UX Research Methods

UX research methods are the structured ways teams learn what users need, what confuses them, and whether a design works — before shipping the wrong thing confidently. The hard part isn't knowing the methods. It's choosing the one that answers the question you actually have.

This page maps the methods that earn their keep, the two axes that organize all of them, and a simple rule for matching a method to a question. If you're trying to decide between researching and just auditing what you have, the Full UX Audit section below draws the line.

Two axes that organize every method

Almost every UX research method sits somewhere on two axes:

• Attitudinal ↔ behavioral — what people say (interviews, surveys) versus what they do (usability tests, analytics). The gap between the two is where most insight lives.
• Qualitative ↔ quantitative — why something happens (small sample, rich detail) versus how much it happens (large sample, numbers you can act on).

Plot a question on those two axes and the right method usually picks itself.

The methods that earn their keep

• User interviews — the why. Best for motivation, mental models, and unmet needs. Useless for "is this usable?" (people are unreliable narrators of their own behavior).
• Usability testing — the can-they. Watch real people attempt real tasks. The single highest-signal method for an existing flow; five users surface most issues.
• Field studies & contextual inquiry — the in-context. Observe people in their real environment, where the constraints you'd never guess actually live.
• Surveys — the how-widely. Cheap to scale, easy to misuse. Great for prevalence, terrible for "why."
• Analytics & session review — the what-actually-happened. Behavior at scale, no recall bias — but it tells you where users dropped, never why.

You rarely need all of them. You almost always need more than one.

Match the method to the question

The discipline is to start from the question and work backward:

1. Write the actual question. "Why do users abandon onboarding on step 3?"
2. Classify it. Behavioral and qualitative → watch people do it.
3. Pick the lightest method that answers it. Five usability sessions beat a 200-person survey for a "why."
4. Triangulate the expensive decisions. When the call is costly, confirm a qualitative finding with a quantitative one before you commit.

Good cheap research beats expensive research done for its own sake — see qualitative UX research on a budget for what that looks like in practice, and the Desk Research course for the synthesis side.

When you don't need research yet

Skip new research when the answer already exists — in analytics you haven't read, in support tickets, or in a heuristic review. If the product is live and leaking and you need a prioritized verdict fast, an expert UX audit (or a quick UX Snapshot) gets you further than another round of interviews. Research and audit aren't rivals — research explains, an audit scores and prioritizes.

A method-selection checklist

• [ ] Have you written the question in one sentence?
• [ ] Is it about attitude or behavior — what they say or what they do?
• [ ] Do you need why (qualitative) or how much (quantitative)?
• [ ] Is there a lighter method that answers it just as well?
• [ ] For costly decisions, are you confirming the finding a second way?

Answer those and you'll spend research budget on questions that change decisions — not on data nobody acts on.

See how we turn research into a prioritized verdict →