P90/P10 measures the ratio between top-end and bottom-end pay, making inequality and pay dispersion immediately visible in a simple, executive-friendly number. Tracked over time and segmented properly, it helps HR diagnose fairness risks, incentive design effects, and structural pay imbalances before they become engagement or retention problems.
P90/P10 is a percentile-based measure of pay spread.
It compares:
- P90 = the 90th percentile value (pay level near the top of the distribution)
- P10 = the 10th percentile value (pay level near the bottom of the distribution)
It is commonly used in compensation analytics as a robust dispersion metric alongside range and standard deviation because percentiles are less sensitive to extreme outliers than minimum/maximum.
Formula
P90/P10 = 90th percentile pay divided by 10th percentile pay.
What does it actually measure?
P90/P10 measures pay dispersion - how far apart higher and lower earners are within a defined employee population.
A simple interpretation:
"How many times higher is pay near the top than pay near the bottom?"
Example:
- P90 = 120,000
- P10 = 60,000
120,000 divided by 60,000 equals 2.0
This means pay near the top is 2× pay near the bottom.
Why HR should care
P90/P10 is not a statistical vanity metric. It acts as a distribution health signal.
It helps HR answer questions like:
- Are we becoming more unequal over time?
- Is our pay architecture widening faster than our job structure?
- Are incentives disproportionately benefiting a narrow group?
- Are we increasing perceived unfairness risk under transparency?
- Is leadership pay diverging from workforce pay?
A useful behavioral anchor:
Employees react to fairness narratives, not compensation math. P90/P10 helps you detect when the narrative is likely to fracture.
Interpretation Guidelines
P90/P10 has no universal "good" threshold. Interpret it within context and segmentation.
| P90/P10 Ratio | What It Suggests | HR Implication |
|---|---|---|
| ~1.2-1.4 | Tight distribution | Potential compression / limited differentiation |
| 1.5-2.0 | Moderate dispersion | Often consistent with healthy differentiation |
| 2.0-3.0 | Wide gap | Higher inequality and fairness sensitivity |
| 3.0+ | Very wide spread | Elevated cultural, retention, and governance risk |
Context multipliers (expect higher ratios):
- Mixed populations (executives + frontline together)
- Variable pay-heavy groups (sales, trading, deal roles)
- Rapid growth or restructuring periods
- Equity-rich reward systems
Rule of thumb: The trend and segmentation matter more than the point estimate.
A Few Real-Life Use Cases
1. Pay Equity & Internal Fairness Signal
Scenario: Employee relations complaints rise; leadership pay is questioned.
HR calculates total compensation:
- P90 = 180,000
- P10 = 42,000
- Ratio = 4.3
That is a very wide spread. It doesn't prove unfairness - but it increases the probability of perceived unfairness.
Action questions:
- Is leadership pay accelerating faster than workforce pay?
- Is bonus design creating disproportionate upside?
- Is transparency making gaps more salient and emotionally charged?
Better diagnostic move: Split the ratio into components:
- Base pay P90/P10
- Bonus P90/P10
- Equity/LTI P90/P10
Often, the widening is driven by variable pay and equity, not base salary.
2. Compression Risk Within a Job Family
Within a single job family and level:
- P90 = 82,000
- P10 = 76,000
- Ratio = 1.08
This indicates severe compression - the pay distribution has little room to differentiate.
What this can cause:
- Promotions feel financially trivial
- High performers see low reward delta
- External hires "leapfrog" internal pay
Actions to validate root cause:
- Time-in-role vs pay positioning (range penetration)
- Performance distribution vs increase outcomes
- Range width and midpoint progression adequacy
- Merit grid design (does it actually differentiate?)
3. Incentive Plan Diagnostics -Is the Spread Earned or Assigned?
Sales group:
- P90 total earnings = 250,000
- P10 total earnings = 70,000
- Ratio = 3.6
Is that healthy? It depends on whether dispersion reflects performance or structure.
Interpretation logic:
- If performance is truly skewed → dispersion may be appropriate
- If territory, quota setting, lead assignment, or manager discretion drives outcomes → dispersion may reflect structural unfairness
Pair with:
- Quota attainment distribution
- Territory value / opportunity allocation analysis
- Pay-performance correlation (do earnings track results?)
Goodhart's Law risk:
If upside is too extreme, people optimize the metric - sometimes via unhealthy competition or rule-bending. This is a plan design governance issue, not a moral failing.
4. Monitoring Culture During Growth
Fast-growing company:
- 3 years ago: P90/P10 = 1.9
- Now: 3.1
Likely drivers:
- Senior leadership hiring at high market rates
- Equity grants concentrated in upper levels
- Entry-level hiring surge lowering P10
What matters most: the trend line and what changed in the reward mix.
5. DEI & Pay Gap Risk Screening -Early Warning, Not Proof
P90/P10 does not prove discrimination.
But it can flag structural conditions where inequity risk rises.
If the ratio is widening:
- Check representation across pay percentiles
- Overlay gender/ethnicity distribution by decile
- Examine promotion velocity and time-in-level patterns
Treat it as an early-warning indicator, not a conclusion.
How It Compares to Other Spread Metrics
| Metric | What It Tells You | Weakness |
|---|---|---|
| Range | Max - Min | Highly sensitive to outliers |
| Standard Deviation | Average deviation from mean | Hard to explain; assumes distribution shape |
| Coefficient of Variation | Spread relative to mean | Too technical for many HR audiences |
| P90/P10 | Intuitive top-vs-bottom spread | Ignores what happens in the middle |
Why practitioners like P90/P10: it is simple to explain and relatively robust.
How to Operationalize It in HR Reporting
Use P90/P10 where it supports decisions:
- Annual compensation review deck
- Compensation committee reporting
- Pay equity diagnostics
- Incentive plan health checks
- Workforce fairness dashboards
Track and present:
- Trend over time (most important)
- By business unit / function
- By job level and job family
- By reward component (base vs variable vs equity)
Recommended visuals:
- Line trend over time
- Side-by-side comparisons across units
- Component split (base/bonus/equity) for the same population
When P90/P10 Misleads
Common mistakes that break interpretation include mixing job levels (e.g., executives and frontline roles), which inflates dispersion and hides actionable insight, and ignoring small sample sizes, where groups under roughly 30 employees can produce unstable percentiles. P90/P10 should always be segmented at minimum by level, job family, location, and employee type, otherwise the ratio becomes too blunt to guide decisions. It is also incorrect to treat a high ratio as automatically "bad," since some commission-heavy or performance-driven roles naturally produce wider outcomes. Avoid using P90/P10 in very small populations, highly standardized pay systems (e.g., union step scales), individual fairness investigations, or when a few atypical payouts dominate the distribution - use compa-ratio distributions, penetration analysis, or cohort comparisons instead.
P90/P10 ultimately helps HR answer whether the organization is widening or narrowing economic distance inside the company - and whether that shift is intentional and properly governed. Widening gaps matter because they change cultural signals, influence retention and discretionary effort, affect perceived fairness in transparent environments, and shape the employer brand and broader trust narrative.