Branding agency buyer personas: test who decides before you shortlist

Branding agency buyer personas are useful only as hypotheses about roles in a decision. A fictional name, stock photograph and confident list of "pain points" can hide the absence of research. Start with an evidence card that can be disproved, and do not describe any role as typical until a suitable study supports that statement.

What to take away

Use roles, not imaginary biographies

A branding assignment may involve several decision positions. In UK organisations these commonly include a marketing or brand lead who drafts the brief and shortlists agencies, a marketing director who owns the budget, a founder or managing director who signs off, and colleagues in procurement or finance who handle terms and supplier onboarding. Treat each as a hypothesis: the economic buyer can approve expenditure, a project sponsor defines the desired change, and an implementation owner must apply the assets after handover. Titles vary, so record what the person actually decides rather than the title on an org chart.

UK trade bodies sit on both sides of this market. The Design Business Association and the Institute of Practitioners in Advertising represent agencies and their professional standards, while the Chartered Institute of Marketing represents many of the buyers who commission brand work. Note membership as a fact about the organisation, not as evidence about an individual's motives.

People affected by the output may encounter the name, language or design without taking part in procurement. These are workflow hypotheses, not portraits of particular people or claims every organisation uses the same titles.

GOV.UK Service Manual guidance on planning user research begins with the problem, assumptions, user groups and decision need. Apply that discipline to each proposed role. Write what the team needs to learn, which recent event would reveal it and what evidence would overturn the hypothesis.

Give each card an evidence ledger

A usable card should contain:

Leave demographic fields blank unless they are relevant, collected proportionately and analysed with a defensible sample. Organisation size, sector and region may help sampling, but they do not reveal an individual's motives. A role can also move between people during procurement, approval and implementation.

Recruit for contradiction

The GOV.UK page on finding participants recommends recruiting actual or likely users, accounting for access needs and watching for recruitment bias. Seek people who used another route, paused a project or rejected external help, not only satisfied agency clients. Record who declined or could not be reached.

The MRS Code of Conduct, current in its 2023 edition, covers participant wellbeing, transparent professional practice and inclusive-data reporting for members and company partners. It is not legislation and does not make a sample representative. Record membership status rather than implying the code applies by default.

Keep inference separate from observation

During analysis, label four columns: what the participant did or supplied; what they said; the researcher's interpretation; and what remains unknown. GOV.UK's guide to analysing research sessions describes moving from raw records to findings, but qualitative themes do not establish population prevalence.

Research files may contain commercially sensitive and personal information. In the UK, personal data is governed by the UK GDPR and the Data Protection Act 2018, and the Information Commissioner's Office (ICO) is the regulator that enforces them. Follow the agreed purpose, access and deletion schedule, informed by GOV.UK guidance on participant data and privacy, with qualified privacy review.

Each research round needs a field record. Name the interviewer, note taker and analyst; set start and end dates; keep question order and version; record which proposed roles responded or did not respond. List the incentive, recruitment supplier, commissioning client and any conflict.

If observers attended or recordings were made, attach the participant-facing notice and permission record; those fields show how the evidence arose without turning the persona into a fictional individual.

Promote a persona hypothesis into a planning input only after several relevant sources support the same decision pattern and meaningful counter-evidence has been considered. As a working rule, not a statistical threshold, count "several" as at least three independent sources from at least two different organisations, plus one recorded contrary case. Otherwise keep the card marked unvalidated and avoid using it to target people or predict demand.

Worked example: one card tested

For example, the hypothesis is that a marketing lead who drafts the brief also decides which agencies are shortlisted. Evidence one: a buyer says they wrote the brief and chose three agencies. Evidence two: a second buyer says the founder named an agency first. Evidence three: a procurement lead says they ran the shortlist.

Only the first account supports the hypothesis; the second and third contradict it, because the founder and procurement each shaped the shortlist. The card stays unvalidated and is reworded: shortlisting is shared between the marketing lead, the founder and procurement. The example is illustrative, not observed.

If you are the buyer, use the card to decide who must be in the room when you read agency pitches, directory listings and client reviews, and ask which role wrote each review or ran the shortlist behind it. The brief template and the shortlisting method show how a tested role card feeds a real agency choice.

Before you act

Common questions

What should a usable persona card contain?

It should name the decision role and qualifying event, plus the geography and organisation eligibility for the market you serve. Add the sampling frame, recruitment route and behaviour to examine. Record evidence observed, source date, analyst, rival explanations and unresolved questions, then the next decision, owner and review date.

Why recruit people who rejected external help?

The GOV.UK guidance on finding participants recommends recruiting actual or likely users and watching for recruitment bias. Seeking people who paused a project, used another route or rejected external help tests the hypothesis against contradiction, rather than confirming it only with satisfied agency clients.

When can a persona hypothesis become a planning input?

Only after several relevant sources support the same decision pattern and meaningful counter-evidence has been considered. Until then, keep the card marked unvalidated and avoid using it to target people or predict demand.