UX research and interviews

1) Why and when to conduct UX studies

UX research reduces uncertainty and risk: confirms user needs, identifies barriers and priorities, helps make decisions about features and design before expensive development.

• Idea/discovery: you need to understand "what to build" (needs, context, JTBD).
• Concept assessment: I would like to check the value and wording of the offer.
• Design validation: before/after prototype, before release.
• Post-release: why metrics sagged/grow, where friction and insight backlogs.

2) Study classes and interview location

Generative (discovery): in-depth interviews, contextual observations, diary studies, task mapping (JTBD).
Evaluation: usability tests, tree-testing, card-sorting, first-click, surveys.
Causal/experimental: A/B tests, bandits.
Interviews are a key method of generative research: revealing motives, contexts, "works" and barriers.

3) Interview types

Unstructured: Free dialog for primary topic mapping.
Semi-structured (recommended by default): there is a guide and blocks, but the order is flexible.
Structured: fixed scenario for comparability of responses (later).

4) Preparation: goals, hypotheses, users

1. Formulate a business question and research goals (what you will solve).
2. Make working hypotheses (H1...Hn) and key behavioral questions.
3. Identify recruiting segments and criteria (by behavior, not demographics).
4. Decide the format: remotely/in person, individually/focus group (for interviews it is better 1:1).
5. Plan logistics: slots, record, consent, reward, backup links.

How many respondents?

For high-quality interviews, the benchmark is 5-8 people per segment until saturation (when new sessions do not bring new topics). For 2-3 key segments, these are usually 12-20 interviews.

5) Recruitment and screening (template)

Principles: hire "by behavior," weed out professional respondents and industry employees, check the criteria through specifics ("Tell us about the latest experience...").

• Q1. Describe the last time you . When was it?
• Q2. What tools/sites/apps did you use?
• Q3. How often are you ? (Weekly/Monthly/Less)
• Q4. Exclusion scenario: "Do you work in UX/marketing/research?" (delete "Yes")
• Q5. Age, country/language, platform (desktop/mobile), write consent.

6) Ethics, confidentiality, consent

Obtain informed consent: purpose, format, record, anonymization, right to refuse.
Minimize PII, store data securely, use anonymous IDs.

With sensitive topics (finance, excitement, health) - soft formulations, the right to "skip."

7) Design guide interview (template)

1. Warm-up (5 min): Context, role, recent experience.

• Real life stories (last time/hardest/most successful).
• "Before-during-after": three phases of the journey.
• Motives, triggers, barriers, compromises, alternatives.
• JTBD wording: "When , I want to ."
• 3. Checking concepts/screens (10-15 min) - if appropriate.
• 4. Completion (5 min): What was unexpected, what was not asked, readiness for a followup.

• "Tell us about the latest case where... What was the starting point?"
• "What alternatives have been considered? Why did you choose this?"
• 'What was the hardest part? How did you cope?"
• "If one step could be removed, what would it be and why?"

What to avoid: suggestive ("Do you agree that..."), double ("When and why..." in one), hypothetical ("And if..." - possible, but at the end), estimated ("Like/dislike?" instead of "How did you do it in practice? »).

8) Conduct: moderation without distortion

Build a report, explain the role, emphasize that "we are not testing you."

Use active listening: echoes, refinements, pauses.

Tricks: 5 Why, "show how you do it," "think out loud."

Monitor neutrality (tone, facial expressions), record what is observed, not interpretations.
Enter timecodes for subsequent decryption.

9) Documentation and storage

Record + transcript, notes on the template: Situation → Action → Result → Emotion → Quote → Opportunity.
Immediately after the session, fill in debrief (5-10 minutes), mark the latest insights and questions.

10) Qualitative data analysis

Coding: open → axial → selection of key themes; agree on a codebook in advance.
Affinity-charting: grouping citations/observations into clusters, then themes/insights.

JTBD synthesis: map "Situation → Motive → Expected outcome → Barriers → Triggers."

Persons/segments: based on behavior and motivation, not just demographics.
CJM/Service Blueprint: stages, tasks, pains, emotions, metrics, backstage processes.
Reliability: inter-coder agreement, source triangulation (interviews + behavioral data + metrics).

11) Insight to Solution

• Insight: Observed pattern ("Players get confused between balance and bonus score").
• Opportunity: how to turn it into value ("Reduce wallets into a single panel and add hints").
• Solution/Hypothesis: "If we do X, we will see the Y-metric."
• Experiments: prototype → usability → pilot → A/B, with clear success criteria.

12) Research quality metrics

Theoretical saturation, triangulation, transparent artifact (replicability).
For quantitative follow-up steps: effect size, test power, confidence intervals.
For usability: task success, task time, SEQ/UMUX-Lite, errors/points of failure.

13) Specificity of domains with sensitive scenarios (finance/iGaming)

Interview about money/rates/privacy with increased sensitivity.
Emphasize anonymity, avoid shame/evaluative language.
Prototypes - with disguised data; discussing risks/limits - through real stories, not abstractions.
Do not forget about compliance (rules of responsible play, age, KYC/AML): exclude inappropriate groups in advance by screening.

14) Study plan template (copy to wiki)

Purpose:

Research questions:

Methods:

Segments and criteria:

Sample:

Risks/Ethics:

Procedure:

Output artifacts:

Research success metrics:

Timeline:

Responsible:

15) Interview guide template (fragment)

1. Introduction (1-2 min): purpose, format, confidentiality.

2. Warm up: "Tell us about yourself and your latest challenge ."

3. Stories: "Describe how you last . What was the trigger? What are the alternatives?"

4. Phase "before": expectations, selection criteria.
5. Phase "in time": steps, tools, barriers, emotions.
6. Phase "after": result, satisfaction, what to improve.

7. Concept/screen check (if any): "What do you expect to see next? Where would you click?"

8. Closing: "What didn't I ask? What was the most helpful part of our conversation?"

16) Moderator checklist

• Goals and hypotheses are formalized
• Screening approved, recruitment launched
• Consents ready, record verified
• The guide is polished, timecodes are stamped
• Neutral wording, pauses, "show how"
• Debrief after each session
• Single codebook and board for Affinity
• Synthesis is decorated in artifacts, insights are tied to data
• Experiment backlog and owner per hypothesis

17) Antipatterns

Interview "about opinions," and not about behavior and stories.
Mixing screening and interviews (respondents "adjust").
Leading questions, no pauses, "protection" of your own solution.
No system analysis (no coding, no Affinity).
Insights without reference to a business issue and metrics.

18) Instruments (by role)

Conducting and recording: Zoom/Meet, Lookback, OBS.
Transcripts and coding: Dovetail, Airtable/Notion, Excel/Sheets.
Synthesis and visualization: FigJam/Miro, Whimsical.
Recruiting: respondent panels, their bases, screening forms.

19) Fast start (one-pager)

1. Describe the business question → 2) select 1-2 segments → 3) screen → 4) conduct 6-8 semi-structured interviews → 5) Affinity + codebook → 6) design insights and CJM → 7) formulate 3-5 experiments with success criteria.