Defining a payroll software audience from payroll events instead of a broad SME label

The useful target is not "UK small businesses". It is a group with a shared payroll workflow, recurring problem, authority to change it and enough similarity for one product to serve responsibly. Employer headcount helps, but pay frequency, worker mix, pension process and operating model often matter more.

Start with the payroll event

HMRC's PAYE guidance for employers says an employer running payroll itself normally reports employee pay and deductions on or before payday. Map who prepares the inputs, checks the calculation, approves payment, submits information and handles corrections around that event.

The buyer and user may differ. A founder may sign a contract, an administrator may run payroll, an accountant may review it and employees receive the outcome. Research each role rather than interviewing only the budget holder.

Segment by operating model

Create separate groups for:

• employers running payroll internally;
• accountancy practices and payroll bureaux serving clients;
• employers using a managed service;
• larger teams operating payroll within a wider HR or finance platform;
• software platforms considering embedded payroll.

These groups have different permissions, volumes and support expectations. A bureau needs client separation and batch control. An employer may value guided setup. An embedded partner needs interface reliability and a clear division of responsibility.

Add workflow qualifiers

Describe each segment using observable conditions: number of active employees, PAYE schemes, weekly or monthly runs, variable pay, starters and leavers, pension providers, approval steps and source systems. HMRC's software selection page notes functional differences such as varied pay periods, payslips and pension deductions.

Workplace pensions add another qualifier. The Pensions Regulator advises employers to check their payroll process and whether software supports automatic enrolment. Record the particular scheme, data format and tasks involved.

Estimate reach without pretending it is adoption

Use the ONS business activity, size and location dataset to understand England's employer structure by region, industry and employment band. Do not label registered enterprises as software buyers. Some have no PAYE scheme, outsource the work or already use an adequate product.

Build a research frame from verified customers and likely users. Include recent switchers, satisfied incumbents, organisations using a bureau, and people who abandoned a purchase. A sample drawn only from your mailing list will overstate familiarity and interest.

Write a testable audience statement

Use this form: "We are researching [role] in [operating model] who handle [payroll pattern], currently use [approach], and encounter [observed problem] during [event]."

Attach evidence: participant count, recruitment source, dates, English locations, completed payroll cycles and paid-pilot outcomes. Define exclusion criteria as carefully as inclusion criteria. The audience is ready for product work when repeated behaviour, not demographic resemblance, supports the same narrow problem and buying route.

Test whether the group is commercially reachable

An internally coherent audience can still be costly to acquire. Record where buyers seek help, which adviser influences them, when contracts renew and what event creates urgency. Compare qualified contacts, completed demonstrations, pilot starts and paid conversions using the same segment definition.

Do not treat every participant who likes the concept as a prospect. Ask what budget would fund the change, who can veto it and what migration evidence is required. Track the time between first concern and a signed decision.

Revisit the segment after several live cycles. If support needs, willingness to pay or required functionality split into distinct patterns, create narrower audiences instead of averaging away the differences. A smaller, reachable group with a repeated job is more useful than an impressive national label.