Why Run a Tree Test?

The primary reason you want to conduct a tree test is to make sure your navigation labels make sense to actual users (and not just the design team). Users need to be able to navigate your content quickly and easily, or they aren’t likely to stick around.

By running a tree test, you can gather important insights into how users approach and interact with your product. It can help you better understand their expectations and whether the categories and labels you’re using align with those expectations. It’s a fast and focused way to catch confusing navigation, misleading labels, and other issues that may impact user experience.

Insights Gained

• Identifying dead-ends, redundancies, or gaps in navigation
• Which categories are clear, and which are confusing
• How users naturally expect info to be organized or grouped
• Whether people struggle to find answers quickly
• Find where users backtrack or give up entirely

When To Conduct

Whether you’re creating something new, or redesigning an existing product, you’ll want to test your IA early and iterate often. Think of tree tests as checkpoints – the earlier you test, the quicker you identify issues, and the easier (and cheaper) it is to make changes.

Tree testing can be done as soon as you’ve sketched out your information architecture and done a card sorting test. As long as you have a defined navigation structure, you can test it.

Not only should testing be done during the design process, but it can also be done periodically throughout the product’s lifetime. As content is added, links are changed, and your website or app evolves over time, IA will need to be updated. Even if you only revisit it at major redesigns, tree testing is a great “non-invasive” way to keep things in check.

Benefits of Tree Testing

• Quick and efficient to set up
• Works without a finished prototype
• Gives hard numbers – data you can act on
• Catches problems before they’re baked into the design

What if you skip it?

What happens if you don’t conduct a tree test? Without it, you might move forward with a navigation structure that users don’t understand. If they don’t understand it, how can they use it? By the time you realize what’s not working, you’ve already built the design around it. That’s a much more expensive fix down the line and can lead to a lot of user frustration and wasted design effort.

Steps to Conducting a Successful Tree Test

1. Define a test objective.

First, get your ducks in a row with a research plan. Outline the basics – your product’s purpose, your target audience, and what you’re wanting to learn with this tree test. Your objective should be a clear and specific statement of what you aim to accomplish.

Are you testing if people understand your top-level categories? Are you comparing two versions of your navigation? The goal you set will shape the questions and tasks your write later.

2. Outline your “tree”.

Create a text-only outline of your site’s structure. No links, no images, just the hierarchy of categories and subcategories. It will look like a “tree” similar to how a multilevel outline does.

Note: Don’t have your information architecture figured out? Go back to card sorting tests to help inform what categories should be in your navigation. You can revisit tree testing once you have a basic navigation structure.

3. Pick a testing tool.

There are many research platforms online that help with tree tests. They let you plug in your tree and tasks and track results automatically. Look for one that’s user-friendly and offers detailed analytics for analyzing test results. (I’ll compare some of the top tools available, below.)

Once you’ve chosen, input your “tree” structure and review it for accuracy and usability before moving on to the next step.

4. Write realistic tasks.

Write short, scenario-style prompts to give to test participants. Think of real actions users would take on your website or app and write them with a specific goal in mind. You can create up to 10, but 6-8 tasks is recommended to avoid participant fatigue.

Tasks can be about finding resources, identifying new categories, testing labels, alternate placements – whatever is related to your objective. Make them clear and realistic so that people make choices like they would in real life.

Note: Avoid using keywords in your tasks that match keywords in the tree! This can skew participant’s choices and throw off the accuracy of testing results.

5. Recruit test participants.

Now that your tree test is prepared, it’s time to find participants! The ideal number of people is about 30-50 per variant. As long as you choose participants that align with your target audience, this should give you a reasonable confidence rate to look for patterns of behavior.

How to Recruit Participants

When looking for test participants, there are many different ways to source them. You can pull from an existing email list, invite people on social media, or find them on recruitment panel websites.

Some tree test tools give you the option of recruiting people within their system, but keep in mind there are additional costs per-participant (plus subscription fee, depending on the tool) if you’re using their built-in user panels.

If you’re looking to self-recruit, a great way to entice potential participants is to offer an incentive of some kind, such as a gift card, or a chance to win a prize. Without an incentive or obligation to participate, people are less likely to contribute, and you’ll need to invite many more users than you require.

Selecting Participants

Be selective when choosing participants and aim for people who resemble your actual users. By doing so, the results gathered will better translate your target audience’s needs. Although the ideal number is around 30-50, a smaller group of highly relevant participants can give you a solid foundation of data to uncover insights or potential patterns.

If you’re inviting participants outside of your existing customers or users, use a screener survey with carefully worded questions to make sure you only receive responses from people who align with your target audience. Ask them general questions such as “How often have you used the following websites/apps?” with a list that includes your product and relevant competitors. If they don’t meet your criteria, the testing tool will politely stop them from taking the tree test.

6. Do a test run. (optional)

Before gathering the bulk of your data, it might be a good idea to run a few “pilot” test sessions. Since most tree tests are done remotely and unmoderated, a potential drawback is the lack of context surrounding user behavior. Conducting moderated “pilot sessions” with a few participants can be a way to mitigate this drawback. That way, you can make sure the test instructions and task phrasing are understandable and notice any nuances around participant’s choices that you might not have otherwise seen.

Note: If you’re unable to moderate pilot sessions, another way to gather additional context is to ask follow-up questions in a short survey after participants complete the test.

7. Run the test and collect data.

Finally, it’s time to conduct the test! Since unmoderated tree tests are done remotely, they’re very easy to do. Simply launch the test using your online tool and share it with participants to complete on their own time. Now just wait for the results to roll in!

