Tree testing guide
Master tree testing with our comprehensive guide. From benefits to overcoming challenges, learn how to optimize your site's navigation effortlessly.
Tree testing guide
A brief introduction to tree testing
Information architecture (IA) as a subject is decades old. For example, the Database for Advances in Information Systems published a paper on building and implementing IA within an organization in 1989, and people were practicing it without putting a name to it for hundreds of years (the Dewey Decimal System in libraries, anyone?).
It wasn't until the internet became a household tool that digital IA was relevant to the masses. The book Information Architecture for the World Wide Web, originally published in 1998 by Peter Morville and Louis Rosenfeld, started helping website owners better organize their information hierarchy for public consumption.
Of course, the internet has changed massively since the days of HTML-only sites, and what's possible now for a 2D screen is pretty much only limited by imagination. But with so many design possibilities, how do you know if your website visitors can navigate effectively to find the information they’re looking for?
The answer lies in UX research – specifically, tree testing. You can use this method remotely or in-person to test how easy it is to find information on your website. Using a simplified version of your site's navigation, you ask participants to find specific information within the tree.
The role of tree testing in optimizing IA
Tree testing helps you get quantitative and qualitative feedback on the IA of your website or app at a foundational level. It’s ideal for validating navigation when developing a new product, for whole-site migrations, or when adding new content pages to your site.
So, how does it optimize IA?
Tree testing helps make sure that the structure and organization of your information is intuitive and user-friendly.
Deciding where information belongs might be intuitive to you when you’ve been working on a product design for a long time, but this doesn't mean that everyone else will find the navigation options clear.
Tree testing allows you to take a step back and see how users behave on your site or app and adjust the labels or categories based on your findings.
Users aren't supposed to notice IA – when it's great, it's intuitive for them, and they don't question it. If users do notice the IA, it's likely because they got frustrated at not being able to find what they were looking for. Tree testing can help you find these flaws before you spend time and resources on development.
What is tree testing?
In the book Information Architecture for the World Wide Web, the authors have multiple definitions of information architecture (IA), although the first one is pertinent to user experience:
The combination of organization, labeling, and navigation schemes within an information system.
Using this definition of IA, we can infer a formal definition of tree testing, which we'd suggest is:
Tree testing is a research method based on information architecture, where you present users with a hierarchical category structure, or tree, to evaluate information navigability and findability.
This definition helps explain the roots of tree testing (excuse the pun), what it entails, and what you'd use it for.
Is tree testing qualitative or quantitative?
Tree testing can be both qualitative and quantitative, depending on your research objectives and the stage of research.
In formative research, the primary focus is on understanding user behavior, preferences, and thought processes. Qualitative tree testing involves observing and analyzing participants' navigation paths, their choices, and any comments or feedback they provide in response to open-ended follow-up questions. This approach helps you identify usability issues, user expectations, and areas for improvement within the IA.
The aim of summative research is to objectively measure the effectiveness and efficiency of the IA. By conducting quantitative tree testing, you can gather numerical data on task success rates, completion times, and directness. This approach offers statistical analysis of the navigation structure’s usability and allows you to compare different designs or iterations.
By combining both qualitative and quantitative approaches, you can gain a comprehensive understanding of users' navigation behavior, uncover usability issues, and make informed decisions to optimize the IA for a better user experience.
What is tree testing used for?
Tree testing is used to evaluate the IA and navigation of a website or app. It helps identify any issues or inefficiencies in the organization and structure of the content.
Here are some common use cases for tree testing:
Evaluating website navigation: Tree testing helps assess the clarity and intuitiveness of a website's navigation structure. It can identify issues such as confusing labels, redundant categories, or missing information.
Testing a new IA: Before implementing a new navigation structure, tree testing can be used to validate its effectiveness. It allows you to gather feedback and make informed decisions about proposed changes.
Assessing content organization: Tree testing helps you work out if content is logically organized and easily discoverable. It can reveal whether users can find information quickly or if they encounter difficulties due to poor categorization or labeling.
