Reading Wage Percentiles: How to Interpret 10th to 90th Percentile Pay Data

Why the Numbers on a Wage Table Can Mislead You — Until You Know What They Mean

Picture this: you're an HR generalist at a 70-person professional-services firm, and your hiring manager just forwarded a LinkedIn job post from a competitor offering $95,000 for the same Accounting Manager role you have budgeted at $78,000. You pull up the BLS Occupational Employment and Wage Statistics (OEWS) data to check where $78,000 actually sits in the market — and you're staring at a row of numbers labeled H_PCT10, H_MEAN, A_PCT25, A_MEDIAN, A_PCT75, A_PCT90. If you've never been taught to read a wage distribution, those columns look like alphabet soup.

The good news: wage percentiles are a precise, learnable tool. Once you understand what each column is saying, you can answer three questions that matter for every compensation decision you make — what the market actually pays, where your roles sit relative to that market, and which benchmark to target when you set a salary band. This article gives you a working framework for all three.

What a Wage Percentile Actually Tells You

A percentile is a rank-order statement about a population. The 10th percentile wage for a given occupation means: 10% of workers in that occupation earn at or below this amount, and 90% earn above it. The 90th percentile means the opposite — 90% earn at or below this amount, and only 10% earn above it.

Applied to BLS OEWS data — which draws wage estimates from a sample of roughly 1.1 million establishments covering approximately 830 occupations nationwide and across around 530 geographic areas — the percentiles describe where real employers are actually paying real workers right now, not what any single job description says.

Here is the full span you will typically see in an OEWS data file:

Column	What it tells you
P10	Floor of the market — new entrants, lowest-paying employers
P25	Below-median market — cost-conscious employers, developing roles
P50 (Median)	The true midpoint — half earn more, half earn less
P75	Above-median market — competitive employers, experienced workers
P90	Top of the market — specialized, high-demand, or premium-paying employers

The distance between P10 and P90 is the wage spread for the occupation. A wide spread (common in management, technology, and sales roles) tells you the market rewards experience and employer quality unevenly. A narrow spread (common in tightly regulated trades or heavily standardized roles) tells you most employers pay similarly regardless of seniority.

Mean vs. Median: Which Number to Use for Band-Setting

The OEWS table also reports a mean wage (labeled A_MEAN or H_MEAN for annual/hourly). Many people reach for the mean first because averaging feels familiar. For compensation decisions, the median is almost always the more reliable anchor — and the difference matters.

Here is why: the mean is pulled upward by the highest earners in the occupation. A small number of executives or specialized practitioners earning two or three times the typical wage can drag the mean significantly above the median, making the occupation look better-paid across the board than it actually is for the majority of workers.

The median — the 50th percentile — is the one statistic that accurately describes what the employer in the middle of the market is paying. Use it as your band midpoint anchor unless you have a deliberate reason to target above or below the market center.

The BLS OEWS all-occupation annual mean wage was $69,770 in the May 2025 release. That figure appears frequently in press coverage, but it reflects mean skew across all occupations. When you are setting a band for a specific role, the occupation-level median from the relevant geographic cut is the number you want.

For a deeper walk-through of how the OEWS data is structured and how to download the right file for your state or metro, see our BLS OEWS benchmarking guide.

The Anatomy of a Percentile Row: A Worked Example

Suppose you are benchmarking a Staff Accountant role (SOC 13-2011, Accountants and Auditors) in a mid-sized metro area. Your downloaded OEWS table shows the following annual wage data:

P10	P25	P50 (Median)	P75	P90
$45,100	$54,800	$70,500	$91,200	$118,600

(These are illustrative figures structured around the shape of typical OEWS accountant data; always pull current figures directly from the BLS for your specific metro area.)

Reading across the row:

• $45,100 (P10): This is the entry floor. Employers paying here are typically offering roles with limited scope, in lower-cost sub-markets, or to recent graduates with no experience. If your offer is at or below this number, you are competing with the bottom tier of the market.
• $54,800 (P25): A below-median but not-floor position. Common among smaller firms, non-profits with constrained budgets, or companies in lower cost-of-labor geographies.
• $70,500 (P50): The market midpoint. An employer targeting this wage will attract and retain a typical experienced accountant in this metro. This is the standard band midpoint anchor.
• $91,200 (P75): Above-median. Employers here are signaling that experience and tenure pay off, or they are competing in a talent-tight segment.
• $118,600 (P90): Top-of-market. Senior accountants with specialized skills (e.g., technical accounting, M&A experience), or employers in high-demand sectors paying a premium.

