Six practices for social scheduling interviews, from candid invitation to missing voices

Social media scheduling customer interviews should reconstruct recent work rather than ask them to predict a purchase. The six practices below form a research method, not a report of completed interviews.

Method: desk review of current UK regulator and government research guidance, mapped to a social publishing workflow. Research date: 6 September 2026. Scope: adults participating through organisations operating under a stated England rule, with UK legal sources labelled. Inclusions: recruitment, participation, questioning, evidence handling, accessibility and analysis practices. Exclusions: medical research, children, covert observation and product sales calls. Ranking: non-ranked; sequence follows the life of an interview. Conflicts: no supplier funded this article and no interview service was tested.

What to take away

1. Give a candid invitation

Name the researcher, commissioner, topic, reason for selection, time, incentive, recording plan and withdrawal route. Avoid implying that an employer expects participation. Government Social Research's ethical assurance guidance addresses informed participation, enabling participation and minimising harm. It is a government standard, not approval of a commercial project.

2. Settle data use separately from participation

An ethical agreement to be interviewed is not automatically a UK GDPR lawful basis. The ICO's research principles page makes that distinction and says there is no single research basis. Record purpose, necessity, lawful basis, transparency and rights after qualified review.

Recruitment and later promotion are also different. Apply the ICO's direct-marketing planning guidance to the actual channel and purpose. Do not add a participant to a marketing audience because they accepted an interview.

3. Anchor questions to the last relevant post

Ask the participant to choose a recent publication and walk through the brief, source assets, edits, approval, scheduled time, platform response and monitoring. Neutral prompts such as "What happened next?" leave room for unexpected answers. A feature-led question like "Would reminders solve this?" embeds the researcher's preferred explanation.

4. Preserve evidence layers

Label the prompt, participant statement, artefact, researcher inference and unresolved point separately. Do not publish a quotation unless wording, context and attribution were authorised. A document can contradict memory without proving deliberate error. Record the disagreement.

5. Minimise and secure the record

Collect only the account, role and workflow details required for the decision. Use synthetic examples where a live post would expose customers or credentials. The ICO's appropriate safeguards guidance prioritises minimisation and discusses anonymous and pseudonymous data. Set access, encryption, transcript, deletion and incident responsibilities before recording.

6. Report recruitment and missing voices

Maintain counts for invited, ineligible, declined, no response, completed, withdrawn and excluded cases. State field dates, geography and respondent roles. Accessibility barriers, workload or recruitment channel may shape who appears. Without follow-up evidence, the reason for non-response is unknown.

Before analysis, freeze the instrument version and a coding note that defines each field. Have another reviewer code a small set without seeing the first interpretation, then discuss differences. Preserve unresolved disagreement and cases that contradict the main pattern.

Delete contact details, recordings and transcripts on the approved schedule, and keep a minimised audit record of recruitment and decisions. If withdrawal was offered after interview, the participant information must explain its practical deadline and what can still be removed.

Finish with a bounded decision: proceed to a synthetic workflow test, revise the eligible population, or stop. A vivid story is not an England prevalence estimate, and stated willingness is not a purchase.

Before you act

Common questions

What should a candid interview invitation include?

It should name the researcher, commissioner, topic, reason for selection, time, incentive, recording plan and withdrawal route. Avoid implying that an employer expects participation. Government Social Research ethical assurance guidance addresses informed participation, enabling participation and minimising harm, but it is a government standard, not approval of a commercial project.

How should evidence be handled after an interview?

Label separately: prompt, participant statement, artefact, researcher inference, unresolved point; publish a quotation only if wording, context and attribution were authorised.

Record any disagreement: a document can contradict memory without proving deliberate error. Collect only the account, role and workflow details needed for the decision, and use synthetic examples if a live post would expose customers or credentials.

What should the final report say about recruitment?

Maintain counts for invited, ineligible, declined, no response, completed, withdrawn and excluded cases. State field dates, geography and respondent roles. Accessibility barriers, workload or recruitment channel may shape who appears, and without follow-up evidence the reason for non-response is unknown. Finish with a bounded decision: proceed to a synthetic workflow test, revise the eligible population, or stop.