Pay Policy Line: How to Build One From Job Evaluation Points

Pay Policy Line: How to Build One From Job Evaluation Points

You scored your jobs. You pulled market data. Now you have two numbers for every job, and they do not agree. The pay policy line is how you reconcile them into one defensible pay structure.

A pay policy line is a straight line that links job evaluation points to pay. It shows what you intend to pay for a job of a given size. You build it by regressing market pay on points, then adjusting the result for where you want to sit against the market. Every grade midpoint on your structure comes off that line.

This guide walks through the full build with real numbers. You will see how to run the regression, how to apply your pay position, how to read the fit, and how to turn the line into grade midpoints and ranges.

TL;DR

• The pay policy line is your market pay line, adjusted for your pay philosophy. It is the bridge between job evaluation points and dollars.
• Build it in four moves: score benchmark jobs, collect market pay, regress pay on points, then multiply by your market position (for example, 1.03 to lead by 3%).
• Check the fit. SHRM's pay structure teaching case treats an R-squared of .95 or higher as the standard. A lower number means your points or your market matches need rework.
• Slice the line into grades, set midpoints at each grade's center, then add range spread and overlap.
• The line only works if the points are sound. A line through bad scores is a clean line through wrong pay.

What a pay policy line is

Two lines matter in structure design, and people mix them up.

The market pay line shows what the market pays for jobs of each size. Job evaluation points sit on the X axis. Market pay for your benchmark jobs sits on the Y axis. You fit a line through those dots.

The pay policy line is that same line after you adjust it for your pay philosophy. If you pay at the market, the two lines are identical. If you want to lead or lag, you shift the policy line up or down.

WorldatWork notes that salary structures can be expressed in terms of policy lines, and defines the related building blocks you will use below. Range spread is maximum minus minimum, divided by the minimum. Midpoint progression is the percentage gap between the midpoints of two adjacent grades, measured against the lower midpoint.

Why bother with a line instead of pricing each job on its own? Because the line smooths out noise. One job's survey data may be thin or odd. The line uses all your benchmarks together, so it keeps internal equity intact while it anchors you to the market. If you want the deeper trade-off, read market pricing vs job evaluation.

What you need before you start

• Job evaluation points for every benchmark job, scored on a consistent plan. If you have not defined your compensable factors, stop here and do that first.
• Market pay for those same benchmarks, aged to a common effective date. Use a weighted average or median of base pay from your survey sources.
• A pay position decision. Lead, match, or lag, and by how much. Write it down.
• At least 15 to 20 benchmarks that span your lowest to highest grades. Fewer than that, and one outlier can tilt the whole line.

Benchmarks should cover the full point range. If all your data clusters between 300 and 500 points, the line will be unreliable at 800.

Step 1: Plot points against market pay

Here is a small example. Seven benchmark jobs, scored by points, with market pay:

Put points on the horizontal axis and pay on the vertical axis. You will see the dots climb roughly in a line. That shape is what a good point-factor plan produces.

Step 2: Run the regression

Fit a straight line using least squares. In Excel, use SLOPE, INTERCEPT, and RSQ, or add a trendline to the chart and display the equation.

For the seven jobs above, the market pay line is:

Market pay = $22,756 + ($125 × job points), with an R-squared of 0.996.

Read the slope as a price per point. Each additional job evaluation point is worth about $125 of annual base pay in this market. The intercept is the notional pay for a job with zero points, which is not a real job. It simply anchors the line.

Test it against the data. A 340-point job predicts $22,756 + ($125 × 340) = about $65,250. The actual market pay was $64,000. That is a small miss, which is what a tight fit looks like.

Check the fit

R-squared tells you how much of the variation in market pay your points explain. SHRM's pay structure teaching case states that R-squared should be .95 or higher, and that a lower value means you should re-examine the job evaluation points.

If your R-squared falls short, look in this order:

1. Bad market matches. A job matched to the wrong survey title will sit far from the line. Fix it with better salary survey job matching.
1. Mis-scored jobs. A job with too many or too few points sits off the line in a predictable direction. Re-score it.
1. Factor weights. If whole families sit above or below the line, your factor weighting may over- or under-credit what the market pays for.
1. Hot-skill jobs. Some jobs pay a premium no point plan explains. Treat them as market-priced exceptions and take them out of the regression.

