A defensible methodology, by design.

The calculation engine is based on Tool 5 of the European Institute for Gender Equality’s framework for equal pay, which is the methodological reference used by EU regulators and most member-state guidance documents. We use the EIGE work-of-equal-value criteria — skills, responsibility, effort, and working conditions — as the four pillars of comparability.

The four pillars are weighted 35 % skills · 25 % responsibility · 25 % effort · 15 % working conditions. These weights are configurable in the platform so an organisation can document a justified deviation, but the EIGE defaults apply unless the customer explicitly changes them through a recorded decision.

Roles are mapped to the International Standard Classification of Occupations (ISCO-08), maintained by the International Labour Organization. ISCO-08 provides a stable, regulator-recognised taxonomy that lets comparisons survive role-title drift across teams and jurisdictions.

The platform performs an initial mapping using deterministic rules and falls back to assisted classification where the title is ambiguous; every assisted mapping is presented to a human reviewer for confirmation and is logged with the reviewer identity. No mapping is finalised without an explicit human decision.

Pay is measured as full remuneration per the Directive’s Article 3 definition: base salary plus variable pay, long-term incentives, allowances, and quantifiable in-kind benefits. The platform captures each component separately so a customer can examine the gap at each layer, which is what auditors typically ask for.

We report two figures side by side:

• Unadjusted gap — the raw difference in average pay between groups. This is the headline number most often quoted in the press and is required as one of the disclosure elements under Article 9 of the Directive.
• Adjusted gap — the residual difference after controlling for the work-of-equal-value factors above (job family, level, experience, working time, location, and the customer’s justified pay-band design). This is the figure the Directive uses to trigger remedial action: where the gap exceeds 5 % in a category of workers performing equal or equivalent work, Article 10 requires the employer to act.

We disclose the regression specification, the controls included, and the sample size for each cell of the analysis. A reader who wants to reproduce the result has everything needed.

• Article 9 — reporting obligation for employers with 150 or more workers (first cycle in 2027, then on a three-year cycle for the 150–249 band, annually above 250).
• Article 10 — joint pay assessment with worker representatives is required when the adjusted gap exceeds 5 % in any category of workers performing equal or equivalent work, and where the difference cannot be justified by objective, gender-neutral factors.

We are transparent about the boundaries of the tool so customers can plan for the work that remains in-house.

• We do not provide the legal opinion on whether a specific adjusted gap is “justified by objective, gender-neutral factors.” That call belongs to the customer’s legal counsel, informed by our outputs.
• We do not replace the negotiation with worker representatives required under Article 10. We provide the data that makes the conversation productive.
• We do not file reports with member-state authorities on the customer’s behalf. We produce the report; submission remains the employer’s responsibility.
• We do not adjust historical pay data. If the customer corrects a payroll error retroactively, the corrected source becomes the input for the next scoring run; prior runs are immutable.

The four pillars decompose into 14 subfactors, each scored on a 0–5 scale — except Knowledge, which uses an extended 0–8 scale to cover the full range from no formal qualifications to globally recognised expertise. Every subfactor below is the verbatim EIGE Tool 5 definition (toolkit Tables 9–22); level descriptors are the regulator-facing anchor text. Click a card to expand the level scale and sector examples.

Methodology version eige-v1.2 · provisional · published 2026-05-18.

Every scoring run captures the inputs, the configuration (weights and overrides), the intermediate statistics, and the final outputs in an immutable record. Runs can be re-opened years later by any role with read access; the audit log shows who ran them, who reviewed them, and what changed between runs.