How to Run a Pay Equity Analysis (Without Getting It Wrong)

Pay equity analysis has moved from a nice-to-have to a boardroom obligation. EU legislation, UK reporting mandates, US state transparency laws, and the litigation risk that accumulates when pay disparities go unexamined have collectively made this a standard HR competency — or at least, it should be.

The problem is that most organizations that run pay equity analyses run them badly. They either produce a raw gap number that confuses dissimilar employees with a discrimination problem, or they control for so many variables that every disparity is explained away and the analysis finds nothing by design. Neither outcome is honest. Neither outcome is defensible.

This guide is about the practical version: what pay equity analysis actually measures, where the methodology breaks down, the legal requirements you need to understand, and the software options worth considering.

What Pay Equity Analysis Actually Measures

Pay equity analysis measures whether employees are paid differently based on protected characteristics — gender, race, ethnicity, age, disability status — when all other relevant factors are held constant. That last clause is where most of the complexity lives.

There are two distinct types of gaps that any rigorous analysis must distinguish:

The raw (unadjusted) gap compares average pay across demographic groups without accounting for role, level, tenure, or geography. If the average woman at your company earns 82 cents for every dollar the average man earns, that is the raw gap. It is a real and meaningful number. It reflects the cumulative effect of occupational segregation, representation gaps at senior levels, and in some cases direct pay discrimination. But it does not tell you whether employees in the same role at the same level are paid differently.

The controlled (adjusted) gap compares pay between employees in like-for-like situations — same job family, same level, same location, similar tenure and performance history — and asks whether there is still a statistically significant difference between demographic groups. This is the number that most directly reflects whether your pay-setting decisions are equitable. A controlled gap of 2% for women versus men in the same role is both legally significant and operationally actionable in a way the raw gap is not.

Both numbers matter. The raw gap tells you something important about your organization’s structure — whether women and people of color are represented at senior, higher-paid levels, whether occupational sorting is a problem. The controlled gap tells you whether your pay decisions are equitable when you remove structural factors. A company can have a large raw gap but a negligible controlled gap (structural representation problem but no direct pay discrimination), or a small raw gap masking a significant controlled gap (senior representation looks okay but pay decisions at each level are biased). You need to know which situation you are in.

Why It’s Harder Than It Looks

Pay equity analysis looks tractable until you encounter three problems that most guides either skip or understate.

The job architecture problem. Comparing pay equity across employees requires that your job architecture — the system of job families, levels, and titles — is internally consistent. If your job levels aren’t applied consistently across business units, managers, or geographies, then your comparison groups are meaningless. A “Senior Manager” in one division may be doing the work of a “Director” in another. An analysis that compares all Senior Managers will produce noise, not insight.

Before running any pay equity analysis, you need to audit whether your job architecture is clean. For many companies, especially those that have grown through acquisition or have decentralized HR practices, this audit is the most significant piece of work in the entire project. Getting it wrong means you are comparing employees who are not actually comparable, which either overstates or understates disparities depending on the pattern.

The variable comp problem. Most pay equity analyses look at base salary. Base salary is only part of total compensation — and often not the largest part for senior employees, salespeople, and anyone receiving meaningful equity grants or performance bonuses.

If your analysis shows that base salary is equitable at every level but stock grants are significantly larger for men in the same role, you have a pay equity problem that your analysis didn’t find. Total Rewards directors who look only at base are measuring a fraction of the compensation package and calling it done. A complete analysis requires total cash (base plus annual bonus) and ideally total compensation including equity awards. Equity grant data is often held in a different system than base salary data, which makes this harder — but not doing it is not a reasonable alternative.

The intersectionality problem. An analysis that looks only at gender misses race and ethnicity, age, disability, and the intersections between them. A Black woman may face a different pay gap than the average for women, or for Black employees, with both disparities present simultaneously. Analyzing each dimension independently and concluding “our gender pay gap is within tolerance” tells you nothing about what happens at the intersection.

This matters analytically and legally. In US employment law, intersectional discrimination claims (based on the intersection of race and sex, for example) are recognized and can arise from compensation data that looks acceptable when each dimension is examined in isolation. Running the analysis by dimension only is not sufficient.

The Methodology: Step by Step

A rigorous pay equity analysis follows a consistent sequence. The steps are not difficult to understand, but each one has failure modes worth naming explicitly.

Step 1: Build your comparison groups. Group employees by job family, job level, and location. These are your “similarly situated employee groups” — the populations within which you will test for pay differences. The quality of these groups depends entirely on the quality of your job architecture. If you have not already audited your levels for consistency, do that first.

Step 2: Define your compensation elements. Decide which pay elements you are analyzing. At minimum: base salary. Ideally: total cash (base plus target annual bonus) and total direct compensation (adding equity grant fair value). Document the decision. If you are only analyzing base salary, that limitation should be explicit in any report you produce.

