Audience research for property creators with populations and permissions kept apart

Good audience research starts by naming whose decision is being studied. A property video can involve a commissioning organisation, the organisation's staff, an adult viewer, a person seeking a home and somebody visible in the footage. They are not one audience, and permission from one group does not stand in for another.

The scope is research for commercial property media in England, without invented segments, quotations or preferences. No official source measures a single England audience for property creators, so a team must produce its own bounded evidence and report who was missed.

What to take away

Separate the populations before recruitment

Use a population map rather than a broad label such as "property people".

An individual can sit in more than one row. Record the role in each session rather than merging answers. An estate-agent employee commenting on production cannot automatically speak for a home buyer, and a research participant's agreement to an interview does not permit their face or words to appear in an advert.

Write a decision question

A useful question names the proposed action. For example: should an England-based studio offer source-checked neighbourhood explainers to residential estate agencies in two named regions? State the buyer, deliverable, geography, intended adult user, decision date and evidence threshold.

The ONS UK business activity, size and location dataset can provide a frame of VAT and/or PAYE registered enterprises by English region and industry. It cannot say which enterprises commission creators. Its quality and methodology information documents registration, classification and head-office-location limits. Save the table, edition, SIC selection, unit and exclusions before drawing a sample.

For adult housing tasks, the English Housing Survey supplies England estimates under its own survey design. It is context, not a recruitment directory. HM Land Registry Price Paid Data is also unsuitable as a contact list: an address in a transaction record is not permission to approach an occupant or expose a home-security detail.

Design the sample honestly

Define eligibility before invitations go out. Organisational research might require a current role in commissioning or approving media for a specified property activity in England. Viewer research might require an adult currently completing a stated task, such as understanding a letting process, without collecting an exact address or financial position.

Record the sampling frame, recruitment route, dates, invitations and eligible replies. Record completed sessions and withdrawals, and give the denominator for each rate.

Do not claim a convenience sample represents England; GOV.UK Service Manual guidance on finding research participants recommends varied recruitment routes and including people with access needs. It is a public-service research method reference, not a private publisher's legal safe harbour.

Recruit people voluntarily. Do not scrape personal profiles, infer that someone is moving home or invite a resident because their address appears in property data. Avoid recruiting through a client's sales team if refusal could feel costly. If an incentive is offered, state its amount and conditions before consent and keep it unrelated to what the participant says.

Keep four permissions apart

Research participation, data processing, promotional reuse and sales follow-up are different decisions.

The GOV.UK guide to informed consent for research says participants must understand the purpose of the research and that they can withdraw their consent at any time. That guide is written for government research teams, so a private operator still needs its own legal and ethics review.

The ICO explains that consent to participate in research is distinct from consent as a UK GDPR lawful basis. A privacy reviewer must document the controller, purpose, lawful basis, recipients, processors, retention and rights for the proposed project. Do not add a pre-ticked marketing permission to the research form.

If an interview extract might become a testimonial, seek a separate, specific publication decision after the participant has seen the proposed words, image, context, channels and duration. CAP's current testimonials and endorsements guidance says marketers need evidence that a testimonial is genuine and permission to use it. Research consent alone is not that advertising record.

Sales contact is another step. The ICO's guidance on identifying direct marketing distinguishes genuine market research from messages containing promotion or designed to enable later marketing. Ask separately, and only after privacy and PECR review, if somebody wants future commercial contact.

Protect property and participant information

Research about homes can reveal addresses, routines, occupancy, access codes and alarm or camera locations. It can reveal planned absences, disputes, finances and on-screen documents, but most studies do not need those details.

Under the ICO's data-minimisation guidance, each field needs a rational link to the stated purpose, and the project should hold no more than that purpose requires. The ICO warns this page is under review following the Data (Use and Access) Act changes.

Use participant codes in notes, and store recruitment contacts separately from session material. Brief observers, restrict downloads, set a deletion date and remove direct identifiers from reports.

GOV.UK's detailed guidance on participant privacy and research data covers processor contracts, access, retention and anonymised reporting.

A location can stay identifiable after a person's name has gone; inspect backgrounds, documents and metadata before sharing a clip.

Use interviews and surveys for different jobs

Interviews can reveal language, sequence and exceptions. Ask about a recent real task rather than requesting a prediction about what someone would buy. Keep the question neutral, record the participant code and role, and distinguish their account from an independently verified property fact.

A survey is useful when the response options and population are already understood. Pilot it with people using different devices and assistive technologies. Record starts, eligible respondents, completions, item non-response, duplicate-handling rules and exclusions. A percentage without its question, denominator and sampling method is not a finding.

Tool procurement comes after the data plan. The platform's public product page cannot decide the controller-processor roles, lawful basis or acceptable data location. Review the precise service tier, contract, sub-processors, access controls, export, deletion, incident terms and accessibility evidence. Run a test with synthetic data only before any participant record enters the system.

Build evidence summaries, not fictional personas

A property creator persona should summarise a documented pattern within one population. Give it a functional label such as "regional agency content approver", not a made-up name, photograph or quotation. List the sample dates, eligible denominator, supporting cases, contrary cases, source type, geography, access requirements and confidence limit.

Keep observation and interpretation in different fields. "Four of seven interviewed approvers required a documented correction route" would be an observation only if those interviews actually occurred and the denominator is reported. "A correction route should be part of the offer" is a team decision.

Do not infer market share, income or preference from a competitor's search position or reviews. Companies House company information can help identify a legal entity, but it does not prove service quality or the truth of marketing claims. The CMA's online consumer reviews work shows why review provenance and platform controls need separate examination.

Analyse contradictions and decide

Before coding, define the unit of analysis: organisation, participant or response. Use a codebook linked to the research question. Keep a negative-case log for evidence that challenges the proposed service. A quote is illustrative, not a vote, and repeated comments from one organisation do not enlarge the sample.

Report recruitment source, eligibility, dates, geography, sample achieved, withdrawals, missing responses and known bias. Separate what participants said, what official records establish and what the editor recommends. Do not translate UK-wide evidence into an England result or generalise beyond the sampled roles.

Property facts and promotional conclusions require their own review under the property creators rules and ethics. An interviewee cannot authorise an unsubstantiated property description. Mortgage or investment language must go to a named FCA financial-promotion reviewer. Any suggestion that the creator will introduce buyers, negotiate or conduct estate-agency work needs a property solicitor's decision. Copyright and appearance rights for quotations, images and recordings also remain on hold.

End with a decision record. Proceed only if the defined population supplied enough relevant evidence, important contradictions were addressed, the proposed deliverable can be reviewed safely, and the team can explain who was excluded.

Otherwise narrow the audience, change the question or stop. Ethical research can show an offer should not launch; it is not a device for manufacturing demand.

Before you act

Common questions

Why can one group's permission not cover another group?

A property video can involve a commissioning organisation, its staff, an adult viewer, a person seeking a home and somebody in the footage. These are not one audience, and permission from one group does not stand in for another. An estate-agent employee cannot automatically speak for a home buyer.

What can the ONS business activity dataset tell a property creator?

It can provide a frame of VAT and/or PAYE registered enterprises by English region and industry. It cannot say which enterprises commission creators. Its quality and methodology information documents registration, classification and head-office-location limits, so save the table, edition, SIC selection, unit and exclusions before sampling.

When can an interview extract become a testimonial?

Only after a separate, specific publication decision, once the participant has seen the proposed words, image, context, channels and duration. CAP guidance says marketers need evidence a testimonial is genuine and permission to use it. Research consent alone is not that advertising record.