Pay equity analysis software and what to check before you buy
Pay equity analysis software is worth buying only if it can tell you which salary in a gap is the wrong one, and for a company under 200 people that takes a sourced market band more than it takes a regression engine. Most tools in this category were built for employers with thousands of staff, and the demos hide that. This is the checklist we would use before paying for one, with the arithmetic behind each question.
The product side is on the pay equity software page. This article covers the buying decision itself, meaning what to test in a trial and which vendor claims to discount.
Does the tool have a market reference
A tool that takes your payroll file and returns an average gap between groups answers a narrow question. It says two groups differ. It does not say whether the lower paid group is below the market or the higher paid group is above it, and that distinction decides what you do. If men are overpaid against the market, you do not fix the gap by raising everyone to their level.
Ask in the trial where the benchmark for each job comes from, how old it is and whether you can see the method. Public wage data such as the BLS Occupational Employment and Wage Statistics program is a legitimate source and is documented on bls.gov. A vendor that cannot name its source is selling a black box you will have to defend yourself.
How many people per job the method needs
Regression-based pay equity analysis estimates the pay difference between groups after controlling for job, level, tenure and location. It needs enough people in each comparison cell to produce a result that means something. A company with 40 employees across 25 different jobs has one or two people in most jobs, and the model has almost nothing to compare.
| Headcount | Typical people per job | Method that holds up |
|---|---|---|
| 10 to 50 | 1 to 3 | Band and compa-ratio, person by person |
| 50 to 200 | 2 to 6 | Band and compa-ratio, plus group averages inside a job family |
| 200 and up | 5 or more in the big families | Regression, run through counsel when risk is real |
The headcount lines are a planning heuristic, not a statistical threshold, and a statistician will want to look at your actual groups. The point is that a vendor who sells you regression at 30 people is selling output the data cannot support.
What a gap costs to close
A useful tool prices the fix. Take twelve account executives with a band midpoint of $90,000. If the six women average $84,600 and the six men average $95,400, the average gap is $10,800. Bringing the women to the midpoint costs $5,400 each, or $32,400 a year before payroll taxes. Raising them to the men's average costs $64,800. The first number is a decision a founder can make in a meeting. A tool that produces neither gives you a finding with no budget attached.
Whether it covers the posted range problem
Pay equity and pay transparency are now one workflow. In states that require a range in the job posting, the range you publish is a public statement about how you set pay, and a gap between it and your current employees is what a claimant looks for. A tool that checks current employees but cannot produce a defensible range for the next opening leaves half the exposure open. The pay transparency laws by state overview shows where this applies.
If your company also pays a network of contractors or sellers and splits payouts between them, a payments platform such as fractionalpay.com handles the money movement. That is a separate problem from deciding what each role should be paid.
The privilege question
A pay equity analysis commissioned through an attorney may be protected as privileged, and one you run yourself in a self-serve tool generally is not. That is a reason to run the simple band review in the open, keep it factual, and bring in counsel before any statistical study of a group gap. The pay equity audit guide covers how to sequence the two.
What to test in a trial
- Load ten real roles and check whether the tool names the source and date of each benchmark.
- Ask for the gap for one role with only three people in it and see whether it warns you about sample size.
- Ask for the cost to close the gap at the midpoint and at the group average.
- Check whether the same data produces a range for a new posting.
- Ask whether you can export the record with the band, the date and the reasoning.
Pay equity analysis software questions
What does pay equity analysis software cost?
Enterprise platforms are sold on annual quotes sized to headcount and are priced for large employers. Self-serve benchmarking for small teams costs far less, and the WageList pricing page shows the plans for teams of 10 to 200.
Is a spreadsheet enough for a pay equity analysis?
For under 50 employees it can be, if the band comes from a documented source and the compa-ratio column is correct. The weakness is the band, not the arithmetic. Most spreadsheets rely on a survey someone downloaded once and nobody can source.
How accurate is a pay equity analysis?
It is as accurate as the job definitions and the benchmark behind it. Two people with the same title can do different work, and no tool can see that. Treat the result as a list of cases to review, not a verdict on any individual.
Start with the band
Build a sourced band for the role you are worried about, then compare each person to its midpoint with the compa-ratio calculator.