A compensation plan self-audit may seem intimidating, and it's true that you may want to seek expertise to help you the first time through. But once the tracks are laid, you can follow them every year or so to learn whether your compensation plans are doing what you want them to--or whether they are not.

Sometimes, an audit of your compensation plans is required, and sometimes it is voluntary. "One of the main reasons a company would audit their compensation plans is if they have a legal obligation to do so," says Thomas. "If a company contracts to provide services to the federal government, for example, the Office of Federal Contract Compliance Programs requires them to conduct a compensation audit, either themselves or through outside counsel or consultants. Other companies do it to ensure that they're living up to their commitment to a fairness standard. Usually it is done to get a deeper understanding of the compensation process within the organization."

Both are good reasons. Either way, says Thomas, the first step is a close review of your compensation structure. You should easily be able to see how and why people are paid. "If you have a compensation structure that is transparent and consistent, it should be relatively easy to identify those factors that determine how people are paid. If you can't identify the reasons somebody is paid what they're paid, then your policy and your structure is probably not transparent, and it's probably not consistent. That should be a red flag. Transparency and consistency are the keys to a good compensation policy."

Begin with the data

One of the key components of a compensation self-audit is organization of the data. You may find that the data you typically gather on employees is not sufficient for an audit. Thomas says that categorizing people properly is critical. "People who are similarly situated, meaning they have the same kinds of job responsibilities or job functions, should be grouped together. You don't want to compare the CEO with his administrative assistant, for example, with regard to compensation. Those two serve very different functions in the company. You need to look for people who are similarly situated, and group them along functional lines, functional titles, levels of reporting responsibility, things of that nature."

The audit may require data that you have not collected, or that would be too time consuming (and therefore expensive) to recover. For example, job experience may be a factor you want to consider. Finding and entering that figure would be costly, in most cases; perhaps requiring a hands-on search through old resumes. Instead, some companies use age at hire as a stand-in, Thomas says. The assumption is that an employee had acquired relevant experience before you hired him or her. This kind of proxy variable can be useful, but is not always accurate, Thomas cautions. "When you use these proxy variables, you have to be very careful, because they don't necessarily measure what you think they're measuring." A close review may be necessary.

Close analysis considers 'edge factors'

Once you have properly grouped your employees, your audit will likely uncover some that, on the surface, should be paid similarly but are not. Thomas says further analysis is a critical component of the compensation self-audit, often revealing distinctions in abilities, experience, or other, less tangible differences. She calls these edge factors. "An edge factor might be experience," she explains. "People with longer tenure are going to tend to be more highly paid than those people who have just recently been hired."

Another edge factor might be a professional designation. But even these may not be relevant in every case. For example, two employees with CPA licenses and identical time on the job may, on the surface, appear to command the same salaries. But what if one is employed in the accounting department, and the other is in sales? The CPA is not relevant to the job skills of the latter, but may be critically important to the former. Therefore, the CPA may explain why an accountant who has this designation is paid more than one who doesn't, but it doesn't explain why one salesperson is paid more than another salesperson.

All of this leads up to the fact that, once you have conducted a compensation self-audit using as much technology as is available to you, you still need review the results using a human touch. "Follow-up is so important," says Thomas. "If your audit reveals that Joe Smith is paid in accordance with everybody else, but John Jones is way out of line, then the person working on the follow-up can take a closer look. You might find that John Jones has an edge factor that you didn't (or couldn't) measure. In other words, there is a reasonable explanation for the difference."

Self-audit regularly

In most cases, Thomas recommends that companies perform a thorough audit of their compensation programs at least annually. But sometimes, a more frequent schedule would be wise. "If you've got rapid expansion, if you're a new entrant (into a particular business sector), if you're expanding your workforce and showing very strong growth, you might want to do this every 6 months," she says. "It really depends on the circumstances of the company, what the budget allows, and what problems they've seen before. If you find that your compensation structure is in a bad state, that it's not transparent and inconsistent, you need to remedy that right away. You might want to do an audit, and then follow it up in 3 months or 6 months to see if the changes you made really did what you wanted them to."

Thomas recommends getting the legal department or outside legal counsel involved in your audit before you even start the process. "You need to get the legal perspective on things, because there are confidentiality issues and attorney-client privilege issues. Talking to the legal department will help you cover your bases and make sure you're doing things correctly."

You may be thinking about the expense involved, and Thomas admits that the first time through can be time-consuming, and therefore expensive. But it's a lot like an investment in a good set of tools. "A lot of the cost is really front end," she says. "Collecting the data, figuring out what data you should be collecting, structuring the model, really digging in to understand how and why people are paid. But as long as you don't have any major changes to your compensation system, you can use all that knowledge in subsequent years. It becomes much easier to do." Once you have the proper tools, you can use them many, many times and get the results you want: a project that works.