Comparing different navigation options: If you’re considering multiple navigation options, tree testing can help compare their effectiveness. By testing different tree structures, you can identify the most efficient and user-friendly option.
Iterative improvements: Tree testing can be used iteratively to refine IA over time. By conducting multiple rounds of testing, you can track improvements and make sure that the changes you make are effective.
What happens in a tree test?
In a tree test, you present participants with a hierarchical structure, similar to a site map, representing the organization of content on your website or app. This structure consists of categories and subcategories. You then ask participants to complete a specific task, such as finding particular information, by navigating through this structure.
You can then analyze how participants navigate through the tree, noting which path they took and how long it took them to find the desired information. You can use this to evaluate the effectiveness of the IA and navigation labels.
Tree tests can be run in-person or remotely. In-person moderated tests are conducted face-to-face with participants, usually in a controlled environment such as a usability lab. You can directly observe participants’ behavior and provide assistance or clarification as needed. While this method offers control over the testing environment, it’s less convenient and more resource-intensive than remote testing.
You can run a remote unmoderated test using a platform like Lyssna. Participants can access the test from their own devices and complete it independently, following the instructions you provide. This method offers the advantage of scalability and convenience, and allows you to recruit a larger and more diverse pool of participants.
Pairing tree testing with other UX research methods
Here are some other key research methods you can use in combination with tree testing to help figure out the navigation of your website or app.
Card sorting
Before tree testing, you can use card sorting to understand how your users categorize and label information. This method is ideal if you're unsure about the language your audience uses and you want to understand how they group information.
There are three main card sorting methods. In an open card sort, you ask participants to group and label cards without predefined categories. They’re free to create their own categories and group cards in a way that makes sense to them.
In closed card sorting, you ask participants to sort labeled cards into predefined categories. Hybrid card sorting combines elements of closed and open card sorting. You give participants a set of cards and a set of predefined categories, but they can also create their own categories.
Card sorting helps you understand how users expect to find content on your website, and can be used to inform your initial IA. You can then use tree testing to highlight any category or labeling issues. Using card sorting and tree testing together ensures that your site’s IA aligns with user expectations and is validated by their actual behavior.
Check out these templates to get started with card sorting:
Improve your information architecture (closed card sorting)
Create an intuitive information architecture (open card sorting)
Optimize IA with card sorting (open card sorting)
Preference testing
Preference testing involves presenting participants with different design variations, features, or options to gather feedback and determine their preferences and priorities.
To integrate tree testing with preference testing, you can first use tree testing to refine the structure and organization of your content based on user behavior and preferences. Once you have a solid IA, you can then run a preference test to gather feedback on specific design elements, features, or content options within the context of your refined IA.
This combined approach guarantees that your final product not only aligns with user expectations regarding navigation but also meets their preferences in terms of design and content choices.
Check out these templates to get started with preference testing:
First click testing
First click testing is used to measure the usability of a website, app, or design by finding out how easy it is to complete a given task. You can use it to evaluate the initial click behavior of users when navigating through the IA.
Typically, you’d conduct tree testing first to evaluate the effectiveness of your IA. Once you’ve refined your IA based on the results, you can then conduct first click testing to assess the clarity and intuitiveness of navigation paths within the IA. This sequential approach allows you to first establish a solid foundation of your site’s structure before refining the user experience through first click testing.
Check out these templates to get started with first click testing:
Navigation testing
Navigation testing is a method used to analyze how users navigate through your website or application given a specific task or goal. The results help you hone critical user flows and improve your IA.
After conducting tree testing to refine the structure and labeling of your IA, you can use navigation testing to assess the usability of navigation elements within that IA. By combining these methods, you gain comprehensive insights into both the overall organization of your content and the ease of navigating through it, providing a user-friendly experience for your audience.
