Audience research
Researching finance audiences without asking about anyone's wealth
Research England finance audiences with ethical recruitment, safe questions and traceable evidence while avoiding personal advice and false precision.
Audience research for finance content should begin with an adult task, recruit people who recently faced it and collect only the information needed to improve the explanation. It should not ask participants to reveal balances or recommend products to them. Public statistics frame the problem; interviews and surveys test a particular editorial decision.
England is the operating focus here, but much of the strongest consumer evidence and financial regulation is UK-wide. Label the geography of every finding and never turn a small English sample into a national claim.
What to take away
- Write the editorial decision before you draft any research question or recruit anyone.
- Use official data to frame the problem, not to invent a persona or segment.
- Recruit people who recently faced the task, not people chosen by marketing demographics.
- Ask about the process a person followed, never about their balances or products.
- Build personas only from repeated coded evidence, and retire them when later work disproves them.
Write the decision before the questions
Research becomes vague when a team starts with "understand our audience". State what must be decided after the round. The choice might be whether a pension glossary needs a worked example, whether readers can check a lender's legal identity, or whether an email format is accessible to people using a screen reader.
The GOV.UK Service Manual recommends capturing and prioritising research questions and turning team assumptions into questions. Use one primary question and a few supporting ones. Remove anything that cannot change an editorial or service decision.
Write a harm boundary at the same time. The session will not assess a participant's financial position, say which product suits them or collect evidence for a sales lead. Prepare an approved signpost if someone asks for regulated, debt or crisis support.
Use official data without inventing a segment
The FCA's Financial Lives 2024 hub publishes reports, tables and technical material from a survey of 17,950 UK adults. The evidence can suggest groups whose experiences differ by capability, resilience, product use or access. It cannot identify the followers of a proposed creator.
Choose a table that matches the reader task and read its base, question wording and exclusions. Record that it covers the UK unless the table provides a valid country or regional split. A difference worth investigating is not yet a persona.
Ofcom's 2025 news-consumption research can inform channel questions because it measures UK adults' use of online and social routes. It covers news across subjects, not demand for personal-finance content. Treat it as distribution context only.
Desk research should end with a recruitment hypothesis: which adults faced the task, when, and what range of access or confidence needs may change the answer.
Recruit people who can answer the question
GOV.UK guidance on finding participants says research should include actual or likely users and account for disability, digital skills and literacy. Write screening criteria from the task, not from marketing demographics.
For a study about checking a mortgage communication, recent receipt of that type of document may matter. A participant's exact rate, lender or balance probably does not. Ask the least intrusive question that establishes eligibility.
Recruit through more than a creator's existing following when the decision concerns new readers. Current subscribers are useful for testing a live service, yet their familiarity can hide onboarding and trust problems. Record every recruitment channel, incentive and exclusion.
The Service Manual suggests small rounds of four to eight participants for interviews or usability work, while surveys and benchmarks generally need far more respondents. This is method guidance, not a universal sample formula. Statistical requirements depend on the population, design, estimate and precision claimed.
Prepare consent and data handling first
Money discussions may reveal hardship, health, employment or vulnerability. The GOV.UK page on research data and participant privacy recommends explaining collection and use early, obtaining informed participation consent, limiting access and deleting details when they are no longer needed.
Participation consent and a UK GDPR lawful basis are separate questions. The ICO's research provisions guidance makes that distinction and notes the extra conditions for special-category data. A privacy professional should review the actual study.
Prepare an information sheet covering purpose, activities, recording, observers, storage, reporting, withdrawal and incentive. Store contact details separately from session notes. Use participant codes and set deletion dates before recruitment begins.
Do not reuse research responses for advertising, testimonials or individual follow-up sales unless a separate lawful and ethical process supports that use. The simplest safe choice is to keep research and marketing systems apart.
Interview the process, not the person's wealth
Ask about a recent task in sequence. What triggered it? Which document appeared? What did the person try? Which term or step caused trouble? How did they decide that a source was genuine? What would they expect a creator to explain, and when should the creator refer elsewhere?
Use fictional or redacted examples for usability work. Never ask someone to open live banking, pension or credit accounts in a session. If a participant volunteers identifying details, stop the recording if necessary and follow the agreed removal process.
GOV.UK advice on taking research notes distinguishes observations from interpretations. A note such as "participant returned to the search result to check the firm name" is evidence. "Participant does not trust banks" is an interpretation unless the person said it and the context supports it.
Do not promise anonymity when an unusual story could identify someone indirectly. Explain how quotations will be selected, paraphrased or withheld.
Use surveys for defined quantities
A survey is suitable when answer categories are understood and the team needs a number for a stated population. It is a poor discovery tool for a task the editor does not yet understand. Run interviews or cognitive testing first so respondents are not forced into misleading options.
