Audience research
Part of Marketing attribution audience research: six populations
Marketing attribution target audience for platform research
Define a marketing attribution target audience for platform research using recent platform tasks, decision roles and documented evidence of authority.
A marketing attribution target audience for attribution research is a defined set of organisations and roles linked to recent use, evaluation or implementation of an attribution platform. It is not a loose category of marketers in England and should not be assembled from assumed demographics. Before recruitment, name the platform or platform type and write an eligibility rule that another researcher can apply to the same frame.
Define the platform audience
Examples of named tools include Google Analytics 4 (GA4), Adobe Analytics and Triple Whale. Record which tool or provider the organisation uses, evaluates or is implementing; “uses analytics” alone is not a sufficiently specific eligibility rule.
The relevant audience differs by platform type. For in-house attribution, recruit people who configure or use attribution reporting; for marketing mix modelling (MMM), include the owner who commissions or interprets model outputs; for incrementality tools or providers, include the person responsible for designing or evaluating tests. Recruit across only the types relevant to the research question, because their tasks and decision roles differ.
What to take away
- Name the attribution platform or type, then define eligibility so another researcher can apply the same rule to the same frame.
- Assign respondent roles to specific questions rather than assuming seniority covers every authority.
- Affected people such as customers or staff are a separate population from research participants.
- A segment is usable only when eligibility is reproducible and every claim traces to evidence.
- A segment cannot support an England population estimate just because several participants agree.
Start with the organisation
Choose an England geography rule. A practical rule may use the contracting entity's operating address for the attribution work, verified against Companies House for the registered office or the ONS local units dataset for an operating site. Check the company number on Companies House and the local unit postcode on ONS before screening.
Head office, respondent home address and customer location answer different questions. Record whichever rule is chosen and do not switch later.
England geography rule
Which address defines England?
Use contracting entity operating address
Do not use head office, home or customer location
Specify organisation activity and event window. Eligibility might require a documented attempt to reconcile channel credit, assess a marketing mix model or plan an incremental test within a stated period. Interest in analytics is not an event. For platform research, a recent task might be configuring or reviewing attribution reporting in GA4 or Adobe Analytics, reviewing attribution output in Triple Whale, commissioning or interpreting an MMM, or designing or evaluating an incrementality test.
For a worked example, an England round on marketing mix adoption could set this rule. The organisation has a Companies House registered office in England. Its records show a named marketing analytics role active in the 12 months to 31 March 2024. It attempted to reconcile channel credit or plan an incremental test in that window.
The ONS business activity dataset can describe registered enterprises and local units by region, but it does not identify attribution users or buyers.
Assign respondent roles to questions
Separate the system user, business owner, procurement participant, budget approver and person responsible for privacy or security review. For platform studies, relevant role titles may include attribution lead, analytics manager, MMM owner or incrementality lead. Do not assume a senior title includes every authority. A respondent qualifies for a question only if the screener records a recent relevant task or evidenced decision role. A job title alone is not proof: ask about a specific recent task and, where practical, corroborate it with a redacted project brief, platform-access or ownership record, approval record or procurement document. Companies House and ONS datasets can verify organisation details or geography, not an individual's decision authority.
Respondent evidence matrix
Research question
- How are events defined?
- Created, approved or maintained spec
- Why was supplier considered?
- Wrote brief, evaluated or approved
- What decision used output?
- Owned or documented decision
- Who may be affected?
- Maps data or operational process
Eligible evidence
- How are events defined?
- Only viewed finished dashboard
- Why was supplier considered?
- Heard about project later
- What decision used output?
- Reports general preference
- Who may be affected?
- Guesses from job title
Exclusion
- How are events defined?
- Why was supplier considered?
- What decision used output?
- Who may be affected?
Use a matrix:
| Research question | Eligible respondent evidence | Exclusion |
|---|---|---|
| How are events defined? | Created, approved or maintained the specification | Only viewed a finished dashboard |
| Why was a supplier considered? | Wrote the brief, evaluated evidence or approved procurement | Heard about the project later |
| What decision used the output? | Owned or documented that decision | Reports a general preference |
| Who may be affected? | Maps the underlying data or operational process | Guesses from a job title |
Affected people are a separate population. Customers, website visitors or staff whose data appear in a model do not become research participants unless they knowingly join the study. Their possible rights and harms still belong in the research plan.
Build and disclose the frame
State whether recruitment comes from customer records, an approved panel, the Market Research Society Company Partner directory or another documented source. List coverage gaps, duplicates and records whose England status is unknown. Keep counts for selected organisations, invitations delivered, screened people, eligible people, completed sessions and usable responses.
| Coverage stage | Count to record |
|---|---|
| Selected organisations | Number in frame |
| Invitations delivered | Number sent |
| Screened people | Number screened |
| Eligible people | Number meeting the rule |
| Completed sessions | Number finished |
| Usable responses | Number retained |
Recruitment frame disclosure
- State recruitment source
- List coverage gaps and duplicates
- Flag unknown England status
- Count selected organisations
- Count invitations delivered
- Count screened and eligible people
- Count completed sessions and usable responses
The GOV.UK participant recruitment guide says to define criteria, recruit people who can answer the research question and check recruitment bias.
The ICO direct-marketing identification guidance says genuine market research is not direct marketing, but promotional content or later marketing to participants needs separate review. Apply the UK GDPR and Data Protection Act 2018 when handling participants' personal data.
Decide when the segment is usable
Use an illustrative threshold of two eligible respondents per role and six organisations per round, and let the named research owner set and record the final number.
A segment can guide another round when its eligibility can be reproduced, every claim traces to evidence, non-response is visible and contradictory cases are retained. It cannot support an England population estimate merely because several participants share a view.
Stop if eligibility relies on inferred traits, one recruitment source omits a material role, or participants cannot understand how their data will be used. The output should be a dated eligibility specification and coverage report, not a claim that the segment represents demand.
Before you act
- Name the attribution platform or type and write an eligibility rule before recruitment begins.
- Record the England geography rule and do not switch later.
- Specify the organisation activity and event window.
- Map each research question to eligible respondent evidence.
- Disclose the recruitment frame and its coverage gaps.
- Set an evidence threshold and have a named owner approve the final number.
Common questions
What makes a target audience for attribution research?
It is a defined set of organisations and roles linked to recent use, evaluation or implementation of a named attribution platform or platform type. Before recruitment, record the tool or type and write an eligibility rule another researcher can apply to the same frame.
How should respondent roles be assigned to research questions?
Separate the system user, business owner, procurement participant, budget approver and privacy or security reviewer. Relevant platform-study roles may include an attribution lead, analytics manager, MMM owner or incrementality lead. Do not assume a senior title includes every authority. A respondent qualifies for a question only if the screener records a recent relevant task or evidenced decision role, corroborated where practical by a project, access, approval or procurement record.
When can a segment be used to guide another round?
When the named research owner has approved the evidence threshold and the coverage report shows the required counts. It cannot support an England population estimate merely because several participants share a view.
