How to build a salary grid that meets EU Pay Transparency Directive requirements

Every employer subject to the EU Pay Transparency Directive needs a salary grid. Not just to publish pay ranges in job advertisements, but to demonstrate that pay reflects the value of work rather than the identity of the person doing it. 

That requires two things: defined pay ranges for every role, and a documented rationale for where those ranges sit. This guide covers how to build both. 

What a salary grid is, and what the directive expects from it 

A salary grid maps roles to pay ranges based on their assessed value. It is the operational output of a job evaluation process — not a spreadsheet of market rates, but a structured framework that reflects the relative worth of roles within the organisation. 

The directive does not prescribe a specific grid format. It requires that pay structures be based on objective, gender-neutral criteria, that pay ranges be available to workers and candidates, and that work of equal value be paid equally. A well-built salary grid satisfies all three requirements simultaneously. It is also the instrument that makes pay gap reporting meaningful. Without a grid, gap data has no structural context for analysis or remediation. 

For a full overview of the directive’s obligations and the role the grid plays in meeting them, see our pay transparency compliance checklist → 

The four-factor framework: skill, effort, responsibility, and working conditions

The directive’s “work of equal value” standard requires that roles be assessed against four factors, drawn from EU equal pay legislation and carried into Directive 2023/970: 

• Skill — qualifications, experience, and competencies required to perform the role 
• Effort — physical and mental demands placed on the worker 
• Responsibility — accountability for people, resources, decisions, and outcomes 
• Working conditions — environment, schedule, hazard exposure, and other conditions of work 

These four factors are the legally recognised basis for comparing roles of different types. Two roles that score equivalently across all four dimensions constitute work of equal value — regardless of job title, department, or the gender of the people historically doing them. 

This is where salary grids built on market rates alone fall short. Market data tells you what comparable organisations pay for a role. It does not tell you whether those rates reflect equal value. A warehouse team leader and a customer service manager may look very different on paper but score equivalently when evaluated against these four factors — and must therefore be paid within equivalent ranges. Only a grid grounded in the four-factor framework can demonstrate that. 

Building the grid: a five-step process

A salary grid built to meet the directive’s requirements is the product of five sequential steps. Each builds on the previous one, as skipping or shortcutting any step undermines the defensibility of everything that follows. 

Before starting, read our guide to building a fair compensation policy →.  

1. Inventory and group your roles 

List every role in the organisation. Group roles that share comparable scope, similar levels of responsibility, and equivalent skill requirements. These groupings do not need to be identical, but they need to be defensible as representing work of equivalent value. 

Most organisations find they have fewer meaningful groups than they expected. Job title proliferation — where small variations in title mask essentially equivalent roles — is one of the most common sources of undocumented pay differences. Consolidating to genuine role groups rather than job title clusters is the first step toward a defensible structure. 

2. Evaluate each group against the four factors 

For each role group, assess the level of skill, effort, responsibility, and working conditions required. This assessment must be documented, consistent, and independent of who currently performs the work or what the market pays for it. 

Many organisations use a points-based scoring system: each factor is scored on a defined scale, roles are ranked by total score, and bands are defined around natural clusters in the score distribution. The output is a ranked hierarchy of role groups based on assessed value, which is the foundation on which the rest of the grid is built. 

3. Benchmark against external market data 

Once roles are evaluated and ranked internally, set the pay ranges for each band using reputable salary survey data for the relevant sectors and geographies. Market data should inform the midpoint of each range. This is equivalent to the rate for a fully competent performer in that role group. 

Where internal equity and market rates conflict — where a market midpoint would place two role groups of equivalent internal value at materially different pay levels — that tension needs to be resolved and documented, not absorbed silently into the grid. 

4. Set band widths and midpoints 

Define the width of each pay band to reflect the real variation in pay appropriate for different levels of development within a role. A new hire sits toward the bottom of the band. A fully proficient performer sits at or near the midpoint. An exceptional performer or long-serving expert may sit toward the top. Typical band widths range from 50% to 80% of the midpoint, though this varies by level and sector. 

The directive requires that pay ranges in job advertisements be genuine. That means bands need to be sized to reflect real pay variation and not set as wide as possible to give maximum flexibility. Artificially wide bands do not satisfy the obligation. 

5. Document the rules for movement within and between bands 

The grid defines the structure. Governance defines how pay moves within it. Document the criteria that determine where a new hire enters their band, what triggers a mid-band increase, what justifies an above-midpoint salary, and what determines progression to the next band. These rules must be objective, gender-neutral, and accessible to workers on request. 

Without them, individual pay decisions remain arbitrary regardless of how well the grid was built. This is the step most organisations skip, and the one that leaves them most exposed when the first pay information request arrives. 

Five mistakes to avoid

Building the grid around market rates alone. Market data tells you what comparable organisations pay. It does not tell you whether those rates reflect equal value. A grid built on benchmarks without an underlying four-factor evaluation cannot demonstrate equal value. It also cannot defend pay differences between roles the market treats differently but the directive may consider equivalent. 

Using job titles as a proxy for role groups. Job titles are often inconsistent and influenced by factors unrelated to the work itself. Two people with different titles may be doing work of equal value. Two people with the same title may not be. The four-factor evaluation determines the group — the title does not. 

Setting bands too wide to be meaningful. A pay range spanning more than 60–70% of its midpoint may be too broad to satisfy the directive’s genuine pay range requirement. Width should reflect real variation in pay for different stages of development within a role, and not maximum flexibility for pay negotiations. 

Failing to place existing employees in the grid. Building a grid for future hires without mapping current employees to it creates an immediate compliance problem. Any employee sitting below the band minimum or above the band maximum has a pay position the organisation cannot defend when the first pay information request arrives. 

Treating the grid as finished. A salary grid requires regular maintenance: market data refreshed annually, role evaluations reviewed when responsibilities change, and pay positions audited annually with outliers documented and justified. A grid that was accurate two years ago may be systematically underpaying certain groups today. 

Maintaining the grid over time 

A salary grid is a living document, not a project deliverable. Three things need to happen on a regular cycle. Market benchmarking should run annually, since pay ranges that drift below market create both retention problems and pay equity risk. Role evaluations should be reviewed whenever responsibilities change materially. For all other roles, a two-to-three year review cycle is the standard. Pay position audits should run annually, checking every employee against their band and documenting any outliers with a clear, objective justification. 

The governance owner identified in your compliance framework is the person accountable for this cycle. Without named ownership, maintenance slips, and a grid that no one is responsible for maintaining becomes unreliable faster than most organisations expect. The same principle applies here as to AI governance: the structure is only as reliable as the governance around it. Read more on building governance that holds → 

The grid is the instrument through which everything in the compensation policy becomes operational. Getting it right is not a one-time exercise. It is the ongoing work of fair pay. 

The Compensation Strategy Report 2026 covers how leading European organisations are approaching pay frameworks this year → 

👉 See how Elevo helps you build and maintain a directive-compliant salary grid → A human approach to developing talent