How do tree testing and card sorting differ in purpose?

ProcrustesPR

Card sorting explores how content could be grouped; tree testing evaluates whether users can find it with those groupings and labels. Card sorting is useful when information architecture ideas are still taking shape, while tree testing is useful when an existing or proposed hierarchy is available. The two methods complement each other, but neither replaces the other: just because users group content together does not mean they will easily find what they need in navigation built around those groups.

ProcrustesPRProcrustesPRDoes my website need a speed score of 100?

Card sorting: understanding which content users think belongs together

Card sorting asks participants to group cards representing pages, links, or other content according to criteria that make sense to them. It is typically used early in the design process to help teams understand how users perceive relationships between content, rather than to directly assess how well a menu works.

  • Use it to explore which content users think belongs together.
  • Use it to gather ideas for what users would call each group.

The results still require judgment: different people may place the same content in different groups. NN/g gives the example of participants putting hard-to-categorize items into “Other,” even though website visitors may avoid such a vague label because it is difficult to predict what it contains. Sorting results offer clues for design, not a ready-made navigation solution.

Tree testing: checking whether a defined hierarchy guides users to content

Tree testing starts with a category hierarchy and labels. Participants are then given tasks and asked to find where they would expect the target content to be in a simplified text tree. The test assesses whether users can find content within that structure. It can also compare different structures or different names for the same category.

NN/g describes socks appearing under both “Footwear” and “Accessories.” Observing whether participants look for socks under “Accessories” can help determine whether both locations are needed, rather than relying solely on the team’s intuition. Analysis need not focus only on whether participants ultimately chose a correct answer; it can also examine which categories they selected and whether they went back and forth.

Choose the method to fit the question, not a fixed formula

If ideas for grouping content have not yet taken shape, card sorting is a better fit. If a complete proposed structure already exists and the goal is to check whether important content can be found, tree testing is appropriate. For a redesign, one possible sequence is:

  1. Conduct usability testing on the existing product to identify confusing aspects of the navigation.
  2. Use card sorting to develop new ways of grouping content.
  3. Use tree testing to evaluate and refine proposed hierarchies before moving on to layout design and development.

Tree-testing results do not represent how the complete website will perform

A text tree lacks cues such as placement, color, and imagery. Large menus may also display several levels of categories at once, unlike a test in which users expand one level at a time. Tree testing therefore cannot replace evaluation of the complete interface.

The research format should also match the question. In qualitative tree testing, a moderator can ask why participants made particular choices. This makes it useful for refining the structure early on, but not for collecting statistics such as success rates. Quantitative research can establish benchmarks and compare alternatives, but it requires larger samples and usually cannot explain the reasons behind behavior in depth.

ProcrustesPRProcrustesPREight SEO essentials for beginners: Website speed

Tree testing and card sorting: perspectives from across the web

What should you do next if tree testing shows that users start looking in different categories?

If first clicks are spread across multiple categories for many tasks, the categories may overlap. It is worth reviewing the organization or using card sorting to explore alternative groupings. If participants choose the correct first click but overall success remains low, you can also examine their final destinations to check whether lower-level categories are too similar. These patterns all require further analysis; a single click is not enough to establish the cause.

What should you examine in tree testing when the same content has multiple possible locations?

If a task has several reasonable correct locations, you can designate multiple correct answers and examine which final destination each participant selected. Because tools typically show only the combined success rate across all correct answers, you also need to inspect the detailed data. If one correct location is rarely selected, you can generally consider removing that cross-listing.

Why might you conduct a small qualitative study before quantitative tree testing?

A small qualitative study can first check whether tasks are easy to understand and whether their wording might bias participants. It also allows you to ask why they chose a particular answer and which labels were confusing. These findings can inform revisions to the main test and ideas for improvements. Establishing benchmarks or comparing success rates across alternatives still requires a separately planned quantitative study with a larger sample.

Is card sorting only useful for site-wide navigation?

No. Card sorting can also help you explore how to group content within a single page. For example, it can reveal how users make sense of content in tabbed panels, helping you determine which items belong together. If the research still reveals no clear groupings, tabs may not be a suitable way to present the content. Consider a single-page layout with subheadings instead.

What limitations does tree testing have when categories also have their own landing pages?

With some tree-testing tools, you can usually designate only items at the lowest level of the hierarchy as answers. On a real website, however, a category may have both its own landing page and child pages. When planning tasks and interpreting results, check how your tool represents this structure. Do not assume a page is missing from the actual website just because the tool cannot show it.

ProcrustesPRWhat Ten Years of SEO Taught Me About Its Real Purpose

References

  1. Card Sorting vs. Tree Testingwww.nngroup.com · 2024-02-23T18:00:00+0000 — Source used in the article above.
  2. Tree Testing Part 2: Interpreting the Resultswww.nngroup.com · 2024-01-19T17:00:00+0000 — Discusses: What should you do next if tree testing shows that users start looking in different categories?; What should you examine in tree testing when the same content has multiple possible locations?
  3. Tree Testing: Evaluate Menu Labels and Categorieswww.nngroup.com · 2023-08-06T18:00:00+0000 — Discusses: Why might you conduct a small qualitative study before quantitative tree testing?; What limitations does tree testing have when categories also have their own landing pages?
  4. Tabs, Used Rightwww.nngroup.com · 2024-08-02T17:00:00+0000 — Discusses: Is card sorting only useful for site-wide navigation?

Related articles

Does my website need a speed score of 100?Eight SEO essentials for beginners: Website speedWhat Ten Years of SEO Taught Me About Its Real PurposeHow Do RWD and AWD Work, and Which Should You Choose for Your Website?What are the benefits and development challenges of SPAs?How do UTF-8 and Big5 differ in their effects on search rankings and user experience?What are single-page websites best suited for, and what are their pros and cons?Keep different topics on one site or split them up? Start with audience overlap and maintenance capacity
ProcrustesPR
ProcrustesPR