12 - Analyzing results

"How many roads must a man walk down, before you can call him a man?" - Bob Dylan

This is what we've been waiting for - actual results from actual users. Now, we just need to figure out what the data means.

Luckily, the basic results from tree tests are usually easy to understand - easier, say, than a typical open card sort. And if we're using tree-testing software, the results have already been summarized and visualized, ready for us to discover how well our tree(s) performed.

Cleaning the data

Backing up raw data, removing garbage, updating correct answers

Sharing the data

Publishing via the tool, sharing logins, uploading results to the cloud

Reviewing overall results

Overall success rate, directness, speed, and overall score

Analyzing by task

Where they went, clicked first, backtracked, slowed down, gave up, etc.

Analyzing by branches

Looking for clusters of right and wrong answers

Analyzing by user group or other criteria

Filtering results using survey questions, and comparing them

Discovering evil attractors

What they are, how to spot, why they happen, how to get rid of them

Key points

Next: Chapter 13 - Communicating results