| SIG Home | About
the SIG | SIG Activities | Resources | Topics Newsletter | Conference | Bookshelf | Toolkit |
| Designing for Usability Don't Forget the Power User |
Most usability studies focus on ease-of-learning rather than on long-run efficiency. Ease-of-learning is an appropriate goal for products that are used infrequently, like many commercial Web sites, automatic teller machines (ATMs), or Microsoft® PowerPoint. However, ease-of-learning should not be the primary goal for products like corporate accounting and purchasing software or CAD software that are used many times a day, often by "power users". For products where most users soon become experts and use the products daily, efficiency should be the primary usability attribute, with ease-of-learning a secondary attribute.
This ease-of-learning focus is due, in part, to
the emphasis on "discount usability" where a small group of users are tested in a lab setting, and
the difficulty is conducting studies over a period of time with the same users.
The bias towards ease-of-learning studies can result in products that frustrate power users of the system.
How do you decide between ease-of-learning and efficiency?
Knowledge of your users and their tasks provides a basis for determining whether your primary usability attribute should be ease-of-learning or efficiency. Consider making ease-of-learning the primary usability attribute if:
A task is done infrequently.
A user's job rating is not directly tied to the volume of the output.
There is little training on how to do the task.
There is not a lot of support.
Turnover in a company is high.
Consider making efficiency the primary attribute when:
A task is frequent (for example, it is done every day or many times a day).
A user's job rating depends on the volume of the output (for example, a user is expected to review 100 voices per week or make 100 airline reservations in a day).
The task is highly repetitive.
The audience is highly technical and well-trained.
What can you do to insure that a product is efficient for power users?
First, make sure that you have adequate information on task frequency, training, and corporate expectations about how much work is expected from each user. This information can be gathered through well-designed questionnaires, focus groups, and field studies. Field studies where a sample of users are observed and interviewed about the details of their work can provide information on task frequency and also on areas in the existing tasks that are cumbersome and inefficient.
Second, if efficiency is your primary goal for a new product or an improved version of an old product, arrange a longitudinal study where users are tested several times over a period of days or weeks. The goal here is to get the users over the initial learning hump and gain some data on the efficiency of the product design.
Third, if you have a bimodal audience where some people use the product infrequently, while others use it often, consider designing an interface that accommodates both groups. One approach would be to combine the "Wizard" technique popularized by Microsoft with a second, more efficient user interface for the heavy-duty users. New users could use the wizard to step themselves through a task while more experienced users could skip directly to a more efficient way to complete the task, like a single input form.
Fourth, make sure that your new product is at least as efficient as the product you intend to displace. A common complaint from users migrating from mainframe to GUI applications is that the GUI application is much less efficient than the original character-cell based mainframe application.
And last, consider design techniques for making the user's interaction with a product more efficient. These techniques include:
Minimizing the number of windows that the user has to navigate to complete frequent tasks.
Providing straightforward keyboard shortcuts for highly repetitive functions.
Allowing the creation of templates for tasks that re-use information frequently.
Providing a command line for very efficient keyboard input. The term "command line" has become almost an anathema for GUI applications. However, a command line is a very efficient mechanism for some tasks.
Providing easy customization of features that affect efficiency. For example, provide a feature that allows experts to turn off various levels of confirmation messages.
Avoiding the use of large graphics that slow system response.
Minimizing the amount of eye and hand movements needed to complete a task.
Avoiding unnecessary transitions between the keyboard and the mouse (don t make users take their hands off the keyboard when they are in the middle of a data entry task).
Allowing users to enter codes rather than entire text strings.
Let's not forget the power users in our quest for making products more usable.
|
|