ESRS S1 Own Workforce: Complete Guide to Policies, Metrics and Worker Characteristics

Start ESRS S1 with a population decision: employees, S1 non-employees and S2 workers must be separated.

ESRS S1 is demanding because the public narrative and the workforce dataset must describe the same people, period and impacts. HR, payroll, health and safety, employee relations, compliance and finance often use different populations. Payroll can exclude agency labour; contractor registers can include workers who belong in S2; and sensitive incident or disability data can be restricted by law.

The safest sequence is boundary first, material impacts second, disclosure architecture third and metrics fourth. That prevents a polished table that cannot be reconciled to the workforce model or due-diligence narrative.

Start ESRS S1 with a population decision: employees, S1 non-employees and S2 workers must be separated.

The narrative layer is not an HR-policy inventory. It should explain how policies cover material impacts, which groups and geographies are in scope, how workers and representatives are engaged, which channels are available, how effectiveness is assessed and how remedy is provided or supported when the undertaking caused or contributed to harm.

Link policy coverage to the material working-condition, equality, safety or rights issue.

Identify whether engagement is direct or through workers’ representatives and how it influenced decisions.

Distinguish employee communication, speak-up, grievance, trade-union and formal legal routes.

Assess channel effectiveness through accessibility, trust, non-retaliation, timeliness, resolution and recurrence.

Distinguish systemic corrective action from remedy for affected people.

S1-5 reports total employee headcount by gender; employee headcount for each country with 50 or more employees among the ten largest countries; permanent, temporary and non-guaranteed-hours employees; permanent-employee turnover; and an explanation of inconsistency with the most representative financial-statement employee number.

S1-6 applies where non-employees in the own workforce are connected to material S1 IROs—for example because flexible labour is critical to the business model, reliance is increasing or impacts are material. Report headcount or FTE and state whether the measure is period-end or average. If exact data are unavailable, use a transparent reasonable estimate and explain coverage and limitations.

Reconcile procurement, contingent-labour platforms, site access and HR records. Prevent duplicate people across invoices, access systems and project logs, and identify remote or decentralised populations that central systems omit.

S1-13 requires a coherent health-and-safety account: management-system coverage, fatalities, recordable accidents and rate per one million hours worked, work-related ill health and days lost where applicable. Define the population, recordability basis and hours denominator; reconcile contractor hours or estimates and explain changes in boundary or capture.

State whether employees and S1 non-employees are covered.

Map local legal categories to a controlled group definition.

Use hours worked or a documented estimate; avoid unexplained paid-hours denominators.

Protect health information and maintain work-relatedness and case review.

Explain acquisitions, disposals, better capture and restatements.

Family-related leave metrics require defined entitlement and take-up populations. The unadjusted gender pay gap uses gross hourly pay of all male and female employees and is not an equal-pay-for-equal-work test. The annual total remuneration ratio compares the highest-paid individual with median employee annual total remuneration excluding that individual. Pay elements, currency conversion, population and period must be controlled.

S1-16 focuses on substantiated incidents identified in the reporting period, not all concerns received. Discrimination and other human-rights incidents are reported separately, alongside fines, penalties and compensation recognised in the financial statements. A restricted register should preserve intake, issue, population, investigation, substantiation basis, remedy, cut-off, finance linkage and privacy classification.

ESRS S1 architecture: material impacts drive narrative disclosures and the applicable workforce metrics.

Confirm material S1 sub-topics and affected worker groups.

Build and approve the employee, S1 non-employee and S2 population register.

Map S1-1 to S1-4 to policies, engagement, grievance/remedy, actions and targets.

Create the S1 metric dictionary with population, unit, period, source, owner and privacy class.

Reconcile S1-5 to HRIS, local HR and financial statements.

Build the S1-6 non-employee data or estimate control.

Calculate S1-7 to S1-15 with metric-specific denominators.

Operate the restricted S1-16 incident register and legal/privacy review.

Complete narrative/metric consistency review and governance approval.

The group classifies employees in S1, tests agency workers for S1 non-employee treatment and leaves independently managed delivery-provider workers in S2. It reports exact employee data, a controlled estimate for a material non-employee population and a separate S2 worker-impact analysis. Health-and-safety rates use hours by population and explain estimated contractor hours.