Many people use card sorting tasks to evaluate information architecture and hierarchies; we also use Treejack to do this as a way of ‘reverse card sorting’. Card sorting tells us how users would group topics, and where they would expect to find them, whereas, Treejack allows you to test this grouping on a range of users to investigate how findable things actually are to users.

How can treejack help improve your navigation/when is a good time to use it?:

  1. To sense check your site’s current navigation; are there clear routes for users to find what they are looking for? You can identify user roadblocks by looking at whether they chose the right route, how quickly and whether they backtracked.
  2. You have investigated your current nav and want to make changes to it to improve usability. You can use Treejack to retest the current nav with the new layout to see if it has the desired effect.
  3. You are building a new website and need to ensure the labelling and routes are usable, Treejack allows you to test it on here before the new site goes live.


  • Optimalsort have a speedy external user recruitment process
  • You can get user data on speed of task completion, task success and routes the users took for each task
  • Optimal sort gathers the data and visualises it in a range of tables, pie charts and pie trees for you, saving you time doing this yourself (see examples below)

Treejack data Treejack data 3

Treejack data 2

You can give users up to 10 tasks on your navigation tree. It is limited to 10 as more than this would encourage user abandonment as it becomes very time consuming. This also limits the amount of data you will need to analyse, making the analysis more manageable. You can customise your tree to be presented with your company’s logo and colour.


  • If you want to test two separate trees against one another to see which one is better, you must do this by setting up two separate Treejack tests, and merge the data yourself to compare the two trees. OptimalSort do not have a function to compare two sets of results, meaning analysis can be time consuming.
  • Recruitment is expensive, on top of the monthly or one off testing fees.
  • Keep an eye on participant quality… some users rush tasks on purpose to complete the research as quickly as possible, creating inaccurate data. However,
  • OptimalSort will replace any low quality participants for free if you get in touch with them and provide them with the details.

If you are keen to gather insight on how easy it is for users to understand and progress through your navigation, we would recommend investing in this tool. It is a great way to identify issues with your existing navigation, and will help you find out what navigation structures are more usable if you are considering changing the navigation on your website.

How to set up your TreeJack test

1. Build your tree

Enter your tree labels in the same order and hierarchy as they appear in your current or planned navigation.

2. Enter your user tasks and define the task answers

Aim to have a variety of tasks that test different parts of your navigation, and try to use language in your tasks that don’t give too much away about your navigation before task completion. You can have multiple correct answers, so if your navigation has duplications, you can account for that. See the example for T4S below.

3. Optional:

You can customise the messages that the users will see at the start and end of the study, and their instructions. OptimalWorkshop has standardised messages already entered so you can use theirs as default!

4. Set up your questions

a. The screening question/s will be shown to all users, whether they are recruited by OptimalWorkshop, or by yourself. These are important to ensure you only recruit users who are relevant to your research, although being over specific can cause recruitment delays and raise recruitment costs. For instance, for T4S we recruited users who were in the market for beds, and for Tiger Sheds, only recruited users who were in the market for sheds.

b. Pre & post study questions. You can set up pre-task questions. These can be to find out more about the user and what they know about your product. You can also set up follow up questions to ask the user how they got on with the tasks, which can be useful to help identify which tasks were most difficult and as such the categories most in need of optimisation. OptimalWorkshop gives you a variety of question formats; single or multi line text box, Radio buttons, Dropdown select, checkbox select, likert items or net promoter score.

Optimal workshop screenshot

Examples of research questions are below:


Pre questions


Post questions Post questions 2

5. Optional

Personalise the appearance of the tree.

Personalise tree

6. Recruit

Recruiting your own users

a. Launch the test (CTA in top right corner). b. Copy the link in the ‘recruit’ section under ‘Your study address’ and send this to your participants.

Recruiting through OptimalWorkshop

a. Launch your test (CTA in top right corner). b. Select ‘Get a quote’ and complete the form. It will estimate the cost of recruitment for you. If you are happy with the cost, select the ‘Okay’ CTA and complete the purchase. c. Wait for your participants to complete the study.

Time4sleep Participant recruitment calculator

Here at Epiphany, we believe user research is fundamental when considering restructuring your navigation. We have used it to evaluate navigation structures, old and new for Time4Sleep, TigerSheds and Glide so far this year.

Happy researching!