Practical learning lesson / 8 MIN READ
Recruit people who can disagree with you
Useful research starts before the first question, with who is invited, what they understand and whether they can participate freely.
AI-assisted research and analysis for the SNS Creative Fieldbook
Sources checked 2026-11-05
Build recruitment, consent and interview conditions that protect participants and allow inconvenient observations to emerge.
What you will leave with: Prepare a bounded research round with a clear recruitment brief, respectful invitation, neutral guide and responsible data plan.
Use this when
- The team only interviews enthusiastic customers or colleagues.
- A campaign idea needs customer discovery before creative development.
- Participants may feel pressured to praise a service they depend on.
This covers ordinary adult customer discovery. Research involving children, vulnerable groups, sensitive topics or clinical questions needs appropriate specialist design and safeguards.
Recruit for the question rather than the desired answer
State what you need to learn and which experiences make someone relevant. If the question concerns abandoning a booking, include people who abandoned as well as people who completed it. Recruiting only loyal advocates may produce useful stories about loyalty but poor evidence about barriers. Define inclusion and exclusion criteria before invitations go out. Keep the screen short and avoid revealing the answer you hope participants will give.
Choose a range of relevant situations rather than pretending a small qualitative group represents the whole market. Consider access, language, device use and ability to attend. A convenient time slot can quietly exclude important customers. Record how recruitment happened so later interpretation can account for selection. Do not use a demographic characteristic as a substitute for the behavior or circumstance the research actually needs to understand.
Make the invitation and consent understandable
Explain who is doing the research, its purpose, the activities, expected duration, any incentive and how information will be used. Give people time and a route to ask questions. The GOV.UK consent guidance emphasizes informed, voluntary participation and clarity about observation, recording, sharing and withdrawal. Use that guidance as a practical reference while obtaining appropriate local privacy and ethics advice for the actual project.
Separate agreement to an interview from agreement to appear in advertising, receive marketing or provide a public testimonial. A participant should not discover later that a research recording became promotional content. State incentive terms fairly and do not make payment depend on praise. Where the researcher also controls a service relationship, take extra care that declining or criticizing does not appear to threaten access, employment or customer support.
Run a conversation that permits contradiction
Open with a simple reminder of the session’s purpose and permissions, then ask about a recent relevant event. Follow the sequence of what happened before introducing the proposed solution. Use open prompts and requests for clarification rather than questions that contain the desired conclusion. The GOV.UK interview guidance supports planned topics, accessible arrangements and follow-up questions. A guide gives coverage without requiring the interviewer to read a rigid script.
When showing an idea, distinguish comprehension from preference and stated intention from observed behavior. Ask what the person thinks it offers, what they would do next and what concerns remain. Avoid defending the work when it is misunderstood. Leave room for silence and an honest negative response. One person’s criticism may expose a useful problem, but neither praise nor criticism becomes a market percentage merely because it was strongly expressed.
Keep the evidence useful without keeping everything
Collect the information needed for the research question and agreed follow-up, not an unrestricted record of the participant’s life. Limit access to recordings and separate contact details from working notes where practical. The GOV.UK research-data guidance recommends protecting participant privacy and reviewing arrangements with a data-protection expert. Tell participants about relevant transcription or research suppliers rather than treating a convenient tool as an invisible back-office step.
In notes, distinguish a quotation, an observed action and the researcher’s interpretation. Preserve enough context to avoid misreading a remark, while removing unnecessary identifying details from shared outputs. A distinctive story may still identify someone even after their name is removed. Set retention and deletion responsibilities before collection begins. Keep evidence of the permissions that apply, and do not promise anonymity if the actual process cannot provide it.
Do not confuse research with a brand performance
Dove’s Real Beauty Sketches and Heineken’s Worlds Apart use structured encounters as compelling brand material. Those films can prompt questions about framing, editing and participant experience. They are not published research protocols that validate an ordinary customer-interview method. A scene designed to deliver an emotional reveal has a different purpose from discovery designed to let the team find out that its hypothesis is wrong.
Close the round by reporting what was heard, who was included, what remained absent and which questions need further evidence. Avoid merging participant consent, production releases and research conclusions into one convenient story. If the team wants to use an interview in a campaign, treat that as a separate, clearly explained decision with appropriate permission. The practical success of the round is better understanding under respectful conditions, not a collection of flattering quotes.
Make the reasoning concrete.
A subscription service wants to understand why customers stop after the first month. The account team proposes interviewing six enthusiastic long-term customers.
Guide
Ask about the last period of use, the moment cancellation became likely and alternatives considered before showing a proposed retention message.
Exercise
Write a neutral invitation, three screening questions and two non-leading prompts. Identify a data field you can remove and explain how you will request separate permission if a quotation is later considered for advertising.
Recruit relevant recent cancellers as well as continuing customers, with clear information that participation does not affect support or account status. The sample is for discovery, not a churn-rate estimate.
Take the questions into your next brief.
Use these prompts in your own notes. Nothing is submitted from this page.
Research question
What uncertainty can an interview actually help resolve?
Recruitment criteria
Which experiences are needed and who might the convenient sample exclude?
Participation information
What will happen, what is optional and how can someone stop?
Neutral guide
Which questions invite a real episode and allow disagreement?
Data plan
What will be recorded, shared, retained and deleted, and by whom?
Reporting limits
What cannot be generalized from this round?
Read with a question in mind.
Dove / Unilever
Real Beauty Sketches →
How does a branded encounter differ from an open research interview?
The contrast between self-description and another person’s description illustrates how framing can shape an emotional story.
What it does not prove: The edited campaign does not validate a research protocol or establish a universal psychological conclusion.
Heineken
Worlds Apart →
Where does a filmed encounter become a broader participation program?
The Human Library partner account confirms the campaign’s connection with public dialogue events.
What it does not prove: That partnership does not disclose every participant safeguard or prove that a marketing team should recreate the filmed setup as customer research.
Watch for the wrong conclusion.
- Recruiting only people likely to praise the brand.
- Making an incentive or service relationship feel conditional on agreement.
- Using interview consent as permission for public advertising.
- Treating edited campaign encounters as validated research methods.
Your next action
Review your next recruitment brief and remove one criterion or question that selects for the answer the team wants.
Sources & scope
Original editorial synthesis with linked campaign evidence. Exercises and calculations are hypothetical, not historical campaign facts. No named human author or reviewer is claimed.
- GOV.UK Service Manual: Getting informed consent for user research →Primary guidance · Supports: Participant information, voluntary participation, recording, sharing and withdrawal in the UK government research context.
- GOV.UK Service Manual: Using in-depth interviews →Primary guidance · Supports: Planning interviews, recruiting relevant users, accessibility, discussion guides and follow-up practice.
- GOV.UK Service Manual: Managing user research data and participant privacy →Primary guidance · Supports: Protecting recruitment and research data; reviewing practices with a data protection expert; supplier and public-reporting considerations.
- Ogilvy: Real Beauty Sketches →Primary agency case · Supports: Creative objective, contrast between self-image and others’ descriptions, and São Paulo agency team.
- Edelman: Worlds Apart →Primary agency case · Supports: Agency objective, distribution, partnership and reported views/sales.
- Human Library: Heineken partnership, 20 March 2018 →Primary partner account · Supports: Partner confirmation of 2017 campaign and five public events.