Check out these templates to get started with navigation testing:
Prototype testing
Prototype testing involves creating a prototype and testing it with real users to validate your design decisions. It’s useful for identifying problems or areas for improvement early in the discovery process, and to make sure that you’re building a product that meets user needs and expectations.
In Lyssna, you can run two types of prototype tests: Task flow and Free flow. Task flow is ideal when you're confident with your design and want to test a flow with a specific set of objectives that ends with a goal screen. Free flow is more suitable when you're in the exploratory phase of a project and are looking for initial feedback on your designs. Depending on what your goals are, both of these approaches can be used to test your IA.
Integrating tree testing with prototype testing involves using tree testing to evaluate the IA and navigation structure of your prototype before testing its functionality and interactions with users. By combining these methods, you make sure that both the underlying organization of content and the user interactions within the prototype are optimized for usability and effectiveness.
Check out these templates to get started with prototype testing:
When to use tree testing in the design process
Tree testing can be used at various stages of the design and product development process to inform decision-making and improve the user experience. Here are some specific instances where tree testing can be beneficial:
Early design stages: You can run a tree test during the initial design phase to validate and refine the proposed IA. By testing different navigation structures, you can gather feedback and make informed decisions about the organization and labeling of content.
Pre-launch: Before launching a new website or app, you can use tree testing to identify any navigation issues or content organization problems. This means you can catch and address potential usability issues before they impact users.
Redesigns or major updates: When redesigning an existing website or making significant updates to the IA, you can use tree testing to evaluate the proposed changes. It helps make sure that the new structure is an improvement over the previous one and that users can easily find the information they need.
Ongoing optimization: You can use tree testing as part of an iterative process to continuously improve IA and navigation. By conducting regular tests, you can track the impact of changes and make data-driven decisions to enhance the user experience.
Comparative analysis: If there are multiple design options or navigation structures under consideration, you can use tree testing to compare their effectiveness. You can then choose the most efficient and user-friendly option based on user feedback and task success rates.
Tree testing benefits
Improved navigation and findability
Tree testing helps you identify whether users can find specific information quickly and in as few clicks as possible. By analyzing the results, you can identify areas where users struggle or take longer than expected to locate information.
For example, in the case of a lifestyle magazine, if a significant number of users mistakenly look for “Storage & organization” under “Home improvement” instead of “Housekeeping,” it indicates a potential reorganization opportunity to improve findability.
Identifying IA issues
Tree testing helps uncover IA issues such as duplicate paths, redundant categories, or dead-end paths that prevent users from finding what they're looking for. By analyzing usability metrics such as success rates, directness, and time to completion, you can quantify the severity of these issues.
Additionally, you can gather qualitative feedback by asking open-ended questions to gain deeper insights into users' frustrations or confusion. This information can guide you in making informed decisions to refine the IA and enhance the overall user experience.
The How tree testing works chapter includes a list of example follow-up questions that you can ask when doing tree testing.
Validating site structure before development
One of the key advantages of tree testing is its ability to validate the structure of your site or app early in the design process, before investing resources in development.
By conducting tree tests, you can identify and address navigation issues and usability problems at a stage where changes are relatively easy to implement. This saves time and effort in the long run, as it prevents the need for major structural changes or redesigns later on.
If tree testing reveals multiple IA usability issues, you can prioritize these changes in subsequent iterations, providing a more user-friendly and efficient product.
How tree testing works
Step 1: Define goals for your research
Before conducting a test of any kind (not just tree testing), you need to define what your goals are. Otherwise, what are you measuring against?
For tree testing, one obvious goal would be to "validate site hierarchy based on categories defined through card sorting."
However, if your product is going through a site migration as opposed to developing a wholly new product, a better goal might be to "test if the existing site hierarchy is still optimal for users" or to "test new predefined categories," where the latter may be necessary if lots of new content has been added.
Defining your goals in this way helps shape the tasks and metrics for the test.
Step 2: Identify the target audience
Once you've defined your research goals, you'll have a better understanding of the audience to recruit for your test.