Step 3: Run descriptive statistics. For each comparison group, calculate the mean, median, and distribution of pay by demographic group. This tells you whether the gaps are driven by a few outliers or are distributed across the population, and whether there is a pattern to investigate. A visual distribution — showing the pay histogram for men and women in the same role — often communicates more than a single gap number.

Step 4: Run regression analysis. Descriptive statistics tell you whether a gap exists; regression analysis tells you how much of the gap is explained by legitimate factors (tenure, performance, location) and how much is unexplained. The unexplained portion — typically the coefficient on gender or race after controlling for other variables — is the number you are trying to identify and address. Most pay equity software runs this analysis automatically, but understanding what it is doing matters for interpreting the output and defending your methodology.

Step 5: Set remediation thresholds. Define what gap is significant. Most companies use either ±2% (a tighter standard common in US federal contractor contexts and in EU compliance frameworks) or ±5% (a more permissive threshold often used in initial analyses). The threshold determines who enters the remediation population. Be explicit about your threshold and document the rationale.

Step 6: Develop a remediation approach. Employees whose pay falls more than X% below the group median for their comparison group, and where regression analysis cannot explain the gap with legitimate factors, should be candidates for pay adjustments. Remediation is not automatic — each individual case should be reviewed to confirm that the data is accurate and that there is not a legitimate explanation the regression missed. But the review process should be systematic, time-bound, and documented.

Legal Context: What You’re Required to Do

The legal landscape around pay equity has changed substantially in the past three years and continues to evolve.

EU Pay Transparency Directive. Effective June 2026 for companies with 250 or more employees, and extending to companies with 100-249 employees in phases through 2031, the Directive requires EU member-state transposition of rules on pay transparency and pay reporting. Key requirements include: employees have the right to know the salary range for their role before being hired; employers cannot ask about prior salary history; employers must provide pay information to employee representatives; and companies with more than 100 employees must report pay gap data by category. If a reported gap exceeds 5%, the employer must conduct a joint pay assessment with employee representatives. This is not voluntary reporting — it is a compliance obligation with member-state enforcement authority.

UK Gender Pay Gap Reporting. UK employers with 250 or more employees must publish gender pay gap data annually on a government portal. Crucially, this is a raw gap requirement — mean and median pay gaps, not adjusted gaps. The UK requirement tells the public about your structural representation problem, not whether your pay decisions are equitable. Companies that focus exclusively on UK reporting compliance are measuring the wrong thing for internal purposes.

US Federal Contractor Requirements. US federal contractors are subject to Executive Order 11246 (now part of OFCCP regulations), which requires affirmative action programs and includes compensation analysis requirements. The OFCCP uses a compensation analysis methodology that looks at similarly situated employee groups — essentially the same framework as a good controlled gap analysis. Contractors who have not done this work in advance of an audit are frequently surprised by what the OFCCP finds.

US State Pay Transparency and Equity Laws. Colorado, California, New York, Illinois, and Washington (among others) now require salary ranges in job postings, which indirectly creates pay equity exposure — when ranges are published, employees can benchmark their pay against posted ranges for comparable roles. California’s SB 1162 added pay data reporting requirements for employers with 100 or more employees. These laws do not require pay equity analysis directly but create significant reputational and litigation risk when pay gaps exist.

Software Options

Dedicated pay equity software exists across a range from point solutions to integrated platforms. An honest assessment:

Syndio is the leading purpose-built pay equity platform for large employers. Its core product runs controlled gap analysis, tracks remediation, and produces audit-ready documentation. It is purpose-built for this problem and is the most defensible choice for companies with significant legal exposure or EU reporting requirements. Price point reflects this — it is an enterprise purchase.

Trusaic focuses heavily on US regulatory compliance, particularly federal contractor OFCCP requirements and California pay data reporting. Its Workplace Equity platform includes analysis, remediation planning, and regulatory reporting. Strong choice for US-centric compliance programs.

Pave is primarily a compensation benchmarking platform, but its equity module includes pay equity analysis against market data and internal equity simultaneously. Useful if market benchmarking and internal equity analysis are happening in the same workflow, less purpose-built for regulatory compliance.

Mercer (Marsh McLennan) offers pay equity consulting and analytics through its Mercer Compensation Survey platform and standalone engagements. For companies that want a consulting firm to run and defend the methodology — useful if you are anticipating litigation or regulatory scrutiny and want external credentialing.

Korn Ferry Pay Equity Manager integrates with Korn Ferry’s job architecture and leveling methodology. If your organization already uses Korn Ferry’s job evaluation system, this is a logical choice because the comparison groups are built into the architecture.

