic rules: get representative customers, ask them to perform realistic tasks, and shut up and let them do the talking.

You only need 5 users to uncover enough usability insights to keep you busy for months. Even though there are only 3 rules, they are routinely violated in many studies.

For example, it’s wrong to test with your friends or colleagues. You need to bring in external users who are representative of the target audience and who don’t know anything about your project. And you can’t just let them fool around: they have to do real tasks. And, of course, you have to keep from biasing their behavior and giving them hints about how to use the site.

That’s why the “shut up” rule is so important. Of course, it’s best to have a big multidisciplinary team with dedicated usability specialists for running the studies, but small teams should still do testing.

It’s cheap, and as long as they stick to the basic methodology, designers can definitely run their own usability studies.

How can one test the usability of websites on mobile devices?

The basic rules are the same as for any studies. There’s a 4th rule, which is to run the test on representative equipment.

For a desktop study, this means using a Mac or PC, and it doesn’t matter much which one you pick. Our biggest decision is which screen resolution to use. For the last several years, we main tested at 1024×768, but we’ve now moved up one screen size for most studies.

For mobile, it’s harder to use “representative” equipment, because phones differ so much more than computers do. In our mobile studies, we test sites on all 3 main classes of mobile devices: “feature phones” (the telecoms industry’s paradoxical name for low-end phones with few features), smartphones (e.g., Blackberry), and touch-screen phones (e.g., iPhone).

We recruit a range of users and then test each user with his or her own phone, which they bring to the study. Sadly, this means that we need to test more users in a mobile study than in a desktop study, because the usability issues are very different for each class of phone.

Ideally, I recommend that sites design 3 different mobile versions, because of these differences. I realize that this is only possible for the richest sites. For everybody else, I hope that they will at least produce a separate mobile version with a mobile-optimized design, because usability does suffer when using desktop-optimized sites on a phone, even when this is technically possible.

The original philosophy of the Web was to emphasize cross-platform design, so that a single site can be used everywhere. But this doesn’t work from a usability perspective, even when one can code the material so that it will display on phones.

Either the site will be too scaled-back for a desktop user or it will be too complex for a mobile user. The two usage scenarios are so different that they require different designs.

If we wish to conduct an affordable usability test, what would be the best way to do this?

The only place you shouldn’t skimp is on recruiting representative users, because if you test the wrong people, you’re testing whether the design works for somebody who won’t actually be using it (or who know too much to be stumped by usability problems, in the case of testing people from within your own company).

Everything else is negotiable and can be done on the cheap. I already said that you can run the study yourself, so that’s “free”, except obviously from the cost of your time, but it only takes a few hours to test the recommended 5 users, and you can actually get away with testing 3 if you’re really pressed for time.

You don’t need any equipment, video cameras, one-way mirrors, or analysis software. You don’t even need a computer, if you’re testing a paper prototype.

Otherwise, a laptop or any other available computer will do, and you can run the test in a small conference room or even a regular office.

You do have to close the door, though, to avoid disrupting the user and to safeguard their anonymity, so you can’t test in a cubicle. Just tape a note to the door saying “Usability Test In Progress: Do Not Disturb”. (And remember to take it down between sessions, or people will stop respecting the sign.)

As far as website and blog navigation goes, is breadcrumb navigation ‘dead’?

No, we frequently see users access the breadcrumbs in testing, either to check where they are in a site or to navigate to a higher level.

So breadcrumbs are definitely useful. Just as important, they don’t harm those users who don’t use them. Some studies have found that many users don’t use breadcrumbs.

But that’s OK, because the breadcrumbs don’t cause any trouble for these users, and since they’re a very lightweight design element, breadcrumbs are worth including for the substantial good they offer to those users who do use them.

For web designers, is it ok to break the rules of usability when creating artisitic portfolio websites and blogs?

Yes. First, the definition of art vs. design allows you to do anything in an art project, because it doesn’t serve a utilitarian purpose.

Even though there certainly would be a business purpose in something like a portfolio site, the standard usability guidelines still wouldn’t be as critical, for two reasons:

First, the target audience would be people with vastly superior Web skills (other designers, Internet managers, and the like). And second, people typically don’t do much when visiting a portfolio other than browse it and admire it.

Thus, they won’t be as dependent on easily-accessed features as users of, say, a home banking site where it would be a disaster if people transferred money to the wrong account.

Amazon.com is regarded as one of the top e-commerce websites. What makes it so successful and do you see any usability mistakes on their site?

Amazon is a great case study in the difference between total user experience and the on-screen user interface.

They owe their success to a lot of off-screen aspects of the total user experience, including comprehensive product selection, informative confirmation emails, and rock-solid fulfillment. They also have reasonably good prices, though never the absolute lowest, which proves that it does work to compete on the quality of the user experience and not just on price.

The screen-design is also good in terms of rich product information, including helpful customer reviews. Amazon was one of the first companies to recognize that it’s OK to include negative reviews: this increases credibility and people will just buy something else instead, so they don’t lose the order, even if they lose that particular sale.

All this said, Amazon is not a good model for other sites, because the pages are overwhelmingly complex with much too many features, many of which don’t help users in considering the current product.

Amazon can get away with this complexity because most users are familiar with its design because they shop there so often. But a first-time user would be baffled. Since most sites don’t have people who shop there as much as they do on Amazon, most sites need a simpler design.

Amazon is also not good at helping shoppers understand a product area. Because it’s such a general store (selling everything) and because of its origin as a bookstore (where there’s really no such thing as a product space; only individual books and authors) Amazon is great at telling people about individual products, but terrible at teaching people how they should think about a product category.

This is the great opportunity for specialized sites: they can educate users about their specialty and offer tools that are optimized for the characteristics of that particular product space.

Should usability be the same for every website, or should it be ‘customized’ based on the target audience (e.g – a technology website vs. a news website)?

Usability is always relative to two things: who are the users, and what are they trying to accomplish with the UI?

That’s why we can’t just have one recommended design and just replace the logo to create a new site.

So, for example, if people are trying to just deal with a small number of things, you could simply list them all.

But if the task required users to consider a large number of options, you would need features to find, select, winnow, and sort the options, plus maybe even some kind of visualization tool.

All of these features would be overkill for, say, a restaurant group with 3 restaurants, but they’d be needed for McDonald’s location finder, which would also need a language selector and other internationalization features.

Similarly, people who are highly skilled in a domain would need a different design than less-knowledgeable users. A classic example is medical information: to maximize usability, you need different designs (following different guidelines) for doctors and for patients.

Most websites these days overload their pages with loads of information, news excerpts, Twitter and RSS feeds. Can heavy content pages still be usable?

Yes, but. The big “but” here is definitely that it is much harder to ensure usability the more features you cram onto a page.

Simplicity is usually the better choice. But if you’re in a situation where your users do demand lots of features, then you need to polish the design through many rounds of iterative usability testing.

You must work harder to solve this more difficult problem, and it’s much more risky to release something complex that hasn’t been tested with users than it is to release something simple.

Exclusive interview for WDD by Walter Apai.

Do you agree with Jakob’s principles? Should most usability rules be enforced all the time? Share your points of view below.

WDD Staff

WDD staff are proud to be able to bring you this daily blog about web design and development. If there's something you think we should be talking about let us know @DesignerDepot.