8. Review test results.

Once tests are complete, you can review the results. Depending on the testing tool you use, it should track and collect data on user behavior, including their answer choices, the paths they take, success rates, time spent on each task, etc.

Analyze the data to see if any patterns emerge among participants. Look for commonalities with a focus on key metrics:

• Success rate: how many people found the right answer?
• Directness: how many found it immediately vs. having to backtrack?
• Path: were there common wrong turns or misunderstandings?
• Time: how long did tasks take for people to complete?

These questions will help you determine where improvements can be made and where to focus your efforts.

9. Make changes and test again.

Last but not least, it’s time to make adjustments! Use the insights you’ve gained from the tree test results to improve your navigation structure and IA. Tweak labels, regroup content, or try a new hierarchy to address any findability issues. Keep your main objective in mind as you make changes. Then, run another test to see if it improved things!

Comparing Popular Tools

When it comes to tree testing tools, there are many options to choose from. To know which one is right for you, you’ll want to consider factors such as how user-friendly their interface is, what helpful features they offer, what analytics they track during testing, and whether their cost will fit within your budget.

 To get a better idea of what’s available, here is a look at my favorite options – UXTweak and Lyssna.

UXTweak

UXTweak is an all-in-one UX research “suite” that supports a wide range of testing methods, recruitment options, analysis features, and moderated interviews. Their interface is user-friendly and allows quick test creation, letting you easily dive into IA validation without a heavy learning curve.

Key Features & Advantages

• Auto-import! UXTweak can automatically pull navigation information from your website instead of you having to upload a file or input it manually. (Although, those are still available options!)
• Tech Integrations: Available integrations include Figma, Google Drive, Microsoft OneDrive, and Google Tag Manager. Condens, Marvin, Notion, and Dovetail allow integration from their side and may require exporting.
• Recruitment database: UXTweak lets you manage test participants with your own research participant database. Easily recruit additional users by using their online panel, adding a widget to your website, importing an email list, or sharing test links.
• Moderated variations: Moderated studies are available without the need to integrate or leave their platform, making them an all-in-one research “suite”.
• Metrics tracked: Success rate, directness, time to completion, first click, paths, and destinations. Offers more in-depth analyses of results, presented on multiple tabs.

Cost

Whether you’re working at a startup or a mega enterprise, UXTweak offers different plans depending on your needs – Free, Business, or Custom. Each plan has different limits on the number of studies, tasks, and monthly responses allowed.

Their Free tier for individuals is a good option for students or freelancers wanting to test out all the different tools, but probably not ideal for meaningful studies. This option is limited to only 15 responses per month, does not give you access to recruit from their user panel, and you can only store up to 10 participants in your database. Tree tests specifically are limited to 10 responses and only 2 tasks. You’ll have to raise your tier to at least the business level at $125 per month to increase response caps or gain access to the user panel (which is an additional expense).

Note: UXTweak says they offer discounts on subscriptions for students and non-profits, but they don’t say how much. You have to contact them directly for info about it.

For more information, visit UXTweak’s pricing page.

Lyssna

Lyssna (previously UsabilityHub) is a clean, easy-to-learn testing platform that offers a similar variety of usability tests and research methods as UXTweak. You get access to tree testing, first-click, five-second tests, preference tests, surveys, and (in higher tiers) live interviews scheduling with calendar and video integrations.

Key Features

• Templates! Lyssna allows for quick and easy setup with a preloaded template library and read-to-use tasks.
• Tech Integrations: Available integrations center around interview scheduling and prototyping to make test moderation possible. They include Microsoft Outlook, Google Calendar, Figma, Microsoft Teams, and Zoom.
• Metrics tracked: Success rate, directness, common paths, and time to completion, with per-task breakdowns, first choice, and destination distributions.

Cost

Lyssna offers several plans depending on you and your team’s research needs – Free, Basic, Pro, and Enterprise. Just like with UXTweak, Lyssna’s plans have different limitations for each tier. As you move up, you can unlock more testing methods, participants, longer test times, and live interview capabilities.

Their Free plan is an excellent option for small teams (up to 3 people), freelancers, and even students. They allow you unlimited studies but cap test duration at a max of 2 minutes. The response cap is marginally better at 15 per test, but you do have access to their global user panel at all tier levels (for an additional fee per participant).

Note: Lyssna offers a 25% discount on their subscriptions for students and educators, as well as non-profits and charities. Visit their website to learn more.

For more information, visit Lyssna’s plan comparison page.

Similarities

Both UXTweak and Lyssna are amazing UX research tools  that offer a wide range of testing methods including tree testing. Each tool has tiered plan options to choose from depending on what your project needs and what fits within your budget. Those tiers will determine what tests are available to you, how many responses you’re allowed, and whether you have access to more advanced features. Additionally, both tools support screener surveys to filter potential respondents when seeking participants, as well as follow-up questions after tests for added context.

Note: Available testing methods depend on which tool and plan you sign up for, but every tier includes tree testing, card sorting, preference tests, five second tests, surveys, first click tests, and prototype testing. Session recordings, website testing, and live interviews are also available depending on your tier level.

Which should you pick?

Both UXTweak and Lyssna have a lot in common and are very useful as research tools, but there are subtle differences that set them apart. While UXTweak is excellent for businesses who want an all-in-one platform that does it all, not everyone can afford the monthly price point to access all of their features. Lyssna, on the other hand, is a simpler and more affordable option for those comfortable using them in conjunction with other tools.

Whichever you choose for your UX research, I’m confident you’ll be in good hands.