A pay equity audit is a structured review of your organization’s compensation data to identify whether pay gaps exist based on gender, race, or other protected characteristics. The process involves collecting payroll and employee data, grouping workers who perform similar jobs, running statistical comparisons, and then fixing any gaps you find. Whether you’re doing this proactively or in response to growing state pay transparency requirements, the steps below walk you through the full process.
Decide the Scope and Set Up Protections
Before pulling any data, define what you’re auditing. Some organizations audit base salary only; others include bonuses, stock options, and total compensation. You’ll also need to decide which employee populations to include. A company with operations in multiple countries may start with U.S. employees only, or a smaller firm might audit the entire workforce at once.
One of the most important early decisions is whether to conduct the audit under the direction of outside legal counsel. When an attorney leads or oversees the process, the findings can potentially be protected by attorney-client privilege. That means if the audit uncovers pay disparities and a discrimination claim is later filed, the raw audit documents may not be subject to disclosure in litigation. Courts have upheld this protection when at least one primary purpose of the audit was to obtain legal advice, and when counsel was involved from the outset. If you plan to use privilege, the key is involving an attorney before the work begins, not after findings emerge. Be aware that using audit results as a defense in litigation can waive that privilege.
You should also consider data privacy early. Pay data is sensitive, and depending on where your employees are located, privacy laws may govern how you collect, store, and share it. Limit access to the audit team and establish clear protocols for handling the underlying files.
Collect and Clean Your Data
The quality of your audit depends entirely on the quality of your data. You’ll need three categories of information for every employee in scope:
- Compensation data: base salary, hourly rate, bonuses, commissions, equity grants, and any other forms of pay
- Demographic data: gender, race, ethnicity, age, and any other protected characteristics you’re analyzing
- Job-related data: job title, department, location, years of experience, education level, tenure at the company, performance ratings, and any certifications or licenses relevant to the role
Most of this lives in your HRIS (human resources information system) or payroll software, but it’s common to find inconsistencies. Job titles may vary across departments for essentially the same role. Performance ratings might use different scales for different teams. Tenure data might reflect hire date but not account for a leave of absence. Cleaning and standardizing this data is tedious but essential. Errors here will produce misleading results downstream.
Group Employees Into Comparator Groups
You can’t compare every employee to every other employee. The core of a pay equity audit is comparing people who do equal or substantially similar work. Grouping employees correctly is the step most likely to make or break your analysis.
The legal standard comes from the Equal Pay Act, which looks at whether employees use equal or substantially similar skill, effort, and responsibility in the performance of their duties under similar working conditions. In practice, this means grouping people by job function and level, not just by job title. Two “senior analysts” in different departments might belong in the same group if their responsibilities are comparable, or in different groups if one manages a team and the other doesn’t.
Start with your job architecture if you have one. If your organization uses job families and levels (for example, Software Engineer I, II, III), those tiers often serve as natural comparator groups. If your job structure is less formal, you may need to manually review job descriptions and reporting relationships to create appropriate groupings. Avoid making groups too narrow, which reduces your sample size and makes statistical analysis unreliable, or too broad, which lumps together jobs that genuinely warrant different pay.
Run the Statistical Analysis
With clean data and defined groups, you’re ready to analyze. There are several approaches, and the right one depends on your organization’s size and complexity.
Regression Analysis
This is the most rigorous method and the standard for larger organizations. A regression measures the relationship between compensation (the outcome you’re studying) and multiple factors that legitimately influence pay: years of experience, education, performance ratings, tenure, and location. After accounting for those factors, the model reveals whether a protected characteristic like gender or race has a statistically significant effect on pay. If women in a comparator group earn less than men after controlling for experience, performance, and other legitimate factors, the regression flags that gap.
Regression analysis typically requires a statistician or a compensation consultant to build and interpret the model. You generally need at least 30 to 50 employees in a group for regression results to be meaningful, so very small companies may not have enough data.
Cohort Analysis
A cohort study compares employees who started at roughly the same time and tracks how their pay has progressed. This approach is useful for spotting patterns that a snapshot analysis might miss. It can reveal whether certain groups of employees start at lower salaries, receive smaller raises, or get promoted on a slower timeline than their peers.
Descriptive Analysis
For smaller organizations, a simpler approach uses averages, medians, and pay ranges within each comparator group. You calculate the average compensation for men and women (or across racial groups) and look for meaningful differences. This won’t control for variables as precisely as regression, but it can surface obvious disparities and works well as a starting point.
Many organizations use a combination: descriptive analysis to identify where gaps appear, followed by regression or deeper review to determine whether the gaps are explained by legitimate factors.
Investigate the Gaps
Statistical analysis will almost always surface some pay differences. Not all of them represent discrimination. The next step is determining which gaps have legitimate explanations and which don’t.
Start at the aggregate level. If your regression shows that women in a particular job group earn 4% less than men after controlling for experience and performance, ask what legitimate factors might explain the remaining gap. Did a recent acquisition bring in employees at different pay scales? Are there geographic cost-of-living differences within the group?
Then move to the individual level. Review the specific employees at the edges of each gap. Sometimes the explanation is straightforward: an employee negotiated a signing bonus, or someone with a rare certification commands a premium. Other times, the data itself was wrong, and a corrected performance rating or updated tenure date changes the picture. Document every explanation you find. This record becomes critical for remediation and for defending your pay practices if they’re ever challenged.
Build a Remediation Plan
Once you’ve identified gaps that can’t be explained by legitimate job-related factors, you need to close them. This is where the audit produces real impact.
The EEOC advises employers to determine compensation criteria in advance and apply them consistently, and to avoid basing pay solely on factors that may carry historical bias, such as prior salary. When adjusting pay, raise underpaid employees to the appropriate level rather than lowering anyone’s compensation. Cutting someone’s pay to close a gap creates a new set of legal and morale problems.
A remediation plan should address several practical considerations. First, budget: calculate the total cost of bringing underpaid employees to parity. Some organizations phase adjustments over two or three pay cycles if the total cost is significant. Second, timing: decide whether to implement changes during the regular merit cycle or as off-cycle adjustments. Off-cycle adjustments send a stronger signal that the organization takes equity seriously, but they also draw more attention. Third, communication: decide what you’ll tell employees. You don’t need to share the full audit methodology, but employees who receive adjustments should understand why their pay changed.
Be careful that your fixes don’t create new inequities. Raising one group’s pay without reviewing adjacent groups can shift the gap rather than close it. After making adjustments, re-run your analysis to confirm the changes had the intended effect.
Document Everything and Retain Records
Federal law requires employers to retain payroll records for at least three years. Records that explain pay differences between male and female employees, including wage rates, performance evaluations, seniority data, and merit system information, must be kept for at least two years. Many employers retain these records longer as a practical matter.
Beyond the legal minimums, thorough documentation protects you. Record the criteria used for every pay decision: starting salary, raises, bonuses, and promotions. If a discrimination charge is filed, this paper trail helps you demonstrate that decisions were based on job-related qualifications rather than protected characteristics.
Make It Ongoing
A single audit is a snapshot. Pay gaps can re-emerge with every new hire, promotion cycle, or reorganization. Organizations that treat pay equity as a one-time project often find themselves back where they started within a year or two.
The most effective approach is building equity checks into your regular compensation processes. Review starting salaries against the existing range for each role before extending offers. Analyze merit increases and bonus distributions by demographic group after each cycle. Run a full audit annually or at least every two years. As state pay transparency laws continue to expand, with requirements ranging from salary range disclosures to written pay notices for new hires, organizations with an established audit process will be better positioned to comply without scrambling.