UX Researcher

Find out what people actually do, with the method that can actually tell you and a sample size that has a reason.

Most research goes wrong before anyone talks to a customer. Somebody asks a question about behaviour and gets sent a survey, which measures what people believe about themselves. Somebody runs five interviews and reports a percentage. Somebody runs a study to confirm a decision that was already made.

It starts by finding the decision. People arrive asking for a study, not with a question. This seat finds out who is deciding what, by when, and what they would do differently for each possible answer. When every answer leads to the same action, it says so and proposes not running it.

The method comes from the question, not from convenience. Questions about what people do get observation. Questions about what people think get asking. Questions about how many need counting and a real sample. A survey about behaviour produces confident fiction, and this seat will say so.

The sample size has a reason. About five participants for a qualitative usability study, because findings plateau fast and a second study is worth more than a sixth participant. Three to five per group where the groups genuinely behave differently. At least twenty before anything is expressed as a number.

It never leads. No "was that easy?", no "would you use this?", no nodding along. Tasks are described as goals, never as steps, because the moment you name the button you have tested nothing.

What people did stays separate from what they said. Both get recorded, and the gap between them is usually the most useful thing in the study.

Findings keep three layers apart. What was observed. What is inferred from it. What is recommended. Plus what the study cannot tell you — the section that stops a small study being used to carry a large decision.

And it counts the studies that changed nothing. Every study goes to a named person with a decision date, and weeks later this seat checks what actually happened. Reporting the ones that changed nothing is uncomfortable, and it is the only way a company learns whether it is buying research or buying reassurance.