Pick the Right Test: A Practical Guide to 7 UX Research Methods

Choosing a research method is mostly about matching the method to the question you're trying to answer. Some methods tell you how people organize information. Others show whether people can find things, whether a design makes a good first impression, or whether a whole flow works. Pick the wrong one and you get tidy data that doesn't answer your real question.

Below are seven common methods, each with the same three questions answered: what it's good for, which methods pair well with it, and the best time to use it.

1. 5 Second Test

Participants view a design for five seconds, then answer questions about what they remember or understood.

What it's good for

• Measuring first impressions and gut reactions
• Checking whether the main message or purpose of a page is clear
• Finding out which elements stand out (and which get ignored)
• Evaluating visual hierarchy, branding, and clarity of value proposition

Some methods to pair it with

• Preference test: after learning what each design communicates, find out which one people favor.
• First click test: once the message lands, check whether people know where to go next.
• Survey: scale up the findings with a larger audience, or add questions about brand perception.

Best time to use it

Read all about 5 Second testing with UserSage here.

2. Card Sorting

Participants group topics or content items into categories that make sense to them, and sometimes name those categories. It can be open (they create the groups), closed (you provide the groups), or hybrid.

What it's good for

• Understanding how users mentally organize information
• Generating ideas for navigation, menu labels, and site structure
• Discovering the language your users use for categories
• Settling internal debates about how content should be grouped

Methods to pair it with

• Tree testing: the classic pairing. Card sorting helps you build a structure; tree testing checks whether people can actually find things in it.
• Survey: gather broader input on terminology and priorities.
• Prototype test: confirm the resulting navigation works in a realistic interface.

Best time to use it

Read all about Card Sorting with UserSage here.

3. First Click Test

Participants are given a task and shown a design. You record where they click first.

What it's good for

• Evaluating whether navigation, labels, and layouts guide people in the right direction
• Identifying confusing or misleading links and buttons
• Quickly comparing the clarity of different layouts

Research has suggested that people who get their first click right are far more likely to complete the task successfully, which makes this a fast check.

Methods to pair it with

• 5 second test: first check that the page communicates clearly, then check that people know where to click.
• Tree testing: if first clicks fail, test whether the underlying structure or labels are the problem.
• Prototype test: follow up on the full task flow after the first click.

Best time to use it

Read all about First Click Testing with UserSage here.

4. Preference Testing

Participants see two or more design options and choose the one they prefer, and explain why.

What it's good for

• Choosing between design directions, layouts, visuals, or copy variations
• Understanding which options people find more appealing or trustworthy
• Getting qualitative reasons behind preferences
• Giving stakeholders evidence to support a decision

A word of caution: preference is not the same as performance. People can like a design that doesn't work well, so avoid using this method alone to judge usability.

Methods to pair it with

• 5 second test: see what each option communicates before asking which one people like.
• Prototype test: check that the preferred option actually works in practice.
• Survey: collect preferences from a larger or more segmented sample.

Best time to use it

Read all about Preference Testing with UserSage here.

5. Prototype Test

Participants interact with a clickable or interactive prototype while completing realistic tasks, often with a researcher observing and asking questions.

What it's good for

• Finding usability problems before development
• Observing how people actually move through a flow
• Testing task completion, errors, confusion, and satisfaction
• Validating interaction design, not just visuals

Methods to pair it with

• Card sorting and tree testing: establish and validate the structure before it appears in the prototype.
• First click test: zoom in on specific screens where the prototype test reveals trouble.
• Survey: follow up with satisfaction or confidence questions, or gather feedback from a broader group.

Best time to use it

Read all about Prototype Testing with UserSage here.

6. Survey

Participants answer a set of questions, which can be quantitative (ratings, multiple choice) or qualitative (open-ended).

What it's good for

• Collecting attitudes, opinions, needs, and self-reported behaviors at scale
• Segmenting users and understanding demographics
• Measuring satisfaction over time (e.g., with a consistent set of questions)
• Prioritizing features or problems based on how many people care

Surveys tell you what people say, not necessarily what they do, and they're only as good as their questions. Keep them focused and avoid leading wording.

Methods to pair it with

• Prototype test or interviews: explain the "why" behind survey results.
• Card sorting: validate labels and categories with a bigger audience.
• Preference test and 5 second test: add supporting questions to a design test.

Best time to use it

Read all about Surveys with UserSage here.

7. Tree Testing

Participants are shown a text-only version of your site hierarchy (no visuals, no layout) and asked to find where they'd locate specific items.

What it's good for

• Evaluating whether your information architecture is findable
• Identifying confusing labels and categories
• Seeing where people go wrong, including which wrong paths they choose
• Benchmarking an old structure against a new one

Because it strips away visual design, any problems you find are about the structure and labels themselves.

Methods to pair it with

• Card sorting: build the structure with card sorting, then validate it with tree testing.
• First click test: once the structure is solid, check that the visual interface supports it.
• Prototype test: confirm the navigation works in context.

Best time to use it

Read all about Tree Testing with UserSage here.

Ready to get started?

Run your first research test with User Sage. Sign up for free and get a first read today.