Audience research
Defining a marketing automation target audience by workflow failure, not firm size
Define a marketing automation target audience in England using account evidence, workflow triggers, buying roles and a transparent recruitment screen.
A marketing automation target audience is useful only when it identifies accounts with the same operational problem and the people involved in changing it. "English SMEs" is not enough. It covers organisations with very different customer journeys, data, staffing and purchasing routes.
No buyer interviews or surveys were carried out for this guide. The process below creates a recruitment and proposition hypothesis, not a measured audience size.
What to take away
- Define the target audience by a repeated workflow failure, not by firm size or sector label.
- Screen accounts for repetition, data, ownership, change pressure and feasibility before recruiting.
- Map the buying group across operator, process, data and commercial roles rather than one requester.
- Keep the population frame, research sample and proposed commercial segment as separate outputs.
- State a decision rule before recruitment so evidence and judgement stay clearly divided.
1. Name the workflow and the failure
Begin with one repeated path, such as enquiry to first response or renewal reminder to booking. Describe the failure in observable terms. "Leads are poor" is an opinion; "enquiries remain unassigned after the response window" can be checked.
Workflow failure and entry condition
- Repeated path: enquiry to first response
- Observable failure: unassigned after response window
- Entry condition: runs the workflow
- Entry condition: failure frequent enough to care
- Entry condition: authority or budget to change
Add an entry condition. An account belongs in the first research frame only if it runs the workflow, sees the failure often enough to care and has some authority or budget to change it. This prevents a sector label from doing work it cannot support.
2. Choose the organisational unit
Decide whether the unit is an enterprise, site or team. The ONS UK business activity dataset covers VAT or PAYE registered enterprises and local units, with geography, industry and employment bands.
Enterprise versus local unit
Enterprise
- Definition
- VAT or PAYE registered
- Geography
- Placed at head office
- England operator
- May be outside England
- Use in frame
- Bound population
Local unit
- Definition
- Individual site
- Geography
- Actual site location
- England operator
- May be in England
- Use in frame
- Bound population
Local units are individual sites. An enterprise is generally placed at its head office, so a nationwide operator may have an England site but a head office elsewhere.
Record the choice, date and exclusions. If unregistered organisations matter, state that the ONS frame is incomplete for this purpose. Official counts can bound a population; they cannot identify platform intent.
3. Screen for operating conditions
Use a short account screen with verifiable answers:
Screen for operating conditions
| Test | Evidence to seek |
|---|---|
| Repetition | A recurring event and a documented current route |
| Data availability | The minimum fields exist and their source is known |
| Ownership | A named role is accountable for the journey |
| Change pressure | Delay, rework, missed control or service problem |
| Feasibility | Time, integration and governance can be investigated |
Do not use technology adoption as a shortcut. The 2026 UK Business Data Survey found variation in digitised-data handling by business size and sector. It studied UK businesses, not England's demand for marketing automation.
4. Map the buying group
Recruit beyond the person who requests a demonstration. A daily operator can explain the work; a process owner defines the desired change; a data or security owner tests whether information may move; and a commercial approver weighs cost, terms and exit. One individual may hold several roles in a small firm.
The goal is not a fixed committee chart. It is to learn where decisions and objections actually arise. Ask each participant who supplied information, who could say no and who would inherit the system after launch.
5. Build a contrast matrix
Place workflow state on one axis and decision role on the other. Add only contrasts likely to change the answer, perhaps organisation size, regulated versus ordinary data, in-house versus agency delivery, or single-site versus distributed operations. Include non-adopters as well as recent adopters so recruitment does not select only people already enthusiastic about software.
The GOV.UK Service Manual on finding participants for user research recommends recruiting actual or likely users and considering relevant variety. It also advises collecting only the minimum personal information needed; delete recruitment data when it is no longer required.
6. Set a decision rule
Before recruitment, state what evidence would move an account into or out of the proposed segment. For example: continue only if the same hand-off failure appears across more than one role, an accountable owner exists, and the data route can be investigated. The threshold is an internal product decision, not a claim that the sample represents England.
Use neutral interview questions to keep recruitment conversations non-leading. Keep three outputs separate: the population frame, the achieved research sample and the commercial segment proposed after analysis. That distinction allows another reader to see where evidence ends and judgement begins.
Three example audience hypotheses
These examples show the output of the method. They are hypotheses to test, not findings, and none comes from buyer research.
- Unassigned enquiries.Accounts: B2B service firms where web enquiries wait for a manual hand-off. Roles: sales operations owner, marketing operator. The platform question is lead routing and scoring, which suits CRM-linked tools (see the B2B nurture guide) such as HubSpot or Salesforce Account Engagement (Pardot).
- Abandoned baskets.Accounts: online retailers whose basket reminders are sent by hand or not at all. Roles: ecommerce manager, CRM lead. The platform question is store-data triggers, which suits ecommerce-led tools such as Klaviyo or Dotdigital (see the ecommerce automation guide).
- Renewal reminders.Accounts: subscription or membership organisations whose renewal messages are sent from spreadsheets. Roles: operations lead, data owner. The platform question is date-based journeys and consent records, which most mainstream tools offer; compare how each handles preference changes.
For each hypothesis, write the failure, the entry condition, the roles to recruit and the evidence that would remove it from your list.
Use the first screen in practice
Choose ten candidate accounts from two deliberately different cells in the matrix. Verify the workflow through public material or a neutral screener, then invite the relevant roles to research. The invitation should describe the study, not pitch the product. A qualified reviewer should approve the personal-data purpose, lawful basis, notice and retention route (see marketing automation data protection) before contact details are processed.
Before you act
- Name one repeated workflow and describe its failure observably.
- Add an entry condition covering workflow, failure frequency and authority.
- Choose the organisational unit and record date and exclusions.
- Screen accounts against the five operating-condition tests.
- Map roles beyond the person requesting a demonstration.
- Set the decision rule before starting recruitment.
Common questions
Why is a sector label like English SMEs not enough?
It covers organisations with very different customer journeys, data, staffing and purchasing routes. A useful audience identifies accounts with the same operational problem and the people involved in changing it, so a broad label cannot do that work.
What does the ONS business activity dataset provide?
It contains VAT or PAYE registered enterprises and local units, with geography, industry and employment bands. A local unit is an individual site, while an enterprise is generally placed at its head office, so a nationwide operator may have an England site but a head office elsewhere.
How should the buying group be mapped?
Recruit beyond the person who requests a demonstration. A daily operator explains the work, a process owner defines the change, a data or security owner tests information movement, and a commercial approver weighs cost and terms. One individual may hold several roles in a small firm.
