How a video marketing agency builds buyer personas from evidence

What to take away

Decide which buyer groups you must serve before commissioning research, because that choice sets the sample, questions and budget.
Build each persona from named sourcesONS leisure and tourism data, your own CRM records, and interview notes.
Treat audience data as personal information where it identifies someone, so consent, retention and access rules apply from day one.
Review personas every six to twelve months, because buying committees move faster than most agencies expect.

Start with the decision

Write down the decision the research must inform. It might be which sector to specialise in, which service to lead with, or which buyer signs the contract.

A video production company chasing household brands needs different evidence from one serving local councils. Naming the decision first stops the project drifting into general curiosity.

Keep the scope to one decision per research round. Two decisions double the sample and halve the usefulness of each finding.

Map the buying group

Most business purchases involve more than one person. List the roles that touch the decision: the buyer who feels the problem, the one who sets the budget, and the person who signs.

Buying group roles and concerns

Buyer

Gain: coverage
Fear: wasted spend
Example: marketing manager

Budget setter

Gain: control
Fear: overspend
Example: finance lead

Signer

Gain: results
Fear: risk
Example: director

For each role, note what they gain and what they fear. A marketing manager may want coverage; a finance lead may want a fixed price.

This map becomes the spine of the personas you build later. The guide to video marketing agencies audience research sets out how the pieces fit together.

Choose data sources

Use three kinds of evidence: desk data, your own records, and direct conversations.

Three evidence sources

Desk datamarket size and behaviour
ONS leisure and tourism statistics
Own recordsenquiry source, sector, deal size, churn
Direct conversationssmall and structured

Desk data covers market size and behaviour. The ONS publishes leisure and tourism statistics that help anyone planning travel or hospitality video content.

Your own records show who actually buys. Pull enquiry source, sector, deal size and churn for the last two years.

Direct conversations fill the gaps, so keep them small and structured rather than wide and vague.

Run interviews properly

Interview six to ten people per buyer role, mixing recent buyers with lost prospects. Ask about the last purchase, not hypothetical ones.

Record the language they use. If buyers say "explainer film" rather than "brand video", your website copy should match.

Recruit from your own pipeline first, then use a panel to fill gaps. Confirm consent before you record anything.

A companion piece on video marketing agencies customer interviews covers question design and note-taking in more detail.

Handle data lawfully

Audience research often collects names, job titles and opinions. Information that identifies a living person counts as personal information under UK GDPR, as the ICO explains in its guidance on what personal information is.

That means a lawful basis, a retention period and a route for people to ask what you hold. Keep interview notes in one place, with access limited to the project team.

If you hire researchers or editors, check the rules in the GOV.UK section on employing people before you advertise the role.

Set the sample size

Match effort to decision risk. A small agency testing one new sector can work with eight interviews plus desk data.

A team planning a full repositioning, for example an agency paying £400 a month for panel access, should expect forty or more responses and follow-up calls.

Write the sample target into the brief so nobody argues about it mid-project.

Glossary

Persona: a written profile of one buyer role, built from evidence.
Buying group: everyone who influences a purchase decision.
Desk research: analysis of existing published data.
Panel: a recruited group paid to answer research questions.
Lawful basis: the UK GDPR condition that allows you to process personal data.

Build and test the personas

Turn each role into a one-page profile: goals, objections, triggers, and where they look for suppliers.

Add a confidence rating to every claim, marking it high, medium or low. Low-confidence claims become the next research questions.

Test the personas with the sales team. If a profile does not match the deals they see, the profile is wrong.

Common questions

How many personas should a small agency keep?

Two or three is usually enough for a team under ten people. More than that and nobody remembers them, so the research stops influencing briefs.

Can we reuse client data for audience research?

Only where your lawful basis and privacy notice cover that use. Marketing consent does not automatically permit research, so check before you repurpose a list.

How often should personas be refreshed?

Review annually, or sooner after a major shift in your main sector. Set a calendar reminder so the review does not depend on memory.

What if the research contradicts our positioning?

Treat it as a finding, not a failure. Note the gap, then decide whether to adjust the offer or the audience.