For example, suppose you're testing in the early design stages. In that case, you'll want to recruit participants who fit the marketing team's Ideal Customer Profile (ICP) to see how potential users navigate the product hierarchy.
If it’s a site redesign or migration, existing users can provide valuable insights based on their prior experience with the product.
Learn effective strategies for recruiting user research participants, including understanding your target audience, recruiting customers and non-customers, and offering incentives.
How many participants do you need for tree testing?
The number of participants you need for tree testing will depend on a few factors, such as the design stage you’re at, the complexity of the tasks, and the time and budget you have available.
However, when aiming for statistical significance and reliable numerical metrics, we recommend aiming to recruit around 40 to 60 participants, with 30 as a minimum.
Step 3: Define the tree
After determining your goals and target audience, you can define the tree structure for your test. If you’re using an existing site structure, this step is relatively straightforward as you can replicate the information.
However, if you’re testing a new product, you can design the tree structure based on the results of a card sort. Use the categories that users find most intuitive as categories in your tree and design paths accordingly.
In Lyssna, you can add parent and child nodes manually or import a tree as a CSV file. You can also download our sample tree CSV file and use it as a template.
Step 4: Write the tree testing tasks
Once you've created a tree, you can write the tasks for your participants. This will largely depend on your goals. For example, if you’re designing a travel booking site and are in the early development stages, you could run a tree test to validate and refine the proposed information architecture.
It’s important to develop tasks that can fully engage users in realistic situations they may come across on your website. Use simple, informal language to set the scene and prompt them to find a solution. This approach helps engage participants and encourages them to process information more deeply, leading to meaningful insights.
For example, instead of writing:
"Select where you’d go to book a European flight.”
You could write:
"You're dreaming of a vacation in Europe but haven't decided on a destination yet. Where would you go to explore flight options for your European adventure?"
Here are some other best practices to follow when writing tree testing tasks:
Use natural language: Write tasks in conversational, plain English to mimic real user scenarios and encourage participants to engage naturally with the test.
Keep tasks clear and concise: Make sure tasks are easy to understand and focused on a specific goal within the tree structure. Avoid ambiguity or complex instructions.
Provide context: Introduce a realistic scenario or context that participants can relate to, helping them understand the purpose of the task and how it applies to their needs.
Avoid leading language: Don’t guide participants toward specific paths or solutions in the task's wording. For example, using our travel site example, you wouldn’t say, “You're planning a weekend getaway to Paris and you need to book a hotel. Please click on the ‘Accommodation’ tab to find available hotels in the city center,” as this guides them to the correct answer. (We explore how to avoid bias in the Common tree testing challenges chapter).
Include varied tasks: Incorporate different types of tasks, such as finding specific items, exploring categories, or performing actions to assess various aspects of the tree structure's usability.
How many tasks should you include in a tree test?
The number of tasks you give your participants can vary depending on the complexity of your navigation structure and the depth of insights you're seeking.
As a general guideline, though, we recommend including no more than 10 tasks in a tree test. This allows your participants to engage with different aspects of the navigation and provides enough data for analysis without overwhelming them. It also avoids participants becoming too familiar with the tree structure during the test, which can bias the results for later tasks.
If your navigation structure is particularly complex, you may consider having fewer tasks to make sure participants can focus on each one effectively.
In Lyssna, we don’t limit the number of tree test tasks you can include in a test. To create additional tasks, simply duplicate the section and select a new correct answer for each task.
Step 5: Write your follow-up questions
Follow-up questions can provide valuable qualitative insights into participants' experiences and perceptions of the tree test, helping to identify areas for improvement and refinement in the navigation structure.
Here are some example follow-up questions to ask after conducting a tree test.
Task difficulty
A linear scale question to gauge the perceived difficulty of tasks can provide quantitative data to complement qualitative feedback. Here's an example: "On a scale of 1 to 5, where 1 is very easy and 5 is very difficult, how would you rate the difficulty of completing the task?".
