Hold the Mayo: Tree-testing the Mayo Clinic website

3 min read Optimal Workshop

For World Usability Day 2013’s theme of healthcare, we continue our look at the findability of  prominent healthcare websites. This time we look at the world-famous Mayo Clinic:

Just as we did for the World Health Organization (WHO), we ran a tree test of the Mayo Clinic’s website, to get a quick measure of the site’s organization and labeling.

We set up the Mayo Clinic’s site structure in Treejack, created 9 tasks that covered common scenarios that the general public might encounter, then used Amazon’s Mechanical Turk to get 80 people to do the tree test.  So, how well did the site structure perform?

Overall results

Like the WHO site in our previous study, the Mayo Clinic site did a respectable job of getting their site visitors to the right place. Almost 7 times in 10, participants got to the right answer:

It’s not all sunshine and lollipops, though. Participants had to work hard to get there. In almost half of their attempts, participants had to back up and try a different path:

Evil attractors (or, the importance of clear first-level headings)

So why did they wander around so much? Was there a pattern to their madness?

Yes, there was.

If we look at the graphs of their paths, across several unrelated tasks, we see that they repeatedly visited the same 4 first-level sections:

  • Patient Care
  • Health Information
  • Research
  • Education

If they picked the wrong section, most of those backtracked immediately (shown in blue below):

Most participants eventually found the right answer, but they had to wander around first.

This is an example of evil attractors – headings that lure users down the wrong path, across several unrelated tasks.

Why do these headings attract unwanted clicks? Because they are too general or vague. For example, think about the heading “Health Information”. Then remember that you’re on the Mayo Clinic website. Just about anything you can imagine looking for on this site could be deemed “health information”.

Or consider the task “Learn how to reduce tension in your daily life.” There are no correct answers in either Research or Education, yet some participants considered this as their research, or part of their education in medical conditions:

This is not a big deal if it only happens in one or two tasks, but when it appears over and over again, we can see our first-level headings are not getting the job done.


In terms of findability, the Mayo Clinic site shows two clear results

  • The first-level headings are not clear and distinguishable, causing a lot of backtracking.
  • The lower-level headings are performing better than their parents, allowing users to find most answers on their second try.

Have you seen other examples of evil attractors, either in Treejack studies or in usability testing? We’d love to hear about them!

This website review was developed, analyzed and written by Dave O’Brien (UX professional).

Want another website review? Read the WHO analysis.