How to Conduct UX Research: From User Interviews to Behaviour Analysis

UX research grounds product decisions in real user behaviour and needs, rather than team assumptions alone. Choose the research method according to the question being asked.

Interviews and usability tests help explain why users do not complete an action. Product analytics may be better suited to showing how many users drop out at each step.

Write the Research Question First

“Let’s understand users” is not a research question. “Which information do new customers struggle to understand on the payment screen?” is a measurable question that can be investigated.

Select the Right User Segment

New users, experienced users, customers who leave and different role groups may view the same problem differently. Participant selection directly affects research findings.

User Interview

Asking users to describe recent real behaviour, rather than hypothetical preferences, can yield more reliable insights in interviews.

Usability Testing

Give users a realistic task and observe where they hesitate, misunderstand or go back. The moderator should not lead them.

Behavioral Data

Funnel
Event
Task completion
Search query
Form abandonment
Retention
Breakdown by device and channel

Combine Qualitative and Quantitative Data

Analytics can reveal how widespread a problem is; interviews can indicate possible causes. Combining the two supports better decisions.

Synthesis

Do not leave notes as a long report alone. Group recurring problems, needs, motivations and barriers into themes.

Connect Research to the Product Backlog

Not every finding requires a product change. Build the backlog by considering evidence strength, impact on users and business importance together.

How to Ground UX Decisions in Data

In this UX work, research questions, participant segmentation and user interviews should not be treated merely as design preferences. Identify the user problem behind each critical interface decision and the behavioural metric that will validate it. Changing a button’s colour alone is not a UX improvement; measurable outcomes include users recognising a task more easily, completing it faster or making fewer errors.

Combine quantitative data with qualitative research. Funnel or event analysis can show where users drop out; interviews and usability tests provide context for why. Measuring the same metrics after a change turns an assumption into a product decision that can be validated. A UX backlog item should therefore record the problem, hypothesis, target user, success metric and acceptance criteria, not just a screen or component name.

Implementation and Measurement Checklist

Has the critical user task been defined?
Is baseline data available for current behaviour?
Have mobile and desktop scenarios been tested separately?
Have error states and empty states been designed?
Have accessibility checks been completed?
Has a success metric been defined for evaluating the change?
Has user feedback been linked to the product backlog?

Performance and technical behaviour in critical user flows are part of UX. Slow API responses, unclear loading states after form submission, data loss or a mobile keyboard covering a field can undermine the real experience even if these problems are absent from the design file. Frontend, backend, analytics and design teams should therefore assess the same user scenario end to end.

Research, Hypothesis and Iteration Cycle

A UX change should begin with a hypothesis that is as clear as possible. Instead of “the user dislikes this screen”, use a testable problem statement such as “uncertainty around research-question formulation reduces task completion”. Analytics data, user interviews, support records, usability tests or observed operational problems can support the hypothesis. The team therefore seeks evidence of the problem before assuming a solution.

Next, define the target behaviour for participant segmentation and user interviews. Depending on context, use task completion rate, time, error rate, funnel progression, form abandonment or the number of user corrections. After releasing the new design, compare the same metrics against the previous baseline. If results do not move in the expected direction, reassess the hypothesis rather than defending the design decision.

For sustainable UX optimisation, research findings should not remain in one-off presentations. Keeping the problem, evidence, proposed solution, user segment, success metric and outcome in the same backlog item lets the team see which design decisions actually work over time. This turns UX work from aesthetic revision into a measurable product-development cycle.

B10 Digital Agency’s UX/UI Approach

B10 Digital Agency treats user experience not just as interface aesthetics, but as a product-development discipline combining research, user flows, data analysis, prototyping, technical feasibility and measurement.

Frequently Asked Questions

How many participants does UX research need?

There is no fixed number. It depends on the research method, user diversity and scope of the problem.

Is it enough to implement what users say they want?

No. A user’s proposed solution may differ from the underlying need.

Does analytics count as UX research?

Product usage data is one quantitative source for UX research.

Is research only done before product development?

No. It can support discovery, development and optimisation throughout the product life cycle.

Make User Experience Measurable

Explore B10 Digital Agency’s UX/UI service to improve user flows across web, mobile and enterprise products through research, prototyping and behavioural data.

Explore Design & User Experience →