Doing Research

Researching user interfaces can be done in order to find bugs in an existing interface, to test a prototype, or to find answers to existing design problems. In any of these situations, research is a highly beneficial activity, even if it is small and quickly conducted.

Though in-depth research is obviously desirable, valuble insights into the usability of particular pieces of software can be generated without lengthy or sophisticated research. There are several ways to do research in a quick and lightweight manner.

Heuristic Evaluation

This is possibly one of the quickest approaches to usability research and can be extremely effective. See the heuristic evaluation page for details.

Comparative Analysis

It is a good idea to investigate how other interfaces have tackled the same design problem that you are facing. Find out other software which has tackled the same problem, gather screenshots and notes on them, and try them out if possible. Analyse the relative merits of each of these interfaces. Heuristic analysis is an efficient approach here. Conductive user testing on them is another possibility.

Literature Reviews

General usability literature can be a valuable resource in approaching your research problem (see this and this page for some recommendations). Finding examples of existing research into your design problem can also be productive.

User Testing

Guerilla User Testing

User testing involves observing someone using a piece of software. It can be used for a number of different purposes:

• To discover bugs in an existing piece of software
• To examine the relative merits of different design approaches
• To test designs through the testing of prototypes

The section on user testing (below) provides details on how the research approach is typically utilised. Though this kind of approach is extremely valuable, it is possible to generate useful results with quick and easy approaches to user testing. Simply observing someone using their computer or a particular piece of software is highly informative. Grab a pen and a notebook, sit back and watch. You might want to ask the user why they are doing something in a particular way, or what it is they are trying to acheive. Individual parts of traditional user testing approaches (again, see below) can be incorporated within a guerilla testing approach, including thinking aloud, post-test interviews, and predefined user tasks.

Full Blown User Testing

User testing is an excellent way of testing an interface, and provides detailed results on a range of aspects of a software system. Though styles vary, a user testing session typically involves:

• Explaining the purpose of the test and how it will work. Getting consent for what you plan to do with the results of the test.
• Getting the user to run through a series of predefined tasks. (Testing a file manager may include tasks like creating a directory called 'foo', then moving files to that directory from another directory called 'bar', for example.)
• Notes are taken on any difficulties the user encounters, as well as any unanticipated behaviour or interesting observations.
• Testing sessions are frequently recording using specialised video and screencast recording software.
• Sometimes, thinking aloud protocol is utilised.
• It is typical to conduct a post-test interview with the user to establish their opinions and feelings on the test.
• Once a number of tests have been conducted, the resulting data is analysed and a report produced listing the usability bugs that have been found in the software along with their severity.