The traditional approach to salary bands assumes resources most SMEs don't have. A Mercer or Willis Towers Watson engagement runs €15–€40k per year for survey access; the consultancy work to translate the survey into your role architecture adds another €30–€80k. Six months is a realistic timeline. The compensation team that operates the structure afterwards is one or two FTE.
For a 200-person company facing a June 2026 transposition deadline, that's the wrong solution. Not because the traditional approach is wrong — it's right for the companies it was built for. Wrong because the assumptions about who is doing the work don't hold.
This article describes a minimum viable salary band structure that an HR Director can build alone with a spreadsheet, an honest view of the existing team, and about three weeks of focused work. The output is not as polished as a consultancy deliverable. It is, however, defensible — which is the standard the directive actually requires.
What the directive actually asks for
The directive does not require salary bands. It requires that pay decisions be explainable on objective, gender-neutral grounds, applied consistently across categories of equal work or work of equal value.
Salary bands are not the obligation; they are the most practical artefact that satisfies the obligation. A documented band tells the regulator (and the works council, and the employee asking under Article 7) where each role sits in a structure, why an offer was at that level, and how variation within the band is explained.
The directive accepts any methodology that meets three criteria: transparent, documented, gender-neutral by design, and applied consistently. The five-step build below produces exactly that.
The five-step build
Total: three weeks of focused work. The output is a documented salary band structure, a placement of every current employee, and a methodology that can be audited.
What this approach does not give you
A €40k consultancy engagement produces a deliverable that is more polished than the five-step build above. Three things the consultancy delivers that the SME approach does not:
- Statistical confidence intervals on each band. The Mercer methodology aggregates across thousands of companies; the SME approach aggregates across maybe a dozen external references per cell. The midpoint is less statistically precise.
- Sector benchmarks at fine granularity. "Senior Software Engineer, fintech, Lisbon, 50–250 FTE" is a slice a survey can produce; the SME approach approximates it.
- Annual refresh as a service. Survey access includes annual updates. The SME approach requires the company to repeat the calibration step each year.
For a 200-person SME, the cost of these three losses is rarely worth the €40k. The directive does not require statistical confidence intervals; it requires documented methodology. The directive does not require fine-grained sector benchmarks; it requires the methodology be applied consistently. The directive does not require annual professional refresh; it requires the methodology to be maintained.
Maintenance — what year two looks like
The structure is not a one-shot artefact. It has to be maintained annually for the directive's documentation to remain current.
The annual maintenance cycle:
- Re-calibrate cells. Repeat step 02 against the past 12 months of hiring data and current market signals. Update midpoints where the change is material (>5%).
- Re-place employees. Promotions, role changes, and tenure increments shift placements. Surface employees who have moved out of band.
- Refresh the rationale catalogue. New rationale categories may emerge (geographic expansion, new role types). Document them before applying.
- Run the per-category pay gap calculation. Under Article 9, the per-category mean and median gaps are mandatory. The category structure built in step 01 feeds directly into the report.
The maintenance cycle is what makes the bands defensible over time. A band structure that exists on paper but has not been updated in two years is harder to defend than no structure at all — it suggests the company built the artefact for compliance reasons but is not actually using it to set pay.
"Worth considering" alternative for SMEs that do want survey data: WTW's lower-tier package, Radford's startup-focused product, or industry-association salary surveys (these are cheaper and often more relevant than the flagship Mercer survey for SMEs). The SME approach above is the lower bound — the minimum that satisfies the directive. A purchased survey can refine step 02, but the rest of the build remains the same.
How this feeds the directive's report
The five-step build is not designed primarily for the report. It is designed primarily for pay decisions. But the same artefact answers the report's questions.
Step 01 (the role grid) is the category structure for Article 9 metrics 08–09. Step 03 (the band values) is the documented compensation basis. Step 04 (employee placement) is the cohort data for per-category gap calculations. Step 05 (methodology document) is what is examined if the report is challenged.
The work done to build the band structure is the same work the directive requires. The structure is the artefact. The defence is implicit.
The consultancy delivers the artefact. The structure delivers the defence. For most SMEs, the second is what the directive actually asks for.
Where the diagnostic starts
Before running the five-step build, an honest view of where the company currently sits is useful. The ReadinessCheck™ surfaces, by axis of the directive's principal obligations, what is already documented and what is missing. The result determines whether the build above takes two weeks or four.
For the company that already has a clean payroll export and an existing role taxonomy, PayGapCheck™ can run the per-category metrics directly from the uploaded data — useful once the band structure exists and the calibration is complete.
The band structure is the foundation. The analytical report is the output.
PayGapCheck™ takes a salary file and produces the analytical report — overall and per-category gaps, quartile distributions, and the metrics Article 9 requires. The band structure built via the five-step process above is the input that makes the report defensible.
Start the analysis →