Media buying agencies audience research for briefs

This guide to media buying agencies audience research separates two subjects that are often muddled. The agency needs to understand the organisations and people who may appoint it. It also needs evidence about the public an advertiser wants to reach. Those are different populations, with different recruitment, data and measurement methods.

A useful programme starts with a decision. It might determine which advertiser segment the agency can serve, what a client brief is missing or which media mix can reach a defined public. Beginning with a decision prevents a tool, persona template or attractive dataset from taking over the study.

A working media research brief template

Set out the research brief so that interviews and audience evidence answer a defined planning question. Include:

What to take away

Distinguish the buyer from the campaign audience

The buyer audience includes people who initiate, approve, use or block an agency appointment. Marketing, procurement, finance, analytics, legal and privacy roles may all matter. The relevant combination changes with budget, sector and operating model.

For a media-buying agency directory shortlist, apply the same buyer-research questions to each organisation. Examples to research include OMD, Wavemaker, Mindshare, Zenith, Carat and Starcom; these are shortlist examples, not a ranking or recommendation.

The campaign audience consists of people or organisations an advertiser seeks to influence. Media research may describe their viewing, listening, reading, travel or online use. A campaign audience can contain millions of people while the agency's buyer audience contains a small number of decision makers. One dataset cannot stand in for the other.

Write two separate questions. "Why do English marketing teams outsource cross-channel buying?" is a buyer question. "Which media can reach adults responsible for small-business software purchases?" is a campaign question. The first may need interviews and procurement evidence; the second may need audience currencies, survey data and campaign records.

Frame buyer research around real events

Broad questions invite rehearsed opinions. Ask participants to reconstruct a recent review, brief, approval, campaign problem or supplier exit. Follow the order of events, the documents used and the people who intervened.

Government Digital Service guidance on planning user research advises teams to define research questions, user groups and suitable methods. Its setting is UK government services, so an agency should adapt rather than claim the guidance validates a commercial sample.

Recruit actual or likely buyers across the relevant roles. The companion GDS guidance on finding participants warns about recruitment bias and excluding groups. Publish who qualified, how people were contacted, whether incentives were offered and which prospective buyers remain absent.

An interview guide can cover the trigger for agency help, the evidence needed to approve a supplier, operating friction, money and access, and the cost of switching. Invite examples. Do not ask participants to speculate about an entire industry from one appointment.

Turn interviews into evidence

During analysis, preserve the event and role attached to each observation. A procurement concern about audit access should not become a general claim that "clients demand transparency" without scope. A channel manager's frustration may reveal a workflow issue rather than a market trend.

The GDS page on analysing a research session recommends team involvement and a disciplined route from observations to findings. Keep raw notes, coded observations, interpretation and service decision distinct. Report a pattern only within the recruited sample.

Personas should emerge after this work. A sound persona records the role's decision, authority, inputs, proof requirements and likely failure consequence. Remove invented biographies and decorative preferences. Where two participants in the same role disagree, preserve the condition that explains the difference instead of averaging them into a fictional middle.

Select audience evidence by medium

For campaign planning, each medium has its own measurement arrangements. Barb describes a hybrid approach that joins people-based panel data with device-based census data for television and video. Its data support planning and buying within the UK television ecosystem.

RAJAR is the official body responsible for UK radio audience measurement. PAMCo governs the published-media currency in Great Britain and integrates readership research with digital audience figures. Route supplies Great Britain out-of-home audience data for adults aged 15 and over. For online audiences, UKOM's measurement currency is Ipsos iris. JICMAIL provides the UK joint industry currency for direct mail, including door drops. Cinema planning also needs a channel-specific measurement source, such as Comscore/Rentrak; check its UK coverage and definitions before comparing it with other audience currencies.

These sources should not be forced into a single spreadsheet without definitions. Covered population, geography, time basis, reach, frequency and access differ. England is contained within some Great Britain or UK datasets but is not automatically a published cut.

Ofcom supplies broader public evidence. Its Adults' media use and attitudes research covers UK adults aged 16 and over, with questionnaires and technical material signposted. This can help challenge assumptions about media behaviour. It does not replace a trading currency or forecast an advertiser's campaign.

