Researching who actually wants a small business grant, and who is merely interested

Audience research for a small business grant should explain who needs the programme, who can qualify, who will try to apply and who may be excluded by its design. Those are different groups. Treating them as one audience can produce an attractive campaign that reaches many people but helps few suitable applicants.

This guide is for grant funders, delivery partners, advisers and publishers working in England. It draws on public-sector research practice, but it does not turn general research guidance into a legal opinion or a guaranteed recruitment method.

What to take away

• Map five separate audiences because treating them as one can reach many people but help few suitable applicants.
• Write down the decision the research must inform before recruiting participants or buying a survey platform.
• Recruit for difference rather than convenience, since samples drawn from successful recipients produce a favourable account of the service.
• Ask about actual events during a recent funding decision rather than abstract questions about what someone might do.
• Keep observations separate from interpretations, because a finding from one round may need analytics or further testing.

Start with the decision

Write down the decision the research must inform. Examples include whether to launch a scheme, which English areas need outreach, what an application form should ask, or why suitable businesses abandon an existing process.

The GOV.UK Service Manual advises teams to turn assumptions into research questions and choose methods that answer the most important questions with proportionate effort. Its planning guidance was written for government services, yet the discipline is useful here: decide what needs to be learned before recruiting participants or buying a survey platform.

A broad objective such as "understand small businesses" will not guide the work. A usable question names the group, behaviour and decision. For example: "Why do eligible manufacturing businesses in the West Midlands start but not submit the application?"

Map five audiences

Businesses with the underlying need

These firms face the problem the grant is meant to address, such as equipment cost, skills gaps or research risk. Some will not meet the formal eligibility rules. Researching them can test whether the programme is aimed at the right problem, but their interest must not be reported as eligible demand.

Businesses that appear eligible

This group fits the published rules based on information available before a formal assessment. Record every assumption. A location, employee count or project date that looks simple in a screening survey may need documentary evidence later.

Applicants

Applicants have taken a measurable step, such as registering on a portal or submitting a form. They are easier to observe than non-applicants, so they can dominate the evidence. That creates survivorship bias: the research misses businesses that never heard about the programme or found the first step inaccessible.

Delivery users

An award brings a second set of tasks. Recipients may need to procure suppliers, maintain records, submit claims and report outcomes. The employee who applies may not be the person who handles these duties.

Intermediaries

Accountants, local business advisers, trade bodies, lenders and specialist consultants may introduce firms to a programme. Their view can explain how information travels, but it should not replace direct research with businesses.

Build an evidence frame for England

Use official business data to understand the population before selecting participants. The ONS business activity, size and location dataset includes enterprise counts by region and industry. Its coverage is based on VAT and/or PAYE records, so an editor must state that it does not represent every informal activity or very small business.

The government's finance and support finder can show how current schemes describe their audiences. Its listings are live and filtered. They do not prove that every relevant business knows about a scheme or can apply.

Add operational evidence where it exists: enquiries, incomplete forms, submitted applications, assessment outcomes, support calls, complaints, award acceptance, claims and project completion. Each stage uses a different denominator. Ten withdrawals among 100 submitted applications is not the same measure as ten withdrawals among 100 awards.

Choose methods that fit the question

Interviews work well when the team needs to understand decisions, language or a complex journey. Observation and usability testing can reveal where an application process fails even when participants say it looks clear. Surveys can estimate how common a defined response is, provided the sample and question design support that inference. Administrative data can describe recorded behaviour, but it rarely explains the reason behind it.

The Service Manual's introduction to user research recommends studying the end-to-end journey, including telephone, post and face-to-face routes rather than only the digital service. This is particularly relevant when a programme aims to reach owners with limited time, confidence or digital access.

Use more than one method when the decision carries material cost. An application-funnel analysis may locate the point of abandonment. Interviews can then explore why it happens. A revised page or form can be tested before it is released widely.

Recruit for difference, not convenience

A sample made entirely from successful recipients will usually produce a favourable account of the service. Include businesses that were rejected, withdrew, did not apply, needed help or completed the process without support. Consider disability, digital confidence, language needs, business age, size, sector and English region where these factors affect the service.

Do not claim statistical representation from a small qualitative sample. GOV.UK's guidance for planning a research round distinguishes small qualitative rounds from surveys and benchmarking that need larger samples. The right number depends on the method and decision, not on a universal quota.

