Many employers question their job evaluation scheme but hesitate to replace it as pay transparency requirements increase. Much of the risk of change comes from committing to a new method before its design is complete. Reversing that sequence gives decision makers a clearer choice and a stronger evidence trail.
The problem: both options carry risk
Employers with an ageing or poorly documented job evaluation scheme face a difficult choice. Keeping it means relying on results they may struggle to explain. Changing it means disruption, uncertain costs and potentially contentious changes to grading and pay.
The case against keeping an unexplained scheme has grown stronger. The EU Pay Transparency Directive requires pay structures that enable work of equal value to be compared using objective, gender-neutral criteria. These include skills, effort, responsibility and working conditions. The criteria must be agreed with workers' representatives where they exist. Workers must also be able to access the criteria used to determine pay and pay progression. Failure to meet specified transparency obligations can shift the burden of proof onto the employer in pay discrimination proceedings.
EU member states had until 7 June 2026 to transpose the Directive. By October 2026, published trackers identified five of the 27 as having fully transposed it. UK employers with staff in EU member states need to understand the national rules that apply to those employees as well as the Directive's headline requirements.
Great Britain already has an important legal test. Under section 131 of the Equality Act 2010, a tribunal must find that two jobs are not of equal value where a job evaluation study has given them different values. The exception is where the tribunal has reasonable grounds to suspect that the evaluation was based on a system that discriminates because of sex or is otherwise unreliable. Poor documentation does not in itself make a scheme unlawful. The difficulty is that it leaves the employer less able to show why the scheme produced the results it did.
Changing method presents different risks. Roles need re-evaluating and some may move grade. Employers must assess the financial consequences, pay protection, consultation, appeals and the effect on employee relations. Managers and trade unions need confidence in the replacement. Faced with that uncertainty, some organisations keep the scheme they already doubt.
Where the risk of change comes from
Much of the avoidable risk lies in the order in which decisions are made. An organisation selects a scheme or software supplier, agrees an implementation programme and begins configuring the system, often from standard settings. It then starts evaluating roles. Problems emerge during rollout. Factor descriptions do not fit parts of the workforce. Approval routes do not reflect how decisions are actually made. Results move groups of roles in ways nobody anticipated.
By then, corrections are expensive. Roles have been scored, expectations raised and contractual commitments made. Changing the method or workflow may require re-evaluation, further consultation and system changes. The organisation starts adapting its decisions to fit the system.
A second problem follows. Choices about factors, weighting, approvals and exceptions are spread across emails, workshops and configuration settings. The reasoning is often not captured in one place. The replacement scheme can become as difficult to explain as the one it replaced.
Both problems arise when the organisation commits to implementation before it has finished deciding what should be implemented.
Reversing the order
The alternative is a defined design stage before the commitment to implement. Three elements are designed together:
The process: who submits roles, who evaluates and approves them, how panels reach decisions, how challenges and appeals work and when re-evaluation is required.
The method: the factors, questions, level descriptions, scoring, weighting and grade structure, including how work across different functions will be compared.
The system: what each user sees and does, the evidence retained, how decisions are routed and what each group needs to report.
The design is developed with the people who will use it: line managers, evaluators, panel members, HR and reward teams and trade union representatives. A working prototype of the screens and workflow shows each group how the process will work for them before anything is built. Each group reviews and signs off the parts it will use. Problems that would otherwise appear during rollout surface while they can still be fixed by changing the design.
A prototype shows how the process and system will work. Establishing that the method is sound needs further work. During design, the group examines whether the factors and questions fit the work done across different occupational groups. It challenges assumptions and records why its factors and weighting fairly reflect the work. Testing the method on real roles follows once the system is built. A pilot on a representative sample of roles exposes unclear questions, inconsistent scoring and unintended effects before the full re-evaluation programme begins. Independent evaluators can score the same roles and compare their reasoning. Differences show where descriptions are ambiguous or guidance is needed. Potential grade movements and their pay protection costs can be modelled from the pilot results.
Decisions are recorded as they are made. A design record identifies the issue, the options considered, the chosen approach, the supporting evidence, the gender-neutrality checks and those who agreed it. It provides a contemporary explanation of how the method and its governance were developed. It does not guarantee legal compliance. Records made at the time are likely to carry more weight than justifications assembled after a challenge.
This sequence gives leadership two decision points. With the design complete, the employer can obtain a more reliable implementation proposal, potentially at a fixed price. Leadership can decide whether to proceed, amend the design or stop. Stopping still leaves a specification and a documented assessment of the existing scheme's weaknesses. After the pilot, leadership sees the likely effect on grades and pay before approving the full re-evaluation programme. Not every consequence can be known in advance. Each stage reduces the uncertainty before the next commitment is made.
A worked example
The following organisation is fictional.
A retail and distribution business employs 3,500 people. Most work in UK stores and warehouses. Around 400 work in two EU member states. Its points-factor scheme dates from 2011. The original designer has left. Approvals are handled by email. Leadership wants a replacement. It is concerned about cost, grading changes and disruption.
It commissions an eight-week design exercise. Store and warehouse managers, HR staff, the reward lead and two trade union representatives review the method and a clickable prototype of the workflow. Three issues emerge.
The first is wording. Warehouse managers find that the draft questions assume office work. The group develops separate question sets for office and operational roles. Both keep the same factors and scoring structure. Each pair of corresponding questions is compared side by side to confirm both describe the same level of demand. Found after rollout, this problem could have required roles to be evaluated again.
The second is approvals. The initial workflow allows one HR approver. Union representatives ask how decisions on roles covered by the collective agreement will be made. Managers identify the need for an appeal route. Both are resolved during design. The record for approvals states:
Decision: how evaluation results are approved.
Options: one HR approver or a three-person panel.
Choice: a panel of three, including a union representative for roles covered by the collective agreement. Agreement from two members is required.
Reason: judgement is shared and worker representation is built into the decision.
Gender-neutrality check: the same membership and voting rules apply to all roles, whatever the gender composition of the people in them.
Confirmed by: the reward lead and union representatives.
An appeals stage is also added to the prototype and reviewed before sign-off. Had these issues emerged during implementation, the organisation might have needed to reopen evaluations already approved.
The third is weighting. The old scheme gave more weight to budget and line management responsibility than to physical effort and working conditions. No record existed to justify the difference. Keeping it would have carried that pattern into the new scheme without a reason the business could explain. The group chooses equal weighting and records why. The Directive states that relevant soft skills must not be undervalued. The group checks that the factors recognise the interpersonal skills used in customer-facing roles.
In week eight, the executive team reviews the prototype, the decision record and the implementation specification and proposal. It approves the build and a pilot on a representative sample of roles. The pilot will check that equivalent demands score comparably across office and operational roles. Its results will also be reviewed for unintended gender effects. HR will model the grade movements and pay protection costs before executives decide whether to approve the full programme.
The design exercise exposed three issues while the organisation still had choices. The pilot tests the largest remaining unknown, the effect on grades and pay, before the full commitment is made.
Conclusion
Much of the risk of change comes from the order in which decisions are made. Designing the process, method and system with their users before committing to implementation brings problems forward to the point where they cost least to fix. Piloting on a sample shows the likely effect on grades and pay before the full programme is approved. Recording each decision as it is made gives the employer the evidence that pay transparency law increasingly expects: a record of how and why its job evaluation process, method and system were designed the way they are.
