Contents4
Sufficient consensus
Usability testing observes people performing tasks, recording completion, time, errors and comments. Five participants surface most interface problems, and think-aloud protocols explain them.
It is fast, cheap and reliably useful.
What the result pages leave out
These belong to the subject. People ask about them. They are missing from the agreed coverage.
- Task selection. The tasks are written by the team, which means the test can only find problems the team already imagined.
- Content is rarely tested. Sessions examine finding and clicking. Whether the text answered the question is often left out of the script.
- The participant's real situation. People in a test are helpful, unhurried and unworried about money.
- What they did not ask. A tester who does not know an option exists cannot miss it, and the report will read as success.
- After the task. Nothing in the standard protocol asks whether the person could act on what they read a week later.
What people actually want to know
- Did they finish the task, and did they get what they came for?
- Which questions did they have that we never wrote a task for?
- What would they have done if this were their own money and their own deadline?
- What did they believe when they left?
Test the answer, not only the interface. The button is the cheap part to fix.