Taxonomy Testing: How to Test Your Site Map

If there’s anything a good user experience (UX) designer knows, it’s that you need to test, test, and then test some more. You’ve probably heard of —and conducted — usability testing: testing users on your current site or on a functional prototype to pinpoint the biggest issues. And if you’ve done any information architecture (IA) work, you may have even done some card sorting: asking users to sort page topics into category buckets in order to create a user-centric site map. But I rarely meet anyone who has done any taxonomy testing. Here’s a little bit more about what it is, why it’s important, and how I do it.

What is taxonomy testing?

Some people call it taxonomy testing. Others refer to it as site map testing, tree testinginformation architecture validation, or a variety of other names. Whatever you want to call it, it’s essentially a UX technique that tests how your website is organized (your site map). It’s conducted on a simple, text-based visualization of your site structure, without any navigational aids or design elements. The purpose is to determine whether you’ve categorized all the topics on your website correctly — defined by how findable each topic is to the average user.

Why is taxonomy testing important?

You might say, “Oh we’re doing prototype testing later, so we don’t need to waste time and money on taxonomy testing right now.” Or maybe you’re planning to do card sorting (or have already done it). However, there are some great reasons to consider taxonomy testing in addition to anything else you’re doing.

Taxonomy testing tells you what regular usability testing can’t, because it eliminates visual distractions.

If you skip taxonomy testing and wait until prototype testing or live website testing, how will you know if your users are having issues because of a faulty layout or design — or because pages aren’t located where they would expect to find them? If there’s anything we should have all learned in 7th grade science class, it’s to test one variable at a time. Taxonomy testing allows you test how topics are organized — independently of everything else.

Taxonomy testing allows you to validate your site architecture before you get into design.

It’s important to test and validate during every step of a big project like a website redesign. Once you’re done with taxonomy testing and have a final site map, you can move into wireframing and design with confidence, because you know there are no information architecture issues you’ll have to compensate for.

Taxonomy testing is a great replacement, validation, or extension for card sorting.

If you’ve already conducted a card sorting exercise, how do you validate it? Waiting until the the site is live will cause the same issues I’ve listed above — while running a taxonomy test is a quick, affordable way to do it now. Or perhaps you didn’t do any card sorting, because you didn’t think your website required it. And simpler websites often don’t. No one will argue that your products should go under the “Products” tab and your case studies under the “Case Studies” tab — and so on, and so forth. However, sometimes we IA specialists still run into those pesky few pages that we’re just not quite sure where to put. Running an entire card sorting exercise just to place those few pages in the correct place is actually much less effective than doing taxonomy testing. Or sometimes even after card sorting, we’re still not sure where to put them, because there was just too much other data. Taxonomy testing is purpose-built for that exact scenario: seeing where users expect to find those few specific topics on your website that you’re just not sure where to put.

Taxonomy testing software can show you user paths and help with cross-linking.

I’ll get into this more below, but with the right taxonomy testing software, you can see where testers may have tried to find a topic first, before they eventually found it elsewhere. In comparison, with digital card sorting, you can’t scientifically track the entire process someone went through before eventually placing things in the final categories. And this makes taxonomy testing an invaluable tool for making cross-linking decisions on your website.

How do you do taxonomy testing?

There are plenty of helpful tools out there if you just do some exploring. I prefer to use Treejack. It allows you to create/upload your site map, build a set of tasks asking users to find specific items in that site map, and then send out a link to everyone you’d like to participate in the test. Treejack then provides you with a comprehensive report to help you decide whether you need to make any changes.

To give you a better idea of how it works, here are the instructions Treejack provides to users before they begin a test:


Treejack instructions

Treejack user instructions


Users are then asked a series of questions, such as this example question I asked during a taxonomy test conducted for The Denver Center for the Performing Arts (DCPA):
DSCPA taxonomy question

A sample question (or task) in Treejack


The final report then shows how successful users were at completing these tasks, as well as whether they were able to find the correct answers without having to backtrack:


Treejack report

The final report from Treejack with overall results


You can also drill down into each specific task. For example, for this question — which has 2 possible correct destination points — you can see 18 people found a correct answer directly, 7 people found it indirectly (had to backtrack), and 2 people failed.


Treejack individual question

Sample Treejack results for an individual task

Or you can view a breakdown of your results in a comprehensive table that displays where all users ended up (the site map is along the left, and the task number is along the top). If any boxes are marked in red, you should probably explore why that is and whether it necessitates a change:
Treejack participant destinations

An overview of all destinations chosen during a Treejack test


If you want to dig even deeper, you can also explore the pietree for each question, which illustrates the exact paths your users took. This is the part I mentioned was invaluable for cross-linking earlier in this post. For example, for the question below, you can see most people found the correct path (green) and reached the correct destination point (yellow). However, some people also went down the wrong path (red) and had to backtrack (blue). A very large path marked in red/blue here would suggest a need to either relocate an entire section of the website, or the need to create visual cues that cross-link to the same page from multiple places in the design phase.
Treejack pathing

A Treejack pietree that illustrates the exact path testers took for a specific task


As you can see, not only is a tool like Treejack helpful when it comes to validating your site map, but it also helps point you in the right direction for any changes you may need to make. If you delve into the results, for example, you might find out that Page G really belongs under Category 3 instead of Category 1, because that’s where a majority of users tried to find it. Or maybe Page G should stay under Category 1, but enough people looked for it under Category 3 that you need a big callout on the Category 3 pages directing you to Page G.

And that’s my case for taxonomy testing.

0 replies

Leave a Reply

Want to join the discussion?
Feel free to contribute!

Leave a Reply

Your email address will not be published. Required fields are marked *