Salary benchmarking data sources compared for small employers and how to blend them
Salary benchmarking data comes from six kinds of source, and none is complete. Federal wage surveys such as BLS OEWS are broad, free and a year or more old. Employer surveys are current and detailed but cost five figures and require you to submit your own pay. Self reported salary sites are fast and biased. Posted ranges are fresh and skew to hard to fill roles. A defensible benchmark names its source, dates it and blends two of them at stated weights.
If you own pay at a company of 10 to 200 people, you will be sold data by people who benefit from one particular answer. This page compares the sources on the things that matter when you have to defend a number: who supplied it, how old it is, what it leaves out, and what it costs. At the end is a rule for what to do when two sources disagree by 10 percent, which happens constantly.
The six sources compared
| Source | Who supplies it | Age | Main bias | Cost | Best use |
|---|---|---|---|---|---|
| BLS OEWS | Federal survey, about 1.1 million establishments over three years | Reference May, published the next spring | Base wages only, no company size or level split | Free | Anchor for most non-executive roles |
| Employer surveys | Participating companies, run by consultancies | Annual cut, often months stale on delivery | Large employers over represented | Quote, often five figures | Executives, specialties, total cash and equity |
| Salary aggregators | Self reported entries from workers | Rolling | Self selected, titles not matched to duties | Free to browse | Sanity check on one role |
| Posted ranges | Employers, in job ads under pay transparency laws | Current | Hard to fill roles, wide ranges | Free, hand collection | Check on freshness and local competition |
| State agency data | State labor departments, often built on OEWS | Annual | Same limits as OEWS, finer geography | Free | Metro level checks for one state |
| Peer HR system data | Clients of a payroll or HRIS vendor | Near real time | Skews to the vendor's customer base | Bundled or quote | Tech companies with equity |
Why the federal data is the usual anchor
The BLS Occupational Employment and Wage Statistics program is the only source in the table whose method is published in full and whose sample is large enough to cover a small metro. For an HR specialist, the May 2025 national annual median is $75,940, with the 25th percentile at $58,610 and the 75th at $99,380. Those three numbers are free, dated and checkable by anyone, including the candidate or employee who asks where your range came from. That is a large part of what makes a number defensible.
The cost is age. A May 2025 reference period applied to pay that takes effect in January 2027 is 20 months old. The standard fix is to age the figure with the BLS Employment Cost Index or a stated annual rate. At 3.5 percent a year the factor is 1.059, and the HR specialist median becomes $80,400, about $4,500 higher. A company that skips the aging step ranks itself a step behind the market on the day the new pay applies. Our salary benchmarking tool page shows that arithmetic for three roles side by side.
Are salary aggregators reliable enough to set pay?
Not on their own. Salary aggregators collect what workers choose to report, so the sample leans toward people who changed jobs recently, titles are matched by name instead of by duties, and a mid sized metro can show a range built on a dozen entries. They are useful in one way: as a quick test of a number you got somewhere else. If your aged federal median and the aggregator's middle value are within about 8 percent, you have confirmation. If they differ by 20 percent, find out why before you post a range. Our look at whether Glassdoor salary data is accurate goes through the biases in detail.
What to do when two sources disagree
Disagreement is normal. Use a written rule instead of picking the number you like. A workable one for a small employer, shown here with illustrative figures for a role priced at $80,400 from aged federal data and $88,000 from the median of a dozen current local postings:
- Pick the anchor. Aged federal data at 60 percent weight, because its method is published.
- Pick the check. Posted ranges at 40 percent weight, because they are current and local.
- Blend. 0.6 x $80,400 + 0.4 x $88,000 = $83,440, rounded to $83,500.
- Flag the gap. The sources differ by 9.5 percent, so mark the role as hot and reprice it at the next requisition, not next year.
- File the record. Source names, dates, weights and the result, so the number survives the next question.
The percentages are a choice, not a law. What matters is that they are written down before you look at the result. Posting laws raise the stakes here. In states that require a pay range in the ad, such as Colorado, the range has to be a good faith estimate, and a record like the one above is what you show if anyone challenges it. The pay transparency guide lists the states and the rules, and market pricing a job walks through the matching step that comes before blending.
When to pay for employer survey data instead
Free sources stop being enough in three situations. You price executive roles, where federal data is thin at the top percentiles. You pay equity or total cash incentives that OEWS does not report. Or you hire a niche specialty, such as a regional clinical lead, with too few federal observations. In those cases a licensed survey is worth the quote, and our comparison of salary survey providers shows who sells what and how access works. Pricing it first is wise: the salary survey cost breakdown shows what a company of your size typically pays and what you must submit in return.
Relocating a hire is one case where the location differential in your data matters most. If a candidate is moving from a cheaper metro, price the role for the destination, then use an AI trip planner such as one that builds a day by day itinerary for the look and see trip so the visit does not become a project of its own.
How WageList uses these sources
WageList anchors every role on BLS OEWS wage data, ages it forward at a stated rate, adjusts for level and returns a midpoint and band with the source and date attached, so the blending and record keeping described above are done for you. It costs $99 a month or $588 a year for teams of 10 to 200, and you do not submit your own pay data to use it. It does not report bonus or equity, so if those drive your offers, pair it with a survey for those roles. The job pricing software page covers the role by role method.
Sources: U.S. Bureau of Labor Statistics Occupational Employment and Wage Statistics, May 2025 national estimates for Human Resources Specialists (13-1071) read from the BLS public data API, and the BLS Employment Cost Index. The blending example uses illustrative inputs.