Milestone map
Milestone map
3 milestones
Design the Research Protocol and Recruit Participants
1–2 weeks (3–4 hrs/week)
Define a user research question for a real product, service, or internal tool: what behaviour, need, or pain point do you want to understand? Write a research protocol covering: the research question, participant criteria (role, experience level, relationship to the product), recruitment method, interview format (semi-structured with open questions), and the ≥10 core interview questions. Pilot the protocol with ≥1 internal colleague before external recruitment — a pilot surfaces ambiguous questions, timing issues, and question ordering problems before they affect real participant data. A research protocol without a pilot run is a protocol that was never tested.
Proof required
Submit your research protocol (research question, participant criteria, ≥10 interview questions with rationale for each), and your pilot run notes: who the pilot was with, what changed after the pilot, and the final revised question list.
What gets checked
- Research question is specific and investigable — 'understand user behaviour' is not a research question; 'understand why registered users do not complete the onboarding flow after day 1' is
- ≥10 interview questions are listed with a brief rationale for each — not just the question list
- Pilot run notes document at least one change made after the pilot — a protocol that changed nothing after piloting was not piloted critically
Common mistakes
- Interview questions that lead the participant — 'don't you think the checkout is confusing?' is not a research question; it is a confirmation prompt that produces biased data; all questions must be open and neutral
- Skipping the pilot run — every interview protocol has at least one question that is ambiguous, too long, or in the wrong order; a pilot surfaces these before they affect real participants
Resources
Foundationstart here
Depthgo deeper
What a verifier looks for
- Confirm research question is specific and investigable — flag vague 'understand users' questions.
- Confirm ≥10 interview questions are listed with rationale — flag question lists without justification.
- Confirm pilot run notes document at least one change — flag if the pilot produced zero changes.
- Confirm interview questions are open and neutral — flag any leading or closed questions.
Conduct ≥5 Interviews and Document Findings
3–4 weeks (depending on participant availability, 3–4 hrs/week of active work)
Conduct ≥5 user interviews following the piloted protocol. Aim for 45–60 minutes each. Record or take detailed notes (with participant permission). After each interview, write a brief debrief note (5–10 minutes) capturing: the 3 most important things learned, any surprises, and any protocol adjustments for subsequent interviews. After all interviews are complete, create a findings document: one participant summary per interview (1 page each) covering: background, key quotes (verbatim), observed pain points, and participant goals. The raw data from ≥5 participants is the foundation for the synthesis in M3 — without documented raw findings, analysis cannot be verified.
Proof required
Submit ≥5 participant summaries (one per interview: background, ≥3 key verbatim quotes, pain points observed, goals stated) and your interview debrief notes for each session.
What gets checked
- ≥5 participant summaries are present — not merged into a single composite
- Each summary includes ≥3 verbatim quotes — not paraphrased summaries of what participants said
- Debrief notes for each session are present — including at least one 'surprise' or protocol observation per session
Common mistakes
- Paraphrasing instead of quoting — verbatim quotes are the primary data of user research; paraphrasing at the recording stage introduces researcher bias before analysis begins
- Merging all participants into a composite summary without individual summaries — individual summaries are required to verify that the insights came from multiple distinct participants
Resources
Foundationstart here
Depthgo deeper
What a verifier looks for
- Confirm ≥5 individual participant summaries are present — flag if interviews are merged.
- Confirm ≥3 verbatim quotes per participant — flag paraphrased summaries.
- Confirm debrief notes for each session include at least one observation about the protocol or a surprise.
Synthesise Findings and Write the Research Report
2–3 weeks (3–4 hrs/week)
Synthesise findings from all ≥5 interviews using affinity mapping or thematic analysis: cluster observations and quotes into themes, identify patterns that appear across ≥3 participants, and separate themes from single-participant observations. Write a research report (1,500–2,500 words) covering: research question, participant profiles (anonymised), methodology, key themes with supporting quotes (attributed to participant numbers, not names), implications for product or design decisions, and ≥3 specific recommendations. Present the report to ≥2 stakeholders — product, design, or engineering colleagues — and facilitate a 30-minute discussion of the implications.
Proof required
Submit your research report (1,500–2,500 words) with key themes supported by verbatim quotes attributed to anonymised participants, ≥3 specific recommendations with rationale, and a stakeholder presentation record: attendees, ≥3 questions raised, and your responses.
What gets checked
- Each theme is supported by ≥2 participant quotes attributed to anonymised participant identifiers (P1, P2, etc.) — not stated without evidence
- Themes are distinguished from single-participant observations — a theme requires ≥3 participants; single-participant observations are noted separately
- ≥3 recommendations are specific and actionable — not 'improve the onboarding' but 'add a progress indicator at step 2 of onboarding because 4 of 5 participants said they did not know how many steps remained'
Common mistakes
- Single-participant observations presented as themes — a finding from one participant is an interesting observation; a theme requires ≥3 participants to be called a pattern; conflating the two produces unreliable recommendations
- Recommendations without connection to the research findings — every recommendation must trace to ≥2 participant quotes; untethered recommendations are not research outputs, they are opinions
Resources
Foundationstart here
Depthgo deeper
What a verifier looks for
- Confirm each theme is supported by ≥2 participant quotes attributed to anonymised identifiers — flag themes without direct quote support.
- Confirm themes are distinguished from single-participant observations — flag single-participant observations presented as patterns.
- Confirm ≥3 recommendations each trace to ≥2 participant quotes with specific rationale — flag recommendations without evidence.
- Confirm stakeholder presentation record names attendees and documents ≥3 questions raised.
- ADVERSARIAL VERIFICATION NOTE: user research synthesis is AI-fakeable at the artifact level (themes and quotes can be generated); the verifier should ask the presenter to recall a specific surprising quote from a named participant and explain what was unexpected about it — genuine researchers retain vivid memories of participant surprises that AI-generated reports do not contain.