Research methods
Which UX research method should you reach for?
A quick field guide to matching the question you're actually asking to the research method that answers it, without over-researching.
The hardest part of UX research isn't running a study. It's picking the right one for the question in front of you - and not reaching for a heavyweight method when a ten-minute test would do.
Here's the shortcut: name the question first, then choose the method that answers it. Most research falls into one of a few shapes.
"How do people group these ideas?"
When you're organising navigation, categories or a taxonomy, you want to know how users mentally group things - not how your org chart does. That's card sorting: give people the items and let them cluster and label them. Pair it with affinity mapping to turn the raw groupings into themes you can act on.
"Can people find the thing?"
If the worry is findability - can someone locate a feature, a page, a setting - you don't need a full usability session. First-click testing answers it directly: show the interface, ask where they'd click to do X, and measure whether the first click lands somewhere sensible. It's cheap, fast, and the first click predicts task success surprisingly well.
"Why did they struggle?"
When you already know that people fail but not why, you need to watch. Moderated usability testing - five users, thinking aloud - surfaces the friction that metrics only hint at. Keep the tasks realistic and resist the urge to help; the silence is the data.
"What do they actually think?"
For attitudes, motivations and mental models, talk to people. Interviews are unmatched for depth, but they're easy to bias - ask about behaviour, not hypotheticals, and let them do most of the talking.
Don't over-research
Every method carries a cost, and the biggest failure mode isn't picking the wrong one - it's adding cognitive load to your own team by researching things you already know. If the answer is cheap to test in production, test it there. If a five-minute first-click test settles the debate, don't schedule a fortnight of interviews.
Match the method to the question, run the smallest study that answers it, and move on. The point of research isn't to research - it's to decide.