Interview moves that turn branding research into evidence, and how to use them on agency references

What can one interview establish?

An interview can establish what a particular participant says they remember and experienced around a decision. It cannot show how common that experience is, prove what caused it, or describe the branding market as a whole. Treat the account as one source among several, not as a market measure.

Keep a participant's account separate from what you directly observed and from your interpretation. Label gaps as unknown rather than filling them with assumptions. That distinction lets readers see whether a point comes from a person's recollection or from your analysis, and stops an inference being mistaken for a verified event.

Which moment should the conversation explore?

Ask about a specific past event the participant can recall, such as a discussion about a name, identity or brand rule. A real episode gives the conversation a sequence to explore, rather than asking someone to predict what they might do in a hypothetical situation.

Let participants choose the example. If they cannot recall a relevant event, record that absence instead of offering a scenario for them to agree with. Define who the research is for and set eligibility before inviting people; define the audience before treating the conversations as evidence about a group.

A delayed project, rejected tender or internal route may help explain one person's account, even where no agency took part. Record these as possibilities raised by that participant, not as findings about demand. If there is no suitable event to discuss, leave the gap visible instead of completing the story yourself.

How do I ask without leading someone?

Use open prompts about what happened, in what order and who was involved. Give the participant room to describe an experience that differs from your expectations, including one where nobody collected evidence. Avoid wording that assumes an agency was needed, a change worked or someone can predict a future decision.

Keep prompts consistent enough to compare conversations, while letting people describe events in their own words. A neutral question does not make its answer a measure of prevalence. If someone mentions a rejected or delayed route, ask what happened without presuming a cause or turning an individual account into a wider pattern.

What can a rejected option tell me?

A rejected option may reveal a constraint in one particular account, but it cannot show that the constraint is widespread or permanent. Ask which alternatives were considered and why they were not chosen at that time, then report the response within those limits.

Follow up on what participants say had to be in place before work could proceed, such as evidence, authority, rights terms, implementation commitment or a procurement step. One person's explanation does not prove what caused an outcome; attribution methods can help keep a reported explanation distinct from evidence of causation.

Whose route through the decision should I trace?

Follow the people and handoffs that participants actually describe instead of inferring authority from job titles. Ask who first noticed the issue, who shaped the brief, who approved spending and who used the output. Record only the stages they can identify, leaving unconfirmed roles unresolved.

Ask which records participants are authorised to discuss. A brief, approval note, asset register or evaluation criterion may place an account in context. If a person, stage or record cannot be identified, mark it as unknown; an organisation chart cannot supply an answer that the participant did not give.

What should a participant know before speaking?

Explain the session's purpose, whether it will be recorded, how information will be used and retained, how withdrawal works, and whether observers will attend. Participation is not permission to use the conversation for marketing, so keep the participation record separate from the research notes.

Keep confidential material and personal details out of an agency repository unless they are needed for the research. Clear handling of participant information helps reviewers assess related practices. Say what the session is for before it begins, and do not imply that taking part guarantees any particular use of the account.

The Magenta Book is an evaluation reference, while the Government Social Research ethical assurance process concerns research ethics in that setting. The Government Social Research Code is another public-sector reference. For further context, consult the UK Data Service's ethical issues guidance and UKRI's research integrity resource.

How can I turn notes into evidence?

Keep a traceable route from each session record to each finding, and label participant statements separately from your interpretation. The people who took part, and those who did not, limit what the themes can support. Do not present interviewees as a complete account of the market.

For each session, retain the field date, interviewer and participant code with the research records. Check the full record before using a line as a quotation, and label each analysis item as asked, observed, inferred or unknown. Those labels let readers distinguish reported information from conclusions drawn from it.

If a participant describes a delayed project, record the delay as part of that account. Do not recast it as proof that agencies are unnecessary or that one cause explains the delay. Check measurement mistakes and place interview material alongside other evidence using measurement and reporting.

How do I use these moves on a branding agency's references?

A reference call with a past client of a branding agency is an interview, and the same moves apply. Ask for one concrete project, not a verdict on the agency. Record what the client says, what you can check, and what you infer.

Useful neutral prompts: What was the brief, and who signed it off? What changed between the first proposal and the final files? Which deliverable needed the most correction, and how was it settled? What did you receive at handover, and what did you have to ask for? What would you have wanted in the contract?

Treat the answers as one account. A single warm reference does not rank the agency, and a single complaint does not disqualify it. Compare several calls against the same prompts, and read them next to the agency's review evidence pack and its supplier paperwork.

What should I do first?

Start by defining who may take part and which past event the conversation should explore. Prepare the recruitment and session records before making contact, so participants can understand the purpose and limits of the research.