Your boss asks for "the market rate" for a job by Friday. You can't pull the answer from one table.

The market doesn't publish one fair wage for a Senior Accountant in Cleveland. You get survey data, make a series of choices about how to use it, and build a number you can defend on Monday. That process is market pricing.

The arithmetic takes little time. A spreadsheet will calculate a precise answer from a bad match, a weak data cut, or an aging rate that makes little sense. Your judgment determines whether the answer holds up.

Let's take one job from a blank page to a defensible market reference point.

Start with salary survey data

Salary survey data is a dataset of pay rates for specific jobs, gathered from many employers and published by a vendor. Mercer, Radford, Aon, McLagan, Culpepper, and ERI are among the dozens of vendors in the field.

Employers submit the data. They report what they pay employees in matched jobs, then the vendor cleans the submissions, aggregates them, and publishes percentiles. The 50th percentile, also called the median, is the value where half of the surveyed employers pay below it and half pay above it. The 25th and 75th percentiles frame the middle of the distribution.

That lineage gives survey data its authority. It also sets the limits. The results depend on the matches employers made and the honesty of their submissions. A survey gives you a useful view of the market. It doesn't settle the question for you.

Match the work, because titles lie

You have a Marketing Manager, and the survey has a Marketing Manager. The shared title proves very little.

Match the job content: the scope, the decisions, the reporting relationships, the budget, and the complexity of the work. A Marketing Manager at a 40-person SaaS startup who runs a team of one and owns a $200K budget has a different job from a Marketing Manager at a consumer brand who oversees eight people and a $6M budget. Pricing the first job from the second can push pay 40% too high. A bad match can also produce an offer that gets laughed out of the room, or leave you underpaying and wondering why everyone quits while the competitor down the road prices the job correctly.

Read the survey's benchmark job description and compare it with your internal job description. A benchmark job is one the survey vendor has defined clearly enough for many employers to match consistently. Use that description as the anchor, regardless of the title on your org chart.

A good match has 70–80%+ overlap in content and scope. At that level, you can use it. Below that, you are forcing a match.

An upward mismatch, where you select a benchmark larger than the job, inflates the market number and leads you to overpay. A downward mismatch makes it harder to keep employees. Both errors hide comfortably in a spreadsheet because the spreadsheet still gives you a number. It always does.

Age the data to today

Surveys run on their own publication schedules. A survey might have a data effective date nine months ago, so its pay rates reflect the market as it stood last September. To bring those rates forward to today, apply an aging factor.

An aging factor is an annual percentage used to account for market movement since the vendor collected the data. Say your company believes the market for the job is moving 4% per year, and the survey data is nine months old. The aging adjustment is 4% × (9 / 12) = 3%. Apply it to the survey median, and $90,000 becomes $90,000 × 1.03 = $92,700.

The multiplication takes a moment. Choosing 4% takes judgment. Set the rate too high and you overpay every job in the company. Set it too low and you drift behind the market, then lose employees in year two instead of year one.

A common approach is to choose one aging rate for the year, use it across the company, and revisit it annually. Creating a different rate for each job suggests more precision than the data can support. False precision causes plenty of trouble in compensation.

Scope the survey cut

Before you blend any numbers, decide which slice of the survey belongs in the analysis. A scope cut filters the data by industry, company size, often measured by revenue or headcount, and geography.

A Senior Software Engineer at a fintech in San Francisco should be priced from data that reflects tech, companies of a similar size, and the San Francisco metro. A national all-industries cut answers a broader question.

Narrow the cut too far and the incumbent count can fall to single digits, which makes the percentiles jump around. Leave it too broad and you average San Francisco with Tulsa, then call the result "the market." Use the widest cut that still represents the job in front of you.

Blend more than one source

One survey rarely deserves the whole decision. Pull at least two sources, and ideally three, then blend them.

To blend surveys, weight each source according to how much confidence you have in it for that job. Look at the quality of the job match, the number of incumbents in the cut, and the age of the data. More incumbents generally produce more stable percentiles.

For example, Survey A has 280 incumbents and a strong match, so you give it 50% weight. Survey B has 90 incumbents and a slightly weaker match, so it gets 30%. Survey C is current but thin, so it gets 20%. Age each survey separately, then calculate the weighted average of the three aged medians.

The result accounts for the uncertainty in each source instead of letting one survey dictate the answer.

Price one job from start to finish

Now take a Data Analyst at a mid-size fintech in Chicago. Mid-size here means roughly $400M in revenue and 1,200 employees.

The survey benchmark is "Data Analyst: Intermediate." Read the two descriptions side by side. The internal job owns data pipelines, supports three stakeholders, and calls for two years of typical experience. The benchmark owns data pipelines, supports multiple stakeholders, and expects two to four years. The content and scope overlap by 80%+, which makes this a solid match.

Next, scope the cut to the financial services industry, companies between $250M and $1B in revenue, and the Chicago metro. Three surveys have data for that cut:

The aging rate for the year is 4%. Apply it to each source:

Now assign the weights. Survey A has a big cut and a good match, so it gets 50%. Survey B is decent and gets 30%. Survey C is current but thin and gets 20%.

Weighted average = (0.50 × $89,760) + (0.30 × $93,730) + (0.20 × $87,365) = $44,880 + $28,119 + $17,473 = $90,472.

The $90,472 composite market reference point is a weighted average of three aged survey medians. It is the best estimate of the market P50, or market median, for this job's pay, aged to today and blended across sources. The calculation keeps the three distributions separate and blends their medians. It doesn't pool them into one distribution and calculate a 50th percentile from it. No single survey owns the answer.

You might use that reference point as the range midpoint, then calculate the minimum and maximum from the company's chosen range spread. Write down every assumption so the next analyst can follow the work.

Know where the judgment sits

The final number rests on four choices: the job match, the scope, the aging rate, and the survey weights. Another analyst could make a different choice at any of those points and defend it reasonably.

That makes $90,472 the center of a plausible range. The last three digits don't make the answer exact.

The strongest comp analysts can explain what would have to be true for any number they produce to be wrong by 5%. They know where the data is soft, and they can tell you how much that weakness matters. A spreadsheet is closer to a ruler with a bend in it than a perfect measuring tape. You learn where it bends and read it accordingly.

Market pricing comes down to a short sequence. Match the content and scope. Choose a cut that represents the job. Age each source forward, then blend the results with weights you can explain. The work should end with a reference point you can defend in a sentence and assumptions another analyst can retrace.

A defensible number reflects the quality of the thinking behind it.