How Many Pay Grades Should You Have?

How Many Pay Grades Should You Have?

Someone on your leadership team has an opinion about this, and it is usually a round number. Eight, because eight feels clean. Twenty, because the old structure had twenty. Neither answer comes from anything you could defend in a pay equity review.

The number of grades in your structure is not a style choice. It falls out of three things you have already decided or measured: the distance between your lowest-paid and highest-paid jobs, how big a step you want between grades, and how many genuinely distinct levels of work your job evaluation results show. Get those right and the grade count calculates itself. Get them wrong and you spend the next three years explaining why a promotion came with a 2% raise. Here is the arithmetic, the four tests a structure has to pass, and what failure looks like in both directions.

TL;DR

• Your grade count is set by pay span and midpoint progression, not by headcount or org chart depth.
• Use this: grades ≈ ln(highest midpoint ÷ lowest midpoint) ÷ ln(1 + progression rate), then add 1.
• Most single-structure employers land between 10 and 20 grades. Under 8 usually means too much manager discretion; over 25 usually means promotions that do not pay.
• Midpoint progression of 8–12% suits professional populations; 15% and up suits steep executive ladders.
• Let job evaluation point scores set the boundaries, then test overlap, promotion increment, market fit, and discretion.

The short answer, by organization size

Size is a proxy, not a rule, but it is a useful sanity check. If your draft is far outside this range for your headcount, something in your inputs is off.

For a reference point outside the private sector, the U.S. federal General Schedule runs 15 grades, GS-1 through GS-15, covering nearly the entire white-collar civil service (OPM). Fifteen grades for more than a million employees should tell you headcount is not the driver. Pay span is.

The formula that sets the number

Your grade count is a function of how far you have to climb and how big each step is.

Say your lowest grade midpoint is $45,000 and your highest is $260,000 — a 5.8x span. If you want each midpoint to sit 12% above the one below it, the calculation is ln(260,000 ÷ 45,000) ÷ ln(1.12) + 1, which is 1.754 ÷ 0.1133 + 1, or about 16.5. Call it 16 or 17 grades.

Change the progression to 8% and the same span needs 23 grades. Change it to 18% and you need 11.

That is the whole trade-off: wider steps mean fewer grades, and fewer grades mean wider steps. You cannot have a flat 8-grade structure and a meaningful promotion increment when your top job pays six times your bottom job. The math will not allow it.

Pick your progression first, because it has a defensible logic behind it. Midpoint progression should roughly match what the market pays for one real level of added responsibility. For most professional and technical populations that is 8–12%. For management and executive tiers, where each step adds substantially more scope, 15–20% is normal. Below 7%, employees will not feel a promotion. Above 25%, you have skipped a level that exists in the market, and someone will sit at a grade maximum for years.

Start from job evaluation points, not the org chart

The formula gives you a target count. Job evaluation tells you where the boundaries go.

If you run a point-factor evaluation, every job carries a score built from weighted compensable factors — skill, effort, responsibility, and working conditions, plus whatever sub-factors your plan uses. Sort all jobs by total points and plot them. You will see clusters and gaps. The gaps are your grade boundaries, and they are the only boundaries you can defend when someone asks why the Senior Systems Engineer sits in grade 11 and the Engineering Manager sits in grade 12.

Three rules for drawing those lines:

• Cut at the gaps, not at round numbers. A 40-point hole between 520 and 560 is a boundary. Splitting a tight cluster at 540 because it is neat creates two grades nobody can tell apart.
• Widen point bands as scores rise. A 30-point difference matters more at the bottom of the scale than at the top.
• Check that every grade holds more than one job. A grade containing one job is not a grade. It is a special deal with a number on it.

The full mechanics are in our guide on converting job evaluation points into pay grades.

Running a point-factor evaluation and want the boundaries drawn for you? See how PointFactors does it.

Four tests your grade count has to pass

1. The overlap test

Adjacent grades should overlap, but not so much that promotion is meaningless. Overlap is the lower grade's maximum minus the higher grade's minimum, divided by the lower grade's full range.

Work it through. Grade 5 has a midpoint of $80,000 and a 40% range spread, putting its minimum at $66,700 and its maximum at $93,300. At 10% progression, grade 6's midpoint is $88,000, with a minimum of $73,300. Overlap is (93,300 − 73,300) ÷ (93,300 − 66,700) = 75%.

Seventy-five percent overlap means a well-paid person in grade 5 already earns more than most of grade 6. Promotions will feel like nothing. Widen progression to 15% and grade 6's minimum rises to $76,700, dropping overlap to 63% — still generous, but now a promotion moves someone. Anything above 70% signals too many grades for your spread. Our guide on pay grade overlap covers targets by job level.