Task completion feedback
Did you find the task easy or difficult to complete?
What challenges did you encounter while completing the task?
Were you able to locate the information you were looking for? If not, why?
How confident are you in the accuracy of your selection?
Overall experience
How would you rate your overall experience with the tree test?
Were there any aspects of the test that you found confusing or frustrating?
Did you feel that the instructions provided were clear and easy to understand?
Suggestions for improvement
Are there any additional categories or subcategories you'd like to see added to the navigation?
Do you have any suggestions for reorganizing the existing structure to improve findability?
Were there any labels or terminology used in the tree that you found unclear or misleading?
Comparative feedback
How does this navigation structure compare to others you've encountered on similar websites?
Are there any websites or apps that you believe have particularly effective navigation systems?
Step 6: Conduct your tree test
The final step is to create the test itself. You can use a platform like Lyssna to create and facilitate your test – we have a ready-made tree testing template you can adapt to get started quickly, and the below video shows you how to set up your test in a few simple steps.
Once you've distributed the test and received the responses, you can move on to analyzing the results. We’ll get to that in the next chapter.
Common tree testing challenges
Addressing potential biases
It's important to recognize that biases can inadvertently influence the outcomes of tree testing, just like any other form of usability testing. One common pitfall is guiding participants toward the correct answer through the wording of test tasks or questions. This can happen when tasks are overly instructional or when they include the label itself, leading participants to choose the correct option more often than they might have otherwise.
To mitigate these biases and protect the integrity of your results, you need to critically assess your test materials before distributing them to your participants. Consider the following questions:
Does the task direct users toward the correct path?
Is the label name included directly within the task?
If the answer to either of these questions is yes, you should revise the wording to minimize bias.
By crafting neutral, scenario-based tasks that don't overtly lead participants toward specific paths or options, you can enhance the validity and reliability of your tree testing outcomes. This approach helps to maintain the integrity of the testing process and makes sure that insights accurately reflect user behavior and preferences.
Dealing with inconclusive or unexpected results
Sometimes, you might get data that seems to conflict, such as a high success rate paired with a lengthy average time to completion. In these instances, it's important to delve deeper into additional metrics to get a better understanding of the user experience.
For example, in this scenario, it might indicate a low directness percentage, suggesting that users navigated through a convoluted path to reach the desired outcome. This highlights a potential usability issue that requires further investigation and refinement.
Alternatively, inconclusive or unexpected results might stem from ambiguously worded tasks. If you’re concerned a task isn’t clear, it’s worth asking another team member to review the questions to check for potential confusion.
In Lyssna, you can add comments to any section or question of a test, on both the test builder and test results pages. This is a useful way to get feedback from your team on a test before you share it with participants, or to discuss your findings.
If there are still discrepancies or the results are confusing, you can use this as evidence to recommend further qualitative testing. User interviews or think-aloud protocols offer opportunities to gain deeper insights into the challenges your participants faced during tree testing.
By supplementing quantitative findings with qualitative exploration, you can uncover underlying issues, understand your users better, and inform design enhancements.
Iterations based on user feedback
Refining your tree structure based on user feedback is a critical step in the iterative design process. Whether your test uncovers numerous minor issues or a handful of significant concerns, iterating on your initial design guarantees that your information architecture is continually evolving to meet user needs and expectations.
While iteration fosters improvement, it can also introduce challenges, particularly regarding version control and data management. To navigate these complexities effectively, it's important to maintain clear organization and separation of your tree structure versions and corresponding data. By treating each iteration as a distinct project until the analysis is complete, you can prevent confusion and streamline the comparison process across different design iterations.
By making iteration a key part of your design process, you can improve and fine-tune your tree structure step by step. This focuses on the user and lets you keep making changes to match their needs and habits.
Elevate your research practice
Join over 320,000+ marketers, designers, researchers, and product leaders who use Lyssna to make data-driven decisions.
