Concepts: Usability
Testing
Topics
Usability testing evaluates the system from the perspective of the end user and
includes the following types of tests:
Usability testing is not a substitute for good design-it's most effective
when combined with User-Centered Design (see Concepts:
User-Centered Design).
Start usability testing early. Early
user testing means early prototyping, typically drawings and mockups described
as low-fidelity prototypes. The hi-fidelity prototypes follow later in the process
(see Activity: Prototype the User Interface).
One way of exposing a user-interface design is to have a business or system
analyst sit with the end user in front of the interface. Walk through a common
scenario; for example, a use case's basic flow with typical values you described
in a use-case storyboard. Encourage the person to ask questions and give comments.
The challenge with this approach is to ensure the information you obtain is
as unbiased as possible. To do this, you need to make sure the questions you
ask are context-free. Take as many notes as you can. If possible, have someone
else do this so you don't interrupt the user's natural flow. (For useful guidelines
on conducting user interviews and workshops, see Guidelines: interviews
and requirements workshops.)
Another way of exposing the user-interface design is to perform use tests.
These are often conducted as a lab or workshop with representatives from the
end-user community. In a use test, real users perform real tasks with the interface,
and the software development staff typically takes a passive, observational
role.
A lot of value can be gained from this type of usability testing, however,
there are a number of challenges that must be faced and tradeoffs that must
be made to get reliable, economic results:
- As a general rule, this approach has the most value if the end-user community
is large, varied, and has a great degree of control over selecting their software
system. In the presence of these factors, the risk of not performing use tests
increases. Often the greater the value in performing these tests, the harder
it is to gain access to, coordinate, and manage this activity with the end
user.
- It's important to identify the most common usage patterns, discounting outlier
and exceptional results, to ensure that the user-interface design decisions
are based on the needs of the majority. To do this, you need both broad and
deep sample data, which usually requires a large amount of gathering and collating
effort.
- Where end users must migrate from an existing legacy system to a new system,
they are often concerned that the new system will provide less functionality
than that provided by the old one. Unfortunately, this issue is seldom raised
directly, and is often concealed by comments such as "I want the new
system to look and feel exactly the way the existing system does".
- Where a significant change in technology is being proposed to an end-user
community, it may be necessary to provide training in the basic use of the
technology before significant value will be gained from use testing. For example,
legacy system users may have had no previous experience using a mouse or working
with a GUI.
Each project team needs to consider these challenges against the unique project
environment they are working within to arrive at the appropriate timing, method,
and approach to usability testing.
It's very important to expose the user interface to others. As the design and
implementation of the interface progresses, you expose the design to increasing
numbers of reviewers, including:
- other project members
- external usability experts
- users
To get valuable feedback, you don't always have to go through formal use tests
where real users perform real tasks. An important class of user-interface defects
are caused by the home blindness of the user-interface designer-anyone
who wasn't involved in the user-interface design should be able to identify
most of these defects.
This is an underestimated way of exposing the design. It has a very fast turnaround
time: project members are already familiar with the application and usually
they're available for a spontaneous usability session without tremendous ceremony
or formality. User-interface designers should do this continuously during the
design activity to cure their own home blindness.
A good usability expert can help to reduce development effort by pointing out
common usability flaws, and usually offers other perspectives on the user interface
based on experience. It can be valuable to involve external usability experts
early in the user-interface design work, so there's sufficient time to refactor
the design to incorporate their recommendations.
Exposing prototypes to users is generally a good use of your time. Because
access to users is often limited, it's worth getting feedback on prototypes
when the opportunity arises. Do this as often as necessary to gain the stakeholders'
approval and to correct any misinterpretation of the stakeholders' needs. This
can occur either during requirements capture or user-interface design. Wherever
possible, avoid exposing the same user to the interface more than once-the
second time, the user will be tainted by your earlier design ideas (similar
to home blindness), and, as such, the value of the activity is diminished.
Also, when you expose a software prototype to end users, be careful to set
the expectations correctly. If you don't, users may have expectations that they'll
experience the full behavior of the functioning system behind the user interface.
Further Reading
See [CON99] and [GOU88]
for information on designing for usability.
|