Pay equity audit: How-to

10 min readPublished May 06, 2022Updated May 09, 2022

Caveat: I am not a lawyer and this post is not legal advice. The information provided is general in nature. Please consult your legal counsel for more information.

In one of my recent roles as an Engineering Manager, I advocated for and successfully conducted a pay equity audit. A big part of the success was a well-planned methodology and timeline.

In this post, I share our audit process and some considerations for each step. For more information about what a pay equity audit is, please refer to my previous post.

0. Find an accomplice

Ok, this first step isn't really part of the audit methodology but it's crucial to the process. If you are thinking about advocating for a pay equity audit in your company, find an accomplice.

As I pointed out in my post about the challenges we faced in conducting a pay equity audit, leadership buy-in is critical. And depending on your organization, that might be your biggest obstacle.

By finding someone who will advocate with you, you can share the risk, responsibility, and labor required to get this initiative off the ground. You can take turns amplifying each other's voices. And you can present a united front in the face of pushback.

My accomplice was Kate Donaldson. And I'm so grateful for her collaboration on this meaningful initiative.

1. Create a series of sampling dates

Once you have approval for a pay equity audit, start by identifying the time period you want to analyze.

If this is the first time you're conducting an audit like this, I suggest running the audit for 2 distinct time periods–one for past data and one for current data. You can use your historical audit to test your methodology. You'll be able to identify any missing data that you would need to collect before conducting the audit for your current employees.

Running a historical audit as well as a current audit can also help identify any patterns. Going forward, you won't need to run 2 audits because you will already have historical data from previous audits.

We chose to look at 2 years' worth of data for each audit. For our historical audit, we chose 3 sampling dates covering 4 years ago to 2 years ago (Date 1: 4 years ago, Date 2: 3 years ago, Date 3: 2 years ago). And we chose a similar group of sampling dates for our current audit–2 years ago to the present-day (Date 1: 2 years ago, Date 2: 1 year ago, Date 3: present-day).

2. Create a list of demographic groups

Next, create a list of demographic groups that you are interested in analyzing for pay equity discrepancies. This list will inform what data you'll need to collect based on what you do and don't have available (in your HR systems or that your employees are willing to share).

When making your list, consider not only the categories that you want to analyze but the specific identities within each demographic group. There are pros and cons to how granular you choose to go.

The more specific you are, the more in-depth analysis you'll be able to do. And your dataset will be more resilient in future years when comparing/contrasting with historical data, especially if you ask new questions.

But, the data is more sensitive and identifiable as you get more specific. Depending on your team size, company culture, and organizational trust, it might make sense to define the identities more broadly. This will protect individuals from being potentially identifiable from data points. And it can make team members feel safer providing data.

When putting this list together, it's a good time to define the vocabulary you're using. Take advantage of the opportunity to educate yourself and your leadership team. Make sure you're using the correct and respectful terms. These definitions will create a shared understanding for your team during the audit.

During our audit, we identified the following demographic groups we wanted to analyze for equity discrepancies:

  • Gender identity
  • Gender
  • Race/ethnicity/cultural background
  • Sexual orientation
  • Health or disability status
  • Age
  • Caretaker responsibilities

Due to the nuance caused by geography, national origin, politics, and various historical contexts around race, ethnic, and cultural identities, we opted to combine these 3 categories into a single demographic category for the purposes of our audit and respectfully allow our employees to identify in whatever way(s) they recognized and felt comfortable with.

3. Create remediation plan

Before collecting any data, create a remediation plan to address any equity disparities for current employees uncovered as a result of the audit.

By planning remediation steps before any data is collected, you'll ensure that the process is fair and just. Knowing results before knowing how you'll remediate negative findings could bias your decisions.

When creating your remediation plan, be sure to answer:

  • When will a remedial pay adjustment take place?
  • Will a remedial pay adjustment include back pay? (this is uncommon)
  • What level of discrepancy will qualify for a remedial pay adjustment? (3%? $1,000? etc.)
  • What is the budget for addressing all remediations?
  • If there are budget or timeline constraints, how will remediations be prioritized?
  • How will remedial pay adjustments be communicated to affected individuals and to the team?

4. Identify key participants

A pay equity audit is a lot of work. To prevent the initiative from stalling out, identify who will be responsible for what during the audit. Define each person's role and assign their tasks and/or deliverables.

There are many different ways to organize this work depending on each person's current responsibilities in their job role. In our audit, this is how we defined the roles and organized the work.

  • Audit lead
    • Responsible for compiling and anonymizing all the data for analysis
    • Best if this is someone who is already responsible for managing the team’s sensitive PII
    • Should have no authority to make salary or promotion decisions
  • Audit analyst
    • Responsible for analyzing the data and finding patterns
    • Best if this is someone who is an objective third-party
  • Audit sponsors
    • Responsible for supporting and promoting the audit
    • Responsible for remediating any findings gathered during the audit
  • Audit participants
    • Responsible for providing data for the audit
    • Only people who opt-in will be participants
  • Legal consultant
    • Educates audit sponsors on the potential liability and risk incurred from doing an equity audit

5. Present equity audit methodology and remediation plan to team

Present the methodology and remediation plan to your team. A pay equity audit is a large and sensitive company initiative.

Be sure to include the sampling dates and demographic data you'll be collecting and analyzing, who will be involved and how, and the expected timeline for the audit.