2. The promotion increment test

Take five real promotions from the last year. Under the new structure, does each produce a raise your managers would be comfortable announcing? If a grade jump yields less than about 6–8% for someone already mid-range, your grades sit too close together.

3. The market fit test

Pull market data for the benchmark jobs in each grade and look at the scatter. If grade 9 holds jobs whose market medians run from $92,000 to $141,000, that grade is doing too much work. A grade whose benchmarks span more than about 30% of market pay will generate constant off-structure exceptions.

4. The discretion test

Ask what range a manager can move within without approval. In an 8-grade structure covering a 6x span, a single band might run $70,000 to $145,000. That is not flexibility. It is an unmonitored pay equity exposure, and it is the documented failure mode of aggressive broadbanding.

What too many and too few look like

The two extremes fail differently. Too few grades fails quietly, in offer letters and in the gaps you discover during an audit. Too many grades fails loudly, the first time someone gets promoted and sees a $1,900 raise.

Most employers still run traditional graded structures rather than broad bands. In the Culpepper survey data SHRM published on salary range structure practices, 78% of companies with formal structures used traditional grades, 11% used broadbands, and 9% ran a hybrid (SHRM). WorldatWork's research on salary structure policies and practices tracks how those choices shift as pay transparency rules and remote work push employers to re-examine structure design.

Worked example: a 600-person software company

Lowest midpoint $52,000 (Customer Support Associate). Highest in the main structure $245,000 (VP Engineering), with the C-suite handled separately. Span is 4.7x. At 11% progression: ln(4.71) ÷ ln(1.11) + 1 = 15.8, so 15 or 16 grades.

Point scores showed clean gaps in 14 places, which gave 15 grades. The two jobs sitting awkwardly between grades 7 and 8 turned out, on review, to be the same job under two titles. That happens constantly — grade boundary work is also job catalog cleanup.

The company then split executives into a second structure at 18% progression, because five exec roles had market medians too far apart for an 11% ladder. Different job families can share a grade structure; populations with different market slopes usually should not.

When to add a structure instead of grades

Add a second structure when a group's pay curve has a different shape, not just a different level:

• Executives. Steeper progression, wider spreads, often no formal maximum.
• Sales. Base bands compress because variable pay carries the load.
• Geography. If zones differ by more than about 15%, apply a differential to one structure rather than building a separate ladder. See geographic pay differentials.

Adding grades to accommodate one unusual population distorts the structure for everyone else.

How to decide, in order

1. Set the lowest and highest midpoints for the population the structure covers.
1. Choose midpoint progression based on what one real level of scope is worth in your market.
1. Run the formula for a target grade count.
1. Draw boundaries at the gaps in your job evaluation point distribution.
1. Test overlap, promotion increment, market fit, and manager discretion.
1. Adjust progression — not the grade count directly — if a test fails.

Step 6 is the one people skip. When the overlap test fails, the instinct is to delete a grade. The better move is to widen progression and let the count fall out of the new math, so the structure stays internally consistent.

FAQ

How many pay grades does a small company need? Five to eight is typical under 100 employees. The constraint is pay span, not headcount: if your highest-paid role earns three times your lowest, six grades at 12% progression covers it comfortably.

Should every job family have the same number of grades? No. Families can share one structure while occupying different slices of it. Support might live in grades 2–7 and engineering in grades 5–14. What they share is the grade definitions and the ranges attached to them.

What midpoint progression should I use? 8–12% for professional and technical populations, 12–15% for management, 15–20% for executive tiers. Under 7% makes promotions feel hollow; over 25% suggests a missing level.

Can I have too much overlap between grades? Yes. Above roughly 70%, a strong performer at the top of a grade earns more than most people in the grade above, which makes promotion economically pointless. Fix it by widening progression or narrowing range spreads, not by moving individuals.

Do pay transparency laws change how many grades I need? Not directly, but they raise the cost of getting it wrong. Posted ranges make your structure public, and a grade whose posted range spans $80,000 invites both candidate distrust and regulatory attention.

How often should I revisit the grade count? Review ranges annually, but revisit the number of grades only when the shape of the organization changes — a major acquisition, a new business line, or a job architecture redesign. Structure churn is expensive and confusing.

What if my point scores do not cluster cleanly? That usually means thin job documentation rather than genuinely continuous jobs. Re-check the job descriptions for the ambiguous roles before redrawing boundaries.

The grade count that survives scrutiny is the one you can trace back to a number: a measured pay span, a chosen progression rate, and a point distribution anyone can read. Show that chain and the structure holds up in a comp committee, in a pay equity audit, and in a conversation with an employee who wants to know why their job sits where it does.

PointFactors scores your jobs against weighted compensable factors and produces the point distribution your grade boundaries should be built on — in days, not quarters. Book a demo or see pricing.

Justin Hampton is founder and CEO of PointFactors.