Audience research
Part of Reputation management audience research: separate the populations
Reputation management target audience: who to recruit
Define a reputation management target audience by recent task, authority and evidence, with UK GDPR rules, research standards and review-platform sources.
A reputation management target audience is recruitable only when it is defined by a recent task, authority and evidence. It is not everyone who reads a review, nor a demographic profile assembled from public clues.
Use a bounded fictional decision: an England home-services marketplace wants to learn whether its review-correction and response route needs redesign. Its reviews sit on platforms such as Trustpilot, Google Business Profile, Reviews.io and Feefo, and its team answers them through reputation-management software. The England operating unit is the part of that business which handles reviews and complaints for customers in England; Scottish, Welsh and Northern Irish cases need their own rule. No real participant or organisation was researched.
What to take away
- Define a recruitable audience by a recent task, authority and evidence, not by demographics or public clues.
- Build a frame record with evidence owner, extraction date, eligible period, source systems and duplicate treatment before inviting.
- Recruit across completed, abandoned, escalated and corrected journeys to avoid describing only easy experiences.
- Screen on specific task performance and timing, not on positive or negative views, to prevent bias.
- Pause recruitment if contact authority, participation information, withdrawal or data minimisation cannot be assured.
Write separate eligibility rules
Start with the people affected by the route: reviewers, complainants and anyone identifiable in the evidence. Their inclusion rule could require an authorised, recent interaction with the England operating unit. Do not infer residence, disability, ethnicity or buying power from a name, profile, postcode fragment or writing style.
Contact authority is a data-protection question before it is a commercial one. Under the UK GDPR and the Data Protection Act 2018 you need a documented lawful basis for inviting someone — legitimate interests or consent — and the ICO expects a privacy notice that names research as a purpose and a clear route for people to object. PECR adds consent rules for marketing by electronic mail, so a research invitation must stay separate from any later request to leave a review.
Eligibility by participant group
Affected people
- Who
- Reviewers, complainants
- Rule
- Authorised recent interaction
- Avoid
- Inferring traits from profile
Operational users
- Who
- Moderator, handler, comms
- Rule
- Record actual task performed
- Avoid
- Relying on job title
Governance roles
- Who
- Budget, privacy, security
- Rule
- Answer governance questions
- Avoid
- Mixing with experience research
Next define operational users. A moderator may verify source status; a complaint handler investigates the service event; a communications colleague drafts a public response. Record the task each person actually performs. Job title alone does not show access, experience or decision authority.
Finally list decision and assurance roles: the budget owner, workflow owner, procurement contact, privacy reviewer, consumer or advertising reviewer, security owner and accessibility specialist. These participants answer governance questions, not customer-experience questions. The GOV.UK user-research introduction emphasises end-to-end tasks and inclusion, although it is government-service guidance rather than a universal commercial standard.
For commercial work the reference standards are the MRS Code of Conduct and its fair data rules, the ICC/ESOMAR international code and ISO 20252 for market, opinion and social research. They set the expectations on consent, withdrawal, anonymity and the separation of research from promotion that government-service guidance does not cover.
Build the frame before sending invitations
Create a frame record with the evidence owner, extraction date, England rule, eligible event period, source systems, duplicate treatment and known omissions. Source systems usually means the review platform that holds the original review — Trustpilot, Google Business Profile, Reviews.io or Feefo — plus the reputation-management tool that tags, routes and answers it, such as Birdeye or Reputation.com, and the CRM or case system. Use case references rather than a copied contact database where possible, and separate research invitations from later promotion.
Frame record fields
- Evidence owner
- Extraction date
- England rule
- Eligible event period
- Source systems
- Duplicate treatment
- Known omissions
A worked example of the frame record
| Field | What the record holds |
|---|---|
| Evidence owner | The named person accountable for the review, complaint and correction records used |
| Extraction date | The date the extract was taken, so the same frame can be rebuilt later |
| England rule | Which operating unit handled the case, and how cases handled elsewhere are excluded |
| Eligible event period | The window of review, complaint or correction activity that qualifies a person |
| Source systems | Review platform, reputation-management tool, CRM or case system the record came from |
| Duplicate treatment | How one person appearing in several systems is counted once |
| Known omissions | Groups the source systems cannot reach, such as customers who never left a review |
Recruit across completed, abandoned, escalated and corrected journeys if the contact route is appropriate. Otherwise the sample may describe only people who found the process easy. GOV.UK's participant recruitment guidance warns that timing, location and recruitment routes can exclude groups and advises deliberate inclusion of people with access needs.
Record every disposition: invited, unreachable, ineligible, declined, consented, completed, partial or withdrawn. A response from a convenient group is not evidence about non-respondents. ONS explains that survey uncertainty includes both sampling error and hard-to-quantify non-sampling errors such as refusal and inaccurate answers.
Screen on facts, not stereotypes
Ask whether the person performed or encountered the specific task, when it occurred, which operating unit handled it and whether they can discuss it without exposing another person's confidential information. Ask about communication or access adjustments directly and provide an accessible route to answer.
Do not screen for a positive or negative view; that would bias the study and could turn research into review harvesting.
Do not tie an incentive to praise, answering every question or public posting, and do not invite participants to leave a review as part of the study. Fake reviews, and reviews that conceal an incentive, are banned under the Digital Markets, Competition and Consumers Act 2024, which the CMA enforces. A qualified reviewer must assess any proposed incentive and the later use of responses.
Apply a decision test
The research owner should approve a short table containing population, eligibility, exclusion, recruitment source, evidence the person can provide, privacy risk, missing group and stop rule. Recruitment pauses if contact authority cannot be shown — a documented lawful basis under the UK GDPR, a privacy notice that covers research and, for any marketing follow-up, PECR consent — or if participation information is unclear, withdrawal cannot be honoured, or the study would collect unnecessary sensitive details.
The outcome is not a claimed England audience size. It is a defensible set of people who can answer this one decision, plus an honest account of who is absent. If key affected people cannot be recruited safely, narrow the question or use service records and accessibility review instead of pretending the frame is representative.
Before you act
- Write separate eligibility rules for affected people, operational users and governance roles.
- Create a frame record before sending any invitations.
- Recruit across completed, abandoned, escalated and corrected journeys.
- Record every disposition from invited to withdrawn.
- Screen on task facts, not on positive or negative views.
- Pause recruitment if contact authority or withdrawal cannot be honoured.
Common questions
How should you define the population for a reputation research audience?
Start with people affected by the route, such as reviewers and complainants, who had an authorised recent interaction with the operating unit. Then add operational users and governance roles. Do not infer characteristics from names, profiles or postcodes. Each group needs a separate eligibility rule based on task, authority and evidence.
What should a frame record contain before invitations are sent?
A frame record should include the evidence owner, extraction date, England rule, eligible event period, source systems, duplicate treatment and known omissions. Use case references rather than a copied contact database where possible. Separate research invitations from later promotion. This record helps ensure the sample is defensible and omissions are documented.
When should recruitment be paused?
Recruitment pauses if contact authority cannot be shown — a documented lawful basis under the UK GDPR, a privacy notice covering research and, for marketing follow-up, PECR consent — if participation information is unclear, or if withdrawal cannot be honoured. Withdrawal is put into practice by dropping the participant's data from the analysis and deleting the case reference held in the reputation-management tool or review platform. The research owner should approve a short table covering population, eligibility, exclusion, recruitment source, evidence, privacy risk, missing group and stop rule.
