How to Convert Job Evaluation Points into Pay Grades

How to Convert Job Evaluation Points into Pay Grades

You finished the hard part. Every job in your organization has a point score, the scores hold up under scrutiny, and your evaluation committee signed off. Then someone in Finance asks the obvious question: "So what does a 487-point job actually pay?"

That gap — between a defensible score and a defensible salary range — is where most job evaluation projects stall. Points measure internal value. Grades and ranges translate that value into money, and they have to reconcile with a labor market that never read your factor plan. This guide walks the conversion end to end: how to find grade boundaries in your own data, how to fit a market pay line, and how to build ranges that survive an audit.

TL;DR

• Sort every evaluated job by point score and look for natural gaps in the distribution — those gaps are your candidate grade boundaries, not arbitrary round numbers.
• Most organizations land on 8–14 grades. Fewer grades means wider ranges and less promotion friction; more grades means tighter internal control.
• Regress benchmark job scores (X) against market pay (Y) to get a market pay line. Aim for an R² of about 0.95 — a lower value usually means the evaluation, not the math, needs work.
• Shift the market line up or down to reflect your pay philosophy, then read the midpoint for each grade off the resulting pay policy line.
• WorldatWork research puts typical midpoint progression at 10–14% and market-based range spreads around 47–58%.

Step 1: Lay out the point distribution

Start with a single sorted list: every evaluated job, its point total, its incumbent count, and its current average pay. Nothing else yet.

Now plot the point totals on a number line. You are looking for two things — clusters and gaps. In a healthy point-factor evaluation, jobs do not spread evenly across the scale. They bunch up around organizational levels, because your factor weights encode what your organization actually rewards.

Say a 900-job manufacturer sees clusters at 180–240, 260–330, 355–430, and 470–560, with visible thinning between each. Those thin spots are your candidate boundaries. Drawing a break at 250 splits two jobs that scored 248 and 252 into different grades and invites an appeal you will lose. Drawing it at 252 — inside a gap where nothing sits — is defensible on its face.

If your scores are smoothly distributed with no gaps at all, that is a signal worth investigating. It usually means your factor degrees are too granular or your weights are too flat. Revisit how you weighted your compensable factors before you force grade lines onto data that does not want them.

Step 2: Decide how many grades you need

There is no universally correct number, but there is a real trade-off, and you should make it deliberately.

Most mid-sized organizations land somewhere in the 9–14 band. If you are leaning wide, read up on broadbanding first — it solves promotion friction but transfers a lot of pay discretion to managers, which is exactly the discretion a pay equity audit will scrutinize.

One practical test: after you set boundaries, look inside each grade. If a single grade contains an administrative assistant and a regional operations manager, your boundaries are wrong regardless of what the math said.

Step 3: Fit a market pay line

Grades give you structure. Market data gives you dollars. Connecting them is a regression.

Take your benchmark jobs — the subset with credible survey matches, typically 40–60% of your job catalog. For each one, you need two numbers: its evaluation point score and its market rate at your chosen percentile. Public data from the Bureau of Labor Statistics OEWS program works for a sanity check, but for structure-building you want an industry survey with real job matching. Our guide to salary benchmarking covers how to age and blend multiple sources.

Then run a simple linear regression with points on the X axis and market pay on the Y axis. The output is your market pay line: an equation that predicts a market rate for any point score, including for jobs that have no survey match at all. That last part is the whole point — it is how you price the 55% of your jobs that no survey covers.

Read the R² before you read anything else. The SHRM teaching case on pay structure design treats roughly 0.95 as the working benchmark for variance explained. If your R² comes back at 0.71, do not adjust the regression — go back to the evaluation. A weak fit almost always means specific jobs are scored inconsistently with how the market values them, and those outliers are visible on the scatterplot. Circle the points sitting far off the line and re-examine those jobs. Usually you find a handful with inflated responsibility ratings or an obsolete job description.

If pay curves upward at senior levels — and it typically does — a polynomial fit will beat a straight line. Compare the R² of both and use whichever explains more, but keep the model simple enough that you can explain it to an executive in one sentence.

Building this in a spreadsheet works, but it breaks the moment you re-evaluate 30 jobs and every downstream number has to be recalculated by hand. PointFactors keeps the scoring, the market line, and the grade assignments connected, so a re-score updates the structure automatically.

Step 4: Shift the market line to a pay policy line

Your market pay line describes what the market pays. Your pay policy line describes what you intend to pay, and the difference between them is your compensation philosophy expressed as a number.

If you have decided to lead the market by 3%, multiply the market line's predicted values by 1.03. If you target the 50th percentile flat, the two lines are identical. Some organizations lag on base and lead on incentives; in that case the policy line for base pay sits below market by design, and you should document the reasoning right there in the structure file. Auditors and skeptical VPs both ask.