Recruitment criteria should be written before the team sees the candidates. Otherwise it is easy to select articulate participants who confirm an existing view. Keep a record of who was invited, who took part and which important groups were missed.

Ask about actual events

Questions about what a person did during a recent funding decision are usually more useful than abstract questions about what they might do. Ask how they first heard about the opportunity, what they checked, which documents they used, where they hesitated and what happened next.

Avoid selling the programme during the session. A researcher who explains every feature before asking for a reaction has changed the thing being tested. Avoid asking participants to approve a proposed solution when the real task is to understand the problem.

When a participant reports a rule or deadline, treat it as evidence of their understanding, not proof of the rule. Verify the factual point against the funder's current page.

Protect research data

Interview notes, recordings, contact details and survey responses may contain personal data. The GOV.UK Service Manual says teams should collect only what they need, restrict access, set a retention period and keep records of participant consent. Its participant privacy guidance also recommends review by the organisation's data-protection or legal specialist.

Participation consent and the lawful basis for processing are related but not identical questions. The ICO's research processing guidance explains that the appropriate lawful basis depends on the organisation, purpose and context. A private business should not copy the lawful basis used by a government department without its own assessment.

Tell participants what will be collected, why, who will receive it and how long it will be kept. Remove personal details from shared findings unless there is a sound, reviewed reason to retain them.

Analyse evidence without manufacturing certainty

Keep observations separate from interpretations and decisions. "Three participants could not find the eligibility page" is an observation from a particular round. "The navigation prevents eligible firms from applying" is a broader interpretation that may need analytics or further testing.

Look for contradictory cases. One business that completed the task easily may reveal a useful difference in experience, prior knowledge or device use. Do not discard it merely because it weakens a neat theme.

The Service Manual analysis guidance recommends analysing soon after a round and involving observers to reduce the influence of one researcher. Record which evidence supports each finding and which questions remain open.

Turn findings into controlled changes

A finding should lead to a named decision, owner and review date. A wording problem may need a content change and usability test. A location mismatch may require altered eligibility or clearer outreach, which the delivery team may not have authority to approve. A documentation burden may need a policy decision rather than a redesigned form.

Track what changed and what measure should move if the change works. Do not present a rise in enquiries as proof that awards reached the intended businesses. The next measure might be eligible applications, completion, award take-up or project delivery.

Research record

For each round, retain the decision question, participant criteria, recruitment route, discussion guide, consent record, raw-data location, analysis method, findings, limitations, decisions and responsible owner. Store personal data separately from the summary where practical.

Questions readers ask

Is audience research the same as market research?

The methods can overlap, but the decision may differ. Grant audience research often concerns eligibility, access and delivery of a specific programme, while market research may examine demand, competitors or commercial positioning.

Should rejected applicants be included?

Yes, where the research question concerns eligibility, clarity or assessment experience. Recruitment and reporting must protect confidential or personal information.

Can application data replace interviews?

No. It can show recorded behaviour, such as abandonment at a stage, but it rarely explains the cause. Interviews alone cannot show how common the behaviour is either.

Can a funder use customer testimonials as research?

Testimonials may provide leads, but they are selected public statements rather than a balanced research sample. They should not support broad claims about the whole audience.

What must be checked before publication?

Reopen all scheme and privacy sources, confirm geography and dates, remove identifying participant details, and have legal or data-protection claims reviewed by a qualified person.

The audience-research draft has no live internal links and stays on editorial hold.

Before you act

• Write down the decision the research must inform.
• Map the five audiences before recruiting participants.
• State the limits of official business data you use.
• Write recruitment criteria before seeing candidates.
• Avoid selling the programme during research sessions.
• Set a retention period and restrict access to research data.

Common questions

Why can applicants dominate the evidence in grant research?

Applicants have taken a measurable step, such as registering on a portal or submitting a form, so they are easier to observe than non-applicants. That creates survivorship bias: the research misses businesses that never heard about the programme or found the first step inaccessible.

What does the ONS business activity, size and location dataset not cover?

Its coverage is based on VAT and/or PAYE records, so it does not represent every informal activity or very small business. An editor must state that limit when using the dataset to understand the population before selecting participants for research.

How should a private business approach the lawful basis for processing research data?

The ICO explains that the appropriate lawful basis depends on the organisation, purpose and context. A private business should not copy the lawful basis used by a government department without its own assessment, and should tell participants what is collected and why.