Ofcom's Online Nation examines what people in the UK do online and their experiences. Platform usage can suggest where further planning work is needed, but user reach is not the same as purchasable, eligible ad inventory.

Use platform planning data for its stated purpose

Platform tools can contribute estimates and observed delivery. Google's Reach Planner guidance says its forecasts use the selected audience, budget, geography and formats alongside modelled market and historical information. Store every setting and forecast date so the plan can later be compared with delivery.

Search planning uses a different signal. Google's Keyword Planner documentation describes keyword ideas, search estimates and forecasts, while warning that campaign performance depends on factors such as budget, location targeting and ad quality. It is a Google product source, not an independent measure of total consumer demand.

Agency analysts should label first-party advertiser data, media currency, public research and platform output separately. Agreement between them can strengthen a planning hypothesis. Disagreement is useful too, because it reveals a definition or coverage problem that needs investigation.

Survey only when a survey can answer the question

A survey can estimate the distribution of answers in a defined population if recruitment and weighting support that inference. It can also collect structured feedback from a known client list. It is less suited to uncovering an unfamiliar procurement journey, where interviews allow follow-up.

Choose software after deciding sample, questionnaire and analysis. Google says Google Forms can create surveys and analyse replies, including through Sheets. Microsoft describes question types and creation routes in its Microsoft Forms guide. Those first-party records establish functions, not research quality.

Pilot questions with people outside the project. Check whether terms such as reach, attribution, programmatic or media value mean the same thing to respondents. Avoid double questions and answer options that assume an agency relationship is already beneficial.

Protect participant and audience data

Research may collect names, roles, recordings, opinions or identifiers. Media planning may involve customer lists, tags and audience segments. These uses need separate purpose, access and retention decisions.

GDS guidance on managing participant privacy covers minimisation, consent records and secure handling for UK government research. The ICO's research processing guidance addresses UK data-protection principles and grounds. Personal-data processing is governed by the UK GDPR and the Data Protection Act 2018. A person's agreement to an interview is not by itself the entire data-protection analysis.

Online advertising adds further rules. The Privacy and Electronic Communications Regulations 2003 (PECR) govern the storage of, or access to, information on users' devices, including rules relevant to cookies and similar technologies. The ICO's storage and access technologies guidance covers cookies, pixels, fingerprinting, consent and advertising measurement. Media audience design should reflect the signals lawfully available rather than forecasting an ideal dataset and asking compliance to repair it later.

Convert findings into a buying decision

End each research round with a short decision record. State the question, sample or dataset, fieldwork period, geography, finding, uncertainty and action. Name the person authorised to make that action. A finding without a decision owner becomes background decoration.

For agency positioning, findings may narrow the target client, remove an unsupported service claim or expose a capability gap. For a client plan, they may alter media allocation, creative requirements, local weighting or the measurement design. Mark any change that relies on a platform estimate or qualitative sample.

Revisit the evidence when the brief, target population, channel definitions or privacy conditions change. Before acting on a recommendation, have a researcher check recruitment and interpretation, a data specialist check audience definitions and a suitably qualified UK privacy reviewer check the data treatment.

Keep the buyer-research file and campaign-audience file separate even when they inform the same plan. Their consent, provenance, refresh dates and permissible uses may differ. The handover should tell a future planner what may be reused, what must be recollected and which unanswered question could materially change the media decision.

Before you act

Common questions

Why should buyer research and campaign audience research be kept separate?

The buyer audience includes people who initiate, approve, use or block an agency appointment. The campaign audience consists of people or organisations an advertiser seeks to influence. One dataset cannot stand in for the other because they are different populations with different recruitment, data and measurement methods.

How should interviews with buyers be structured?

Ask participants to reconstruct a recent review, brief, approval, campaign problem or supplier exit. Follow the order of events, the documents used and the people who intervened. An interview guide can cover the trigger for agency help, evidence needed to approve a supplier, operating friction, money and access, and switching costs.

What should be considered before combining audience data from different sources?

Sources such as Barb, RAJAR, PAMCo and Route should not be forced into a single spreadsheet without definitions. Covered population, geography, time basis, reach, frequency and access differ. England is contained within some Great Britain or UK datasets but is not automatically a published cut.