Keeping your team informed builds trust that you are considering the impact of the audit and will take care of their data. Gathering their feedback ensures their concerns are heard and that they are active participants in the process as early as possible.

Listen to your team's feedback and incorporate that into your plan. If the feedback is substantial, take the time to iterate on your plan and present the updates to your team before moving forward.

6. Collect historical data

If you choose to run a historical audit, collect the data you have available for each sampling date you defined in step 1. If you are not running a historical audit, skip to step 9.

Because this is historical data, you will most likely be limited to the data you have available in your HR systems. That's ok. This limited data set can act as an MVP for your data analysis. And it can help you hone in on data you especially want to make sure to collect for your next audit.

Once collected, have your Audit Lead anonymize the dataset. Remove any employee names, employee ids, and other identifiers.

7. Provide anonymized, historical data to Audit Analyst

Give this anonymized dataset to your Audit Analyst for analysis. Be sure to allow time in your timeline for them to complete a thorough analysis and compile their findings.

This step is mostly a waiting game for everyone except the Analyst :)

8. Present findings to the team

Once your Audit Analyst has completed their analysis, schedule time to present the findings to the team. Even if there are negative findings that cannot be remediated (due to the historical nature of this audit).

It's important to demonstrate transparency and integrity with your team early during this process. When they see how the data is being used and presented, they may feel more comfortable participating in the current audit.

If there are data shortcomings, share those. E.g. we couldn't analyze based on sexual orientation because we don't collect that data in our HR systems. This can help the team understand why some data seems to be missing in the historical analysis. And will help them understand why certain questions are included in the upcoming demographic survey.

One of my favorite things that we did during our audit was having our Audit Analyst present the findings himself. During his presentation, he educated the team on different statistical analysis strategies and explained how the math worked. The team really enjoyed this. Understanding the underlying math made them connect more with the whole process.

9. Collect current data

Now it's time to start the current, present-day equity audit. Participating in the audit should be opt-in for all team members. If someone doesn't want to participate and provide their data, respect that.

Gather demographic data, salary data, job titles, promotion dates, and other salary adjustment dates for each participant.

9a. Gather data from current team members

If you choose to send out a survey for Audit Participants so they can self-report, make the questions multiple choice and allow team members to select as many options as they want.

Always provide an "Other" option so they can fill in their answer if none of the existing options fit them.

And always provide a way to opt out of a question. Someone may feel comfortable disclosing some identities but not others.

Do not attach name, email, employee id, or any other identifier to self-reporting responses.

If your team members are self-reporting, be sure to provide a due date for your team to complete their responses.

9b. Gather data from HR systems

When we ran our audit, we provided the option for team members to choose to have their salary data, job titles, promotion dates, and other salary adjustment dates pulled from our HR systems. In this case, the data would not be fully anonymous since our Audit Lead would have to get this data from the HR system for them. But many employees on our team chose this option for convenience (and because our Audit Lead already had access to this data).

If you will be using data from HR systems for your audit, plan time for your Audit Lead to pull this information.

9c. Combine provided data into a dataset

If you chose a hybrid approach of both self-reported data and data from your HR systems, your Audit Lead will need to combine all the information into a single dataset.

Once collected and combined, have your Audit Lead anonymize the dataset. Remove any employee names, employee ids, and other identifiers.

10. Provide anonymized, current data to Audit Analyst

Like before, give this anonymized dataset to your Audit Analyst for analysis. Because you will likely have a more robust dataset for this part of the audit, they will need more time for the complex analysis.

This step is another waiting game for everyone except the Analyst :)

11. Present findings to the team

After your Audit Analyst has completed their analysis, present the findings to the team.

If there are findings that need to be remediated, reaffirm your commitment to the remediation plan that you already shared with the team. Remind them how affected individuals will be communicated with and when they can expect action.

In my experience, the team is interested in all the findings including the positive ones! Charts and graphs were my team's favorite. They had a lot of questions about what the data could and couldn't tell us.

12. Enact remediation steps as necessary

Follow through on your remediation plan. Hopefully, you did not find any evidence of pay discrepancies. But if you did, now is the time to address them so that you have a more fair, inclusive organization going forward.

Because you already did the work to outline the remediation process, this should be relatively painless for your managers and leadership team. It will be straightforward to know who needs a pay adjustment and how much to adjust.

13. Conduct retrospective on the audit process

After you've had a little time to digest everything you learned, retro the audit process. Discuss what went well, what didn't, and what you'd like to change for next time.

You should plan to conduct these audits regularly–every year or at least every couple of years. Having a retrospective while the latest audit is still fresh helps make improving the next audit easier.

When we retro-ed our audit, we had suggestions about identities to add to the survey that we had overlooked as well as some formatting improvements.

Our team validated how much they appreciated having the process and timeline communicated before its start. And we also learned that the education lesson about statistical methods was so fun that we need to make that a feature in all future audits 🤩


A pay equity audit is a big undertaking. It requires a lot of planning and effort from everyone involved. But it has a huge payoff. You'll build trust with your team, a more inclusive and fair culture, and a reputation for equitable labor practices.

Investments like this make it easier to recruit and retain talented people. Good people make great teams.

Are you considering running a pay equity audit? Have questions about how I did it? Feel free to reach out! You can email me or find me on Twitter.

Well-Rounded Dev

Liked this post? Subscribe to receive semi-regular thoughts by email.

    I won't send you spam. Unsubscribe at any time.