Your $78,000 budget from the opening scenario sits between P50 and P75 — approximately the 60th–65th percentile, depending on the exact distribution. That is a competitive position, not a lagging one. The $95,000 competitor offer is operating near P75. Neither is wrong; they reflect different talent strategies.

If you need help matching the role to the right SOC code before pulling these numbers, the guide to matching jobs to SOC codes walks through the classification process step by step.

How to Choose a Target Percentile for Your Organization

Selecting a target percentile is a compensation philosophy decision, and it should be deliberate, documented, and consistent across your job levels rather than re-litigated role by role.

Common approaches:

• P50 target (market-median positioning). The most widely used strategy for organizations with balanced talent competition. You are aiming to attract workers who are typical for the occupation — not underpaying, not overpaying. Band min/mid/max are built around the median as the midpoint.

• P60–P75 target (lead-the-market positioning). Used by organizations competing in tight labor markets, high-turnover occupations, or roles where replacement cost is high. The band midpoint is set above the market median, so employees at full performance are paid above the majority of peers.

• P40–P45 target (lag-the-market positioning). Viable only when offset by other strong total-compensation elements (equity, schedule flexibility, benefits). Risks are real: employees can easily verify their market position against public pay transparency postings, so a lagging cash position needs a credible non-cash story.

The critical discipline is consistency across job levels. If you target P50 for individual contributors and P75 for managers without a documented rationale, you will eventually face questions from employees — and in pay-transparency jurisdictions, those questions may arrive in writing. Define your target percentile by level or job family and hold to it.

Once you have a target percentile selected, the next decision is whether to benchmark against national, state, or metro-level data. A metro-area rate for a high cost-of-labor market can be substantially different from the national figure for the same occupation. See national vs. metro wage data for guidance on which geographic cut to use.

Building a Salary Band from a Percentile Anchor

With a target percentile chosen as your midpoint, you build a salary band by applying a range spread above and below it. The range spread — expressed as the percentage from minimum to maximum as a fraction of the minimum — defines how much room employees have to grow within a band before hitting the ceiling.

A common approach:

1. Set the midpoint at your target percentile (e.g., P50 = $70,500 for the Staff Accountant example above).
2. Choose a spread. A 50% spread (minimum is 80% of midpoint; maximum is 120% of midpoint) is a common default for professional individual-contributor roles.
3. Calculate:

• Minimum: $70,500 × 0.80 = $56,400
• Maximum: $70,500 × 1.20 = $84,600

That gives you a band of $56,400–$84,600, anchored to the market median. An employee hired at $78,000 sits at roughly 92% of midpoint — a healthy position with room for continued growth before hitting the ceiling.

The ratio of an individual's salary to the band midpoint is called the compa-ratio (short for comparative ratio). A compa-ratio of 1.00 means the employee is paid exactly at midpoint; above 1.00 means above midpoint; below 1.00 means below. Tracking compa-ratios across your workforce surfaces pay equity patterns quickly — a cluster of employees from one demographic group sitting consistently below 1.00 is a signal worth investigating before it becomes a compliance issue.

For a fuller explanation of how compa-ratio integrates with band structure, see our piece on the band positioning indicator explained. And for the complete process of building salary bands from scratch — from SOC matching through range-spread decisions — see how to create salary bands.

Putting Wage Percentiles to Work in Your Next Compensation Review

Reading wage percentiles correctly changes what compensation reviews look like in practice. Instead of anchoring to what you paid last year (or what a competitor posted on LinkedIn), you are working from a documented, reproducible benchmark that you can defend to a hiring manager, a candidate, or an auditor.

The workflow, summarized:

1. Match the role to the correct SOC code.
2. Download the current OEWS file for your geographic level (national, state, or metro).
3. Read across the percentile row — P10 through P90 — to understand the full wage distribution, not just the median.
4. Identify where your current salary or offer sits in that distribution.
5. Compare that position to your documented target percentile for that job level.
6. If the position is materially off-target, flag it for correction in the next review cycle or compensation action.

That is repeatable, auditable compensation benchmarking — the kind that holds up when an employee asks why they are paid what they are paid, and the kind that positions you well as pay-transparency requirements continue to expand.

If you want a practical reference for every step of the OEWS workflow — from file download to percentile interpretation to band construction — sign up for the Job Band Builder newsletter. We send one focused compensation-management resource each month: no noise, no vendor pitches, just structured guidance for HR professionals building compensation programs without a dedicated comp team. ```