How to Investigate Unexplained Pay Gaps at Work

Learn how to investigate unexplained pay gaps with clear data, valid comparators and documented action that supports fair, defensible decisions at scale.

A Finance Director asks why two experienced analysts, working in the same function, are paid £8,000 apart. The immediate instinct is to look for a simple explanation: tenure, performance, a difficult hire or a market adjustment. But how to investigate unexplained pay gaps is not a matter of finding the first plausible reason. It is a disciplined process of testing whether the difference is real, relevant, consistently applied and evidenced.

For UK employers, that distinction matters. A pay outcome can be commercially understandable yet still expose weak controls, inconsistent decision-making or potential equal-pay risk. If the organisation cannot show why a difference exists, who approved it and whether similar cases were treated consistently, the gap is not simply an HR query. It is a governance issue.

Start by defining the question precisely

“Unexplained pay gap” can describe several different problems. It may mean a difference between two named employees, a pattern within a job family, a gender pay gap in a business area, or an apparent inconsistency against a salary band. These require different analyses and should not be blended together.

A gender pay gap measures average or median pay differences across groups of men and women. It does not establish whether people doing equal work are paid differently. Equal-pay analysis, by contrast, examines whether employees carrying out like work, work rated as equivalent or work of equal value receive materially different contractual terms, and whether any difference can be objectively justified.

Set out the investigation question in one sentence. For example: “Why does base salary vary by more than 12 per cent among employees in Grade F project management roles with comparable scope?” This creates a usable boundary. It prevents a review from becoming an open-ended search through payroll data and makes clear what evidence is needed.

Establish a reliable pay-data baseline

An investigation is only as credible as the data beneath it. In many organisations, salary sits in payroll, job history in the HR system, performance outcomes in separate files and offer rationales in email. That fragmentation creates false gaps and missing explanations.

Build a point-in-time record for every employee in scope. At a minimum, include base salary, contractual hours, full-time equivalent pay, allowances, bonus opportunity, job title, job family, grade, manager, location, employing entity, start date and relevant pay-change dates. Add the information that may legitimately explain a difference, such as sustained performance outcomes, specialist capability, market premium, accredited qualifications or a documented change in role scope.

Normalise the data before drawing conclusions. Compare full-time equivalent pay rather than raw annual pay where part-time arrangements are present. Separate pensionable salary from overtime, commission and one-off payments. Check that employees transferred through acquisition, moved between entities or changed hours have not been assigned an inaccurate grade or salary basis.

This stage often produces the first finding: the issue is not an unexplained gap but a data-quality problem. That still requires action. An incorrect job code or missing pay-review record can undermine reporting, budget control and confidence in the wider compensation process.

How to investigate unexplained pay gaps with valid comparators

The most common error is comparing people who appear similar but are not performing comparable work. A job title is rarely enough. “Senior Manager” may cover different accountabilities, team sizes, technical demands and commercial exposure across an organisation.

Choose comparators based on the work, not the label. Start with employees in the same grade and job family, then test whether their roles have similar purpose, level of responsibility, decision-making authority, skills and working conditions. Where a formal job-evaluation framework exists, use its outcome. Where it does not, create a structured comparison using the organisation’s own role architecture and document the criteria.

Be equally careful not to define the comparison group so narrowly that meaningful patterns disappear. If women are consistently appointed at the lower end of a grade while men enter nearer the midpoint, analysing only identical teams may conceal a recruitment or progression issue. Review the relevant population at more than one level: named comparators, grade, function, location and entity.

A practical analysis should show the distribution, not just the average. Look at range penetration - where an employee sits between the minimum and maximum of their salary band - alongside salary itself. This can reveal compression, where long-serving employees sit close to new hires, or band drift, where pay has moved beyond the framework without a recorded exception.

Test explanations, rather than accepting them

Once a difference is confirmed, gather the decision trail. Ask what event created the gap: a starting-salary offer, annual review, promotion, retention adjustment, market correction, transfer, change in hours or discretionary payment. Then ask whether the stated reason is supported by contemporaneous evidence.

A valid explanation should be specific, relevant and applied consistently. “Market conditions” may be legitimate where there was a documented shortage for a particular skill at a particular time. It is not a sufficient explanation if no market evidence was retained, the premium was never reviewed, or comparable hires were treated differently without reason.

The same applies to performance. A higher salary may be defensible where an employee has demonstrably broader accountabilities or sustained, evidenced contribution. A manager’s informal recollection of someone being “exceptional” is weak evidence, particularly where performance processes are inconsistently applied.

Test each explanation against three questions:

• Is it connected to the role or the employee’s evidenced contribution?
• Was it known and approved when the pay decision was made?
• Has the organisation applied the same principle in comparable cases?

If the answer is no, record the gap as unsubstantiated rather than forcing a retrospective narrative. Retrospective justification is a warning sign. It may increase legal exposure and makes future decisions harder to defend.

Separate immediate remediation from root-cause action

Not every unexplained gap calls for the same remedy. Where an employee is below an appropriate pay level with no valid explanation, a correction may be required promptly. Consider the effective date, budget implications, contractual position and whether related employees need review. Obtain legal advice where equal-pay risk is identified or the facts are contested.

But correcting one salary does not necessarily repair the control failure that created it. Trace the gap back to its source. Common causes include managers making offers outside bands, unclear promotion criteria, inherited salaries after acquisition, unreviewed market supplements, inconsistent performance calibration and missing approval records.

This is where compensation governance becomes operational rather than theoretical. Introduce clear thresholds for exceptions, defined approvers and a required rationale. Require managers to select from controlled reasons rather than entering vague free text. Set review dates for premiums and red-circle arrangements. During pay review, flag employees whose proposed increase would create compression or take them outside band expectations.

There is a trade-off. Tight controls can slow urgent hiring or retention decisions if every exception requires multiple approvals. The answer is not to remove control, but to make the path proportionate. A time-limited, pre-approved market-premium process is faster and more defensible than an undocumented manager concession.

Create an audit trail leaders can use

A completed investigation should produce more than a spreadsheet with highlighted cells. It needs a record that another person can understand months later, including the scope, data date, comparator rationale, findings, evidence reviewed, decision made, approvals and follow-up date.

For board and executive reporting, focus on the questions that require oversight: how many gaps were identified, where they cluster, how many lack documented rationale, the cost of remediation and whether the underlying control weakness has been addressed. Avoid presenting a single headline figure without context. A small number of high-risk unexplained differences may require more attention than a larger group of clearly documented market premiums.

A live compensation record makes this process materially stronger. Orgress brings pay history, salary bands, decisions, approvals and reporting evidence into one auditable view, so teams can investigate the journey behind a number rather than reconstructing it from disconnected files.

Make investigation a regular control, not a crisis response

The strongest organisations do not wait for a grievance, a gender pay gap reporting deadline or a board challenge before examining pay differences. They review exception decisions at the point they are made, monitor band position through the year and run targeted analysis before annual pay review recommendations are finalised.

Frequency should reflect risk. A rapidly growing business hiring into scarce roles may need monthly scrutiny of offers and salary exceptions. A stable workforce may operate quarterly checks alongside formal annual equal-pay and gender pay gap analysis. What matters is that reviews are repeatable, owned and linked to action.

When a leader asks why two people are paid differently, the organisation should not need to search inboxes or rely on memory. It should be able to show the relevant facts, the reasoning at the time and the decision it is prepared to stand behind.