Women founder funding audience research: inclusive recruitment and decision-led analysis

Good audience research begins with a funding decision that happened, nearly happened or was deliberately avoided. It does not begin by assigning women founders a shared level of confidence, appetite for risk or need for mentoring.

For an England-focused service, the aim is to understand how route, stage, sector, place and personal circumstances shape discovery, eligibility, evidence, terms and outcomes. Gender belongs in the design, but it is not the entire segment.

This guide sets out a practical mixed-method approach for a publisher, programme, adviser or marketplace. It is not permission to collect every personal detail a participant is willing to share. The research purpose, lawful processing and participant welfare need their own review.

What to take away

Turn the product question into research questions

Start with a decision the organisation can make. Examples include:

Write what evidence would change that decision. "Learn about female founders" is too broad. "Identify where pre-trading founders in three English regions stop during route selection, and why" gives the team a population, behaviour and next action.

Separate exploratory and measurement questions. Interviews can reveal why a founder rejected an offer; a suitable survey or administrative dataset may estimate how often that reason appears in a defined population. One cannot automatically perform the other's job.

Define people through behaviour and context

Use a funding episode as the entry criterion: considered external finance, started an application, sought terms, pitched, received a decision or managed an award. Add business stage, route, amount, purpose, legal form and region.

Ask participants how they describe their founder and gender identity where this is relevant and proportionate. Explain why the field exists and allow non-response. Do not infer gender from a name, voice, photograph or company biography merely to fill a dashboard.

State how the study treats women-led, women-owned, women-founded and mixed teams. If a live scheme uses its own eligibility test, retain that exact language rather than rewriting it as the study definition.

The British Business Bank's Small Business Finance Markets report warns that demographic data collection is incomplete and inferential methods may underreport women-owned and women-led companies. Missingness is therefore an analytical finding, not an invitation to guess.

Build a sampling frame that includes absence

A customer list contains users, not the whole potential audience. Accelerator alumni overrepresent people who found and entered a programme. Investor portfolios contain successful transactions. Social-media recruitment can favour visible, digitally confident founders.

Build the frame across several routes:

The GOV.UK Service Manual's participant recruitment guidance advises teams to recruit actual or likely users through varied approaches. It calls for inclusion of disabled people, people using assistive technology and those with limited digital skills or literacy.

For women founder funding, include successful and unsuccessful applications, partial offers, self-funding, discouraged non-applicants and funded delivery. Vary English region, sector and company age. Add intersectional groups where the question, recruitment and sample allow responsible analysis.

Offer remote, telephone and suitable in-person options. Ask about accessible formats, interpreters, support, travel and timing. Caring work and business operations may make a standard weekday slot exclusionary. Pay a fair incentive for research time, not favourable opinion.

Interview a recent funding journey

Use a timeline. Begin when the need for cash became visible, then move through route discovery, eligibility, preparation, contact, decision and aftermath. Ask what the participant did and saw before asking how she felt about it.

Strong prompts include:

Avoid "Did a lack of confidence stop you applying?" It embeds the explanation. Ask "What led to the decision not to apply?" and follow the evidence.

Do not turn an interview into advice. If a participant asks whether to accept finance, separate the question from the session and direct her to an appropriate independent or authorised source. The researcher should not exploit trust to sell a product.

Compare cases without flattening them

Review notes soon after each session. Record observations under a stable journey: trigger, discovery, eligibility, evidence, terms, decision and consequences. Then add interpretation and confidence separately.

Look for repeated patterns and sharp contrasts. A loan applicant may stop because repayments threaten household resilience; an equity founder may reject a term because of governance; a grant applicant may miss scope. Calling all three "fear of finance" would destroy the decision value.

The Service Manual provides a practical process for analysing a research session, including team observation and evidence-based findings. Adapt the method to the organisation's risk and research governance.

Invite a second analyst to challenge themes and look for disconfirming cases. Keep quotes linked to protected source notes, but publish them only with permission and after checking whether details could identify a founder.

Design a survey after discovery

Use interview language to write clear, neutral questions. Define the period, route and unit. "Have you had trouble getting funding?" produces less useful evidence than separate questions about consideration, application, decision, amount, timing and terms within a stated period.

A practical questionnaire can cover:

Randomise answer order where order effects matter, include a write-in option for incomplete lists and separate "none" from "prefer not to say". Pilot the full journey on mobile, keyboard and assistive technology. Time it with real participants.

Do not overload the survey with demographic fields in the hope that every cross-tab will become useful. Write the analysis plan and minimum subgroup sizes first. Small cells may be unreliable and identifiable.