Some teams use a curve instead of a straight line when pay rises faster at senior levels. A straight line is the right default. Use an exponential fit only when the data clearly bends and you can explain why.

Step 3: Apply your pay position

Now convert the market line into the policy line. Multiply each predicted rate by your market position factor.

Say your philosophy is to lead the market by 3%. The SHRM case does exactly this, noting that each predicted rate is multiplied by 1.03 when the firm leads the market by 3 percent. In our example:

Policy pay = 1.03 × ($22,756 + $125 × points)

If you lag the market, use a factor below 1.00. If you lead for specific families, build separate lines for those families rather than inflating the whole structure. Your compensation philosophy should say which choice you made and why.

Step 4: Turn the line into grades

The line is continuous. Your structure is not. You need to slice the points into grades and give each grade one midpoint.

1. Choose grade point bands. Equal-width bands (for example, 100 points) are the simplest. Our guide on converting job evaluation points into pay grades covers band design and cut-score rules.
1. Read the midpoint off the line. Use the center of each band. A grade covering 450 to 549 points centers at 500 points, so its midpoint is about $87,800.
1. Set the range spread. Spread of 40% around that midpoint gives a minimum of about $73,170 and a maximum of about $102,430. Check: $73,170 × 1.40 is about $102,440.
1. Check midpoint progression. The next grade down (350 to 449 points, centered at 400) has a policy midpoint of about $74,940. Progression is $87,814 ÷ $74,939 − 1, or about 17%.
1. Check overlap. With a 40% spread and 17% progression, adjacent ranges overlap by roughly 57%. That is high for a professional structure. Cut the spread or widen the bands if you want less overlap. Pay grade overlap explains the formula and the trade-offs.

Notice how the grade design choices interact. A 100-point band on a $125 slope gives you a 17% progression. A 150-point band would give roughly 25%. Wider progression means fewer grades and less overlap. Tighter progression means more grades and smoother promotions.

For the next layer, see how to build a salary structure from scratch and how to set salary range midpoints.

Where PointFactors fits

The line is only as good as the points behind it. PointFactors scores every job against weighted compensable factors using one consistent method, so your regression starts with clean data. See how it works in a short demo.

Common mistakes

• Regressing on too few jobs. Five benchmarks give you a line. They do not give you a reliable one. Aim for 15 or more.
• Mixing effective dates. Age all survey data to the same date before you regress. Otherwise you fit a line through data from two different markets.
• Including non-benchmark jobs. Unique jobs with no market match belong on the structure, not in the regression.
• Letting the line replace judgment. The line is a guide. A job that sits 12% off the line deserves a conversation, not an automatic correction.
• Forgetting to update. Re-run the regression every time you refresh market data. A structure that has not moved in three years has drifted from the market. Handle drift with an annual salary structure adjustment.

FAQ

What is the difference between a market pay line and a pay policy line?

The market pay line is the regression of market pay on job points. The pay policy line is that line after you adjust it for your pay philosophy, such as leading by 3%. If you pay at the market, the two are the same.

How many benchmark jobs do I need?

Use at least 15 to 20 benchmarks spread across your full range of points. A small set can work for a quick check, but the line becomes unreliable as soon as one job is mis-scored or mis-matched.

What R-squared is good enough?

SHRM's teaching case treats .95 or higher as the target. If you fall below that, re-examine your job evaluation points and market matches before you adjust the line.

Should I use a straight line or a curve?

Start with a straight line. Use a curve only when the data clearly bends and you can explain the business reason, such as steep executive pay. Most professional structures fit a line well.

Can I build the policy line without job evaluation points?

You can fit pay to job levels or grades instead, but you lose the continuous scale. Points let you place any new job on the line immediately, which makes the line a tool for pricing jobs you did not benchmark.

How often should I update the pay policy line?

Re-run it whenever you refresh market data, usually once a year. Between refreshes, hold the line steady and age pay with a structure adjustment factor.

Does the pay policy line help with pay equity?

Yes. Because every job is priced from one line using points, similar jobs land at similar midpoints. That gives you a consistent rule to defend. It does not replace a pay equity audit of what people actually earn.

Build your line on defensible points

A policy line turns your job evaluation into a pay structure managers can explain. PointFactors gives you consistent, factor-based scores to feed it. Start free with PointFactors and we will build your line from your own job data.

By Justin Hampton, founder and CEO of PointFactors.