Jakob Nielsen's Alertbox, February 6, 2000:

Novice vs. Expert Users

It is time to take expert user performance more seriously on the Web.

Web usability has traditionally been focused on increasing ease of learning for the novice users. This makes great sense and should continue to be the main goal. Remember Jakob's Law of the Internet user experience: users spend most of their time on other sites than your own. Thus, users rarely learn enough about any given site to become true expert users.

Learnability was also the focus back in the ancient days when the field of human-computer interaction was established (1983). Classic papers like Jack Carroll's LisaLearning demonstrated that even the best commercially available personal computer was much harder to learn than claimed by Apple. Learnability was a great concern in the early 1980s for several reasons:

A reaction against the elitist attitudes of previous software which had been designed for super-geeks and nobody else. Unix was a classic example.
The emergence of new theories and methods to explain and study novice user behavior. For example, the paradox of the active user.
The emergence of commercially successful graphical user interfaces (e.g., the Macintosh) and the ensuing explosion in the number of average people using computers.

The pendulum had to swing from a focus on geeks in the 1970s to a focus on average users in the 1980s.

By the late 1980s we had a handle on how to design for the novice user. Not that all software designers followed the rules, of course :-(

It was time for the pendulum to swing again.

Much research in the late 1980s and early 1990s concerned expert user performance. Once people learn how to use a system, then what? Famous case studies looked at telephone company directory assistance user interfaces and concluded that a slightly more optimal configuration of command keys would save the American telephone companies around $10 million per year.

Transaction throughput and the support of skilled users performing complex tasks became the goals of many human-computer interaction experts.

Another classic example was the design of the "launch abort" button (and associated status tracking displays) for the mission director in the space program launch control center. This single button would cost many millions of dollars every time it was pushed. But if it were not pushed when it should have been, people would die.

Enter the Web in the early 1990s, and the pendulum really swung back to a focus on the novice user.

Web users are notoriously fickle: they take one look at a home page and leave after a few seconds if they can't figure it out. The abundance of choice and the ease of going elsewhere puts a huge premium on making it extremely easy to enter a site. There is no such thing as a training class for a website. In fact, a website with a help system is usually a failed website.

At the same time as intuitiveness became the main goal of Web design, there were also reasons to care less about the performance of experienced users. First, as mentioned above, most sites don't have very many expert users. Second, the website is not paying the user's salary. Who cares how much time a user spends on performing a task? As long as users buy, it doesn't matter whether they do so slowly. It's their nickel (or their employer's). Finally, the Internet has motivated huge numbers of less-technically-minded people to start using interactive systems (say, WebTV), and such users need even simpler systems.

Expert Performance on the Web

These are all valid reasons to continue to support novices, but the pendulum will soon start swinging a little bit in the other direction, even if it won't swing all the way back to a single-minded focus on experts:

Some websites engender sufficient loyalty that users return frequently and begin using them on a daily basis.
The Web is starting to be used for mission-critical applications, where it does matter how fast people perform.
There will likely be huge growth in Internet-based applications that are not really websites but where users perform daily tasks across the Internet. For example, online calendars and maybe even entire office suites.

Increased attention to expert performance has four implications:

It becomes even more important to design smooth navigation paths and fast-loading pages. Traditional tricks from the GUI world may be revised on the Web: for example, shortcuts for the experienced user that are invisible or downplayed for the novice user. Maybe even training wheels interfaces where the average site visitor gets a simplified design that is easy to learn and loyal users get an advanced design that is more powerful.
The need to stop using Web browsers as the platform for Internet-enabled applications, except when they are targeted purely at casual users. Frequent users will need an optimized interface that takes full advantage of the device they are using.
In-depth content and advanced information should be added to sites to provide the depth expected by experts.
Usability tests must study users over time as they develop expertise in using the site or service.

Learn More

For more on the distinction between different classes on users and how to get the right users to come in for usability studies, see my reports with 233 tips for recruiting test users.

Previous: January 23, 2000: Saying No: How to handle missing features
Next: February 20, 2000: Does the Internet Make Us Lonely?

See Also: Complete list of other Alertbox columns