If we’re using an online testing tool, we’ll probably need to do some data clean-up before analyzing the results. |
Overall scores (based on factors such as success rate, directness, and speed) give us a quick idea of our tree’s effectiveness and how it compares to other trees. |
Task scores show us specifically where the tree is performing well or poorly. |
Low task success rates show us where participants got confused, disagreed about where the answer would be found, or were lured down the wrong path by a promising (but wrong) intermediate heading. |
First clicks are especially important to analyze, because they greatly affect success rates. |
Finding where people backtracked helps us find evil attractors and dead ends. |
Spotting areas of our tree that slowed participants shows us where subheadings need to be clarified. |
Look for patterns of behavior across several related tasks. |
Look for patterns of success or failure across different sections of the tree. |
We can compare the results of different user groups (or other criteria) if we run separate tests for each, or if we run a single test and filter results by a corresponding survey question. |
Next: Chapter 13 - Communicating results