Roughly nine in ten organizations tie midpoints to the 50th percentile, so leading or lagging is a real choice you should be able to defend — not a default.

Step 5: Set the midpoint for each grade

Now the conversion actually happens. For each grade, take the midpoint of its point range and plug it into your pay policy line equation.

A worked example. Suppose your policy line is:

Predicted base pay = $312 × (points) + $19,400

Grade 5 spans 253–330 points, so its point midpoint is roughly 292. That gives you $312 × 292 + $19,400 = $110,504, which you round to a clean $110,500 midpoint.

Repeat for every grade, then check the midpoint progression — the percentage jump from one grade's midpoint to the next. WorldatWork's salary structure research has consistently found progressions clustering in the 10–14% range, with about a fifth of organizations at 15–19%. Progressions that widen as grades ascend are normal and healthy: 9% between grades 2 and 3, 16% between grades 11 and 12, because senior labor markets are simply steeper.

If two adjacent grades come out 4% apart, you have one grade too many — merge them. If a jump lands at 28%, you are missing a level and people will get stuck.

Step 6: Build the range around each midpoint

Range spread is the distance from minimum to maximum, expressed as a percentage of the minimum. WorldatWork found average spreads for market-based structures running from about 47% at the narrow end to 58% at the wide end.

Working from a $110,500 midpoint with a 50% spread:

• Minimum = midpoint ÷ 1.25 = $88,400
• Maximum = minimum × 1.50 = $132,600

Let adjacent grades overlap by roughly 20–40%. Overlap is a feature, not sloppiness — it lets a strong performer at the top of grade 5 out-earn a new hire in grade 6, which is exactly what you want. Zero overlap forces a promotion every time someone becomes genuinely more valuable.

Once ranges exist, compa-ratio becomes your ongoing management tool, and our salary structure guide covers maintenance cadence and annual aging.

Step 7: Test the structure before you announce it

Model every current employee against the new structure and count three things:

1. Below minimum (green-circled). Budget the cost of bringing them to minimum. This is usually non-negotiable in a pay equity context.
1. Above maximum (red-circled). Decide the policy now — freeze, lump-sum in lieu of increase, or grandfather — not later, under pressure.
1. Grade changes. Anyone whose new grade differs from their old one needs a manager conversation scripted in advance.

Then run the demographic cut. If green-circled employees skew toward one gender or race group, you have found a pre-existing pay problem, and implementing the structure is your chance to fix it cleanly rather than inherit it.

If the total remediation cost exceeds about 1% of payroll, phase implementation over two cycles. Above 3%, revisit whether your policy line is positioned realistically for your budget.

FAQ

How many benchmark jobs do I need for a reliable regression? Aim for at least 20–25 well-matched benchmarks spread across the full point range. The spread matters more than the count — 30 benchmarks all clustered between 200 and 350 points will not predict pay for a 700-point executive role.

What if a job's market rate is far above what its points predict? That is a hot-skill or scarcity premium, and it is legitimate. Handle it with a documented market premium or a separate pay range for that job family, not by inflating the job's evaluation score. Inflating the score corrupts your internal equity for every job scored against the same factors.

Should grade point ranges be equal width? Not necessarily. Equal-width ranges are simpler to explain, but real distributions rarely cooperate. Wider ranges at senior levels usually reflect reality, since executive scores spread out further. Follow the gaps in your data first, then smooth for consistency.

Can I skip job evaluation and build grades straight from market data? You can, and plenty of organizations do — but you will have no defensible way to price jobs the surveys do not cover, and no internal logic to point to when someone challenges a pay decision. See market pricing vs job evaluation for how to combine both.

How often should I rebuild the structure? Age the ranges annually against market movement. Rebuild the underlying grade boundaries every three to five years, or whenever a reorganization materially changes what jobs exist.

Do points-to-grades conversions hold up under pay equity review? They hold up well, provided the factor plan is gender-neutral and applied consistently. A documented chain — factor definitions, scores, grade boundaries, market line, ranges — is precisely the evidence regulators look for.

Get the structure built without the spreadsheet sprawl

The conversion from points to grades is not conceptually hard. It gets hard when the inputs change — a reorganization, a new job family, a re-scored role — and every downstream number has to be rebuilt by hand from a workbook only one person understands.

PointFactors runs the whole chain in one place: score jobs against weighted compensable factors, fit the market line, set grade boundaries, and regenerate ranges when inputs move. Book a demo and bring your current point scores — we will show you the grade structure they imply in the same session. Or check pricing if you would rather start on your own.

Justin Hampton is founder and CEO of PointFactors.