Compare survey routes by recruitment control, duplicate prevention, accessibility, data location, export, deletion and weighting. An open social poll can test wording but cannot represent England. A screened panel may support a defined estimate only when its sampling and exclusions are documented.
Avoid collecting special-category or detailed financial information because it might become analytically useful. The ICO says special-category processing must be necessary for a specific purpose and proportionate.
Publish the questionnaire, fieldwork dates, eligible population, recruitment source, achieved sample, incomplete-response rule and limitations beside any figure. Separate descriptive results from causal claims.
Build personas only after analysis
Personas should summarise repeated evidence about a task. They should not begin as fictional biographies. Avoid stock photographs, invented salaries and quotations assembled from several people.
The Service Manual's analysis method recommends extracting observations, grouping them and deciding findings with the team. Maintain traceability from a persona field back to coded evidence. Note how many participants supported it and where the sample is thin.
A useful persona records the trigger, present route, barrier, access need, trusted source and safe editorial response. It also lists contradictions. Two adults at the same life stage may need different material because their task and experience differ.
Update or retire the persona when later work disproves it. It is a working design aid, not a description of every person in a demographic group.
Research competitors through public evidence
Fix a topic, date range and sample before reviewing other creators. Record public titles, claims, source links, disclosures, correction routes and calls to action. Do not infer revenue, audience residence or customer outcomes from engagement counts.
Financial content needs an additional verification layer. FCA guidance explains how to check a firm and its permissions. Resolve the legal entity behind a named product or partnership. Authorisation does not mean that every product, promotion or statement has FCA approval.
Commercial influence should be visible in the audit. The ASA's social-media advertising guidance covers affiliate material and prominent identification. Use the page as an assessment criterion, then send any legal conclusion to a qualified reviewer.
Competitor research should produce an unanswered reader task or a standard the publication must meet. It must not produce copied wording, a lookalike structure or a league table based on subjective impressions.
Turn findings into a publish or stop decision
After each round, separate observations, findings and actions. Count participants only within the declared sample. Record dissenting cases and missing groups rather than forcing a clean story.
Approve an article when eligible adults share the task, the proposed material improves a specific step, suitable primary sources exist and participants understand the boundary between information and personal advice. Assign an owner and date for each resulting change.
Pause when recruitment relies on intrusive data, when the sample comes only from loyal followers, when a finding cannot be traced to evidence or when the desired response would require an individual recommendation. More responses will not repair a badly defined question.
Publication record
The research note should state who planned and ran the work, why it was done, the dates, locations, recruitment routes, incentives, participant criteria, method, privacy controls and limitations. Name the analyst and qualified reviewers. Do not invent participant quotations or round percentages from a qualitative sample.
Before publication, reopen every regulator and method source. Confirm that financial-promotion, privacy and advertising guidance remains current. Keep suggested internal destinations as article IDs until those pages have final URLs.
Before you act
- State the decision the research must inform.
- Write a harm boundary and an approved signpost.
- Check each table's base, wording and geography.
- Prepare consent, privacy and deletion plans first.
- Use fictional or redacted examples in sessions.
- Publish the questionnaire, sample and limitations with any figure.
Common questions
How many participants does the guidance suggest for interviews?
The Service Manual suggests small rounds of four to eight participants for interviews or usability work. Surveys and benchmarks generally need far more respondents. This is method guidance, not a universal sample formula, because statistical requirements depend on the population, design, estimate and precision claimed.
Can I reuse research responses for marketing or testimonials?
No. Do not reuse research responses for advertising, testimonials or individual follow-up sales unless a separate lawful and ethical process supports that use. The simplest safe choice is to keep research and marketing systems apart, and to explain in the information sheet how quotations will be selected, paraphrased or withheld.
What should a persona contain?
A useful persona records the trigger, present route, barrier, access need, trusted source and safe editorial response. It also lists contradictions. Maintain traceability from each persona field back to coded evidence, note how many participants supported it and where the sample is thin, then update or retire it when later work disproves it.
In this guide
- Defining a finance audience by money task and trigger, not age or income aloneDefine a finance audience by a money task and its trigger, use public data, recruit for range, protect participants and test the definition before commissioning.
- Finance creators buyer personas: an evidence-led approachBuild finance creators buyer personas from observed evidence, with a UK partnership example and safeguards for sensitive data.
- Interview prompts for finance audiences that never invite an advice requestSeven interview prompts help England finance creators study adult money tasks without soliciting advice requests or collecting avoidable sensitive data.
- Finance creators survey tools: four routes comparedCompare four finance creators survey tools for UK audience evidence, each weighed by privacy exposure, recruitment control and bias risk before you buy.
- Auditing rival finance content for sources, permissions and disclosuresAudit finance content competitors in England through public claims, sources, permissions, disclosures and corrections without copying their work.