Workday Pay Equity Dashboard is available to Workday HCM customers and runs basic pay equity analysis within the existing HRIS. It is not as methodologically rigorous as purpose-built tools but is sufficient for initial internal analysis at companies that are not yet at the regulatory complexity threshold.

When do you need a dedicated tool versus a spreadsheet? For companies under 500 employees without federal contractor status or EU reporting obligations, a well-structured spreadsheet analysis — pulling pay and demographic data from your HRIS, grouping by job family and level, running basic regression in Excel or Python — is both defensible and practical. For companies with EU reporting requirements, federal contractor status, active OFCCP audits, or more than 1,000 employees, purpose-built software is worth the investment both for accuracy and for the audit trail it produces.

Common Mistakes

Controlling for too many variables. A regression model that controls for every possible legitimate factor — performance rating, time since last review, hiring manager, office location, years in role, years in level — will explain away most disparities and find nothing. This is called “over-controlling,” and it is not analytically neutral. It reflects a choice to assume that every factor used in pay decisions is itself free of bias. If your performance ratings are themselves biased (which is a documented problem), controlling for performance ratings in your pay equity analysis does not remove discrimination — it buries it.

Not including variable compensation. Already discussed above, but worth repeating as a standalone mistake: base salary analysis is insufficient for any company where bonus, equity, or commission is a material part of total compensation.

Treating it as a one-time audit. Pay equity analysis is not a project — it is a process. Your workforce changes constantly: new hires, promotions, departures, and salary adjustments. A pay equity analysis that is current in January is out of date by July if your workforce is growing or changing. At minimum, run the analysis annually. Best practice is continuous monitoring with a remediation process triggered automatically when new gaps cross threshold.

Anchoring remediation to the wrong comparator. Some companies identify a pay gap and address it by reducing the pay of the higher-earning comparison group. This is both legally inadvisable and operationally damaging. Remediation means raising the pay of the underpaid group, not equalizing downward.

Not protecting the analysis under legal privilege. A pay equity analysis that finds significant gaps is a document that, absent proper structure, can be discoverable in litigation. Working with outside counsel to structure the analysis under attorney-client privilege does not change the methodology, but it changes the risk profile of the output.

Frequently Asked Questions

How often should we run a pay equity analysis? Annually at minimum, tied to your merit cycle or total compensation review. If you make off-cycle adjustments, promotions, or new hire offers at a meaningful volume, you should also review pay equity impact at those decision points. Some companies with sophisticated HRIS analytics run continuous monitoring against threshold triggers. Once a year is a floor, not a best practice.

What do we do when the analysis finds gaps? First, validate the data — confirm that the comparison groups are correct and that the data is accurate. Second, review individual cases to determine whether there is a legitimate explanation the regression didn’t capture. Third, develop a remediation budget and timeline. Fourth, document the process. Gaps that are found, validated, and remediated with a documented process are a significantly better legal position than gaps that are not found at all.

Does legal privilege protect our analysis? It can, if structured correctly. Having outside employment counsel retain the analysts or the software vendor and directing the analysis in the context of legal advice seeking can protect the work product. This requires intentional structure from the start — retrofitting privilege protection to an analysis that was already conducted doesn’t work. Talk to employment counsel before starting the analysis if your exposure is significant.

Do we have to disclose what we find? In the EU, yes — the Pay Transparency Directive requires reporting pay gap data and mandates remediation processes when gaps exceed thresholds. In the UK, yes — public reporting is mandatory. In the US, public disclosure is generally not required, though California pay data reporting requires submission to the state DFEH. Voluntary disclosure of methodology (not necessarily results) is increasingly expected in ESG reporting frameworks. Legal counsel should be involved in any decision about what to disclose publicly.

What is a “material” pay gap? There is no universal answer. Many companies use ±2% or ±5% as remediation thresholds. EU guidance under the Pay Transparency Directive focuses on a 5% threshold triggering mandatory joint assessment. US OFCCP uses regression-based statistical significance tests rather than fixed percentage thresholds. The threshold you choose should be documented and applied consistently — don’t set the threshold after seeing the results.

Pay equity analysis is not optional anymore — not legally, not reputationally, and arguably not ethically. But doing it badly is arguably worse than not doing it at all: it creates a false sense of compliance while leaving real disparities unaddressed and potentially creating legal liability through discoverable documents that show the company identified and dismissed problems it found.

The right approach is methodologically rigorous, covers total compensation, repeats annually, and is connected to a real remediation process with budget and accountability. That is not a high bar to clear — but it requires treating pay equity as a process, not an audit.

Related reading:

Best Compensation Management Software — Tools that automate the pay equity analysis process
HR Metrics Every CHRO Should Track — Compensation equity as a key board-level metric
People Analytics Guide — The data infrastructure that makes pay equity analysis possible