Choose the collection tool deliberately

A hosted form offers routing and export. An interviewer-administered survey supports participants who prefer conversation. A panel can fill quotas quickly but represents people enrolled in that panel. An in-product question captures immediate behaviour while excluding non-users. A controlled spreadsheet may suit a small discovery pilot but becomes fragile at volume.

Compare accessibility, sample control, validation, hosting, processor terms, retention, deletion, export and cost. Test the tool rather than accepting an accessibility claim in marketing.

Maintain one approved questionnaire version and a change log. If wording changes during collection, preserve the old response population and explain the break. Do not merge answers to materially different questions.

Handle personal data and participant consent

Give participants a clear information sheet covering the study, organisation, recording, incentive, storage, access, reuse, quotation and contact route. Obtain voluntary agreement to take part. Keep that ethical participation decision distinct from the organisation's UK GDPR lawful basis.

The ICO's research-processing guidance explains that participation consent does not automatically mean consent is the appropriate lawful basis for personal-data processing. The right basis depends on purpose, organisation and context. The ICO also notes current research guidance is subject to review following legal change, so it needs a fresh publication-day check.

Create a data map for recruitment details, recordings, transcripts, survey records, demographic fields and incentives. Restrict raw access, separate contact information from responses, set retention dates and document deletion. Pseudonymisation reduces risk but is not necessarily anonymisation.

Financial journeys can reveal debt, income, credit, business distress and third-party identities. Collect only what answers the research question. Redact account details, investor names and unrelated personal information from working notes.

Build personas that retain evidence

Create a persona only when research reveals a stable, distinct decision pattern. Replace fictional biographies with an evidence card containing funding job, stage, acceptable obligations, missing proof, access needs, behaviour, sources and confidence.

Include a boundary case. A persona for a first-time debt comparer should state that an equity-ready founder with similar demographics belongs elsewhere. This stops teams using a label as a universal script.

Review personas against new rounds and service data. Retire any card that no longer predicts a different content or product decision. Never use a persona as a proxy for creditworthiness, investment quality or protected characteristics.

Research competitors through their actual model

Group alternatives before comparison: lender, investor, grant programme, accelerator, adviser, marketplace, publisher or community. Record the paying party, source of capital, legal entity, audience definition, eligibility, price, revenue model, claims and post-funding service.

Verify companies at Companies House and search the FCA Register when relevant activity may require authorisation. These records establish identity and status, not customer quality.

Open every named scheme. In the UK, women founders commonly encounter Start Up Loans (British Business Bank), Innovate UK's Women in Innovation, British Business Bank investment programmes and NatWest's Back Her Business. A closed award may still rank well in search. The Women in Innovation 2025 to 2026 opportunity clearly carries a closed status and historic deadline; it can inform programme research but not a current-opportunities page.

Ask providers to correct factual descriptions. Disclose paid relationships and never use a false application or copied protected content to complete desk research.

Connect findings to decisions

Every research round should end with a short evidence statement:

Combine qualitative findings with funnel data only when populations and definitions align. Ten detailed interviews can reveal an exclusion mechanism. They cannot estimate its prevalence across England. A large survey can estimate a defined response pattern but may not explain the cause.

Publish the method beside any headline: recruitment routes, dates, sample, questionnaire, incentives, geography, missing data and analytical limitations. Protect identities in small cells and quotations.

A six-week England research cycle

Week one defines the decision, audience, data map and safeguards. Week two recruits contrasting journeys and pilots materials. Weeks three and four conduct interviews and same-day analysis. Week five pilots the survey or service change. Week six synthesises evidence, makes a decision and identifies the excluded group for the next round.

Do not wait for a perfect national sample before improving an obvious barrier. Also do not market an exploratory finding as representative. Confidence should match the method.

Research governance should name a responsible researcher, an ethics or inclusion reviewer, and the organisation's data protection lead. In the UK, the Market Research Society's code of conduct provides a recognised baseline for participant welfare, consent and transparency. Findings that could influence financial services should be reviewed by qualified oversight before publication.

Before you act

Common questions

Why should research begin with a funding decision rather than a founder trait?

Good audience research begins with a funding decision that happened, nearly happened or was deliberately avoided. It warns against assigning women founders a shared level of confidence, appetite for risk or need for mentoring, because gender belongs in the design but is not the entire segment.

How should researchers recruit participants beyond existing customer lists?

Build the frame across general and specialist networks, local authorities, universities, community groups, finance providers, open calls and people who started but did not complete a service journey. Include successful and unsuccessful applications, partial offers, self-funding, discouraged non-applicants and funded delivery.

What should happen before running a survey on women founder funding?

Design the survey after discovery, using interview language for clear, neutral questions. Write the analysis plan and minimum subgroup sizes first, because small cells may be unreliable and identifiable. Pilot the full journey on mobile, keyboard and assistive technology, and time it with real participants.