TheJobsMarket
Negotiating a Salary Offer

How to Build a Pay Case From Published Data

An argument that survives being examined beats a confident one, and building it takes about twenty minutes.

Short answer

Find your occupation code, read the distribution for your own metropolitan area rather than nationally, establish where the offer sits in it, and name a specific number with that as the basis. A case built on a published table can be checked by the person you are asking, which is exactly why it works better than a figure from a salary site with no method.

Four numbers, in this order

The published median for your occupation in your metro. Not the national figure — geography moves pay more than almost anything else you control, and a metro figure is harder to wave away.

The full percentile spread. The tenth through ninetieth, so you can say where the offer sits rather than whether it is above or below one point.

Advertised ranges for comparable roles locally. Current, checkable, and in pay transparency states, plentiful.

What the employer itself advertises for roles at your level, which in those states is public.

Match yourself to the occupation, not the title

Wage data is published by occupational code, defined by duties. Titles are invented by employers, so looking up your title averages across genuinely different levels.

Describe your week, read the occupational definitions, and pick the code that matches what you actually do. If your role spans two, look up both — that is a stronger argument, not a complication.

Place the offer honestly

Below the twenty-fifth percentile for somebody experienced is a question worth asking about. Between the median and seventy-fifth is a normal offer for somebody established. Above the ninetieth is unusual and usually means the role is larger than the title suggests.

Being honest with yourself here matters, because an argument that misplaces you is easy to check and collapses the moment somebody does.

The sentence that does the work

“The May 2025 wage survey puts the median for this occupation in this metro at X, and the seventy-fifth at Y. Given the scope we discussed, I was expecting something closer to Y.”

Named source, stated period, specific geography, and a placement you have justified. Every part of that is checkable, which is exactly why it is difficult to dismiss.

Why data beats need, every time

An argument built on your rent, your commute or your circumstances invites a conversation about your circumstances. An argument built on the market invites a conversation about the market, where the employer’s own advertised ranges are on your side.

It also survives the person you are talking to. A recruiter can take a market argument to a hiring manager and a compensation team; they cannot take your household budget anywhere.

Cite it properly or not at all

Name the source, the reference period, the geography, and whether it is median or mean. Survey data describes a period about a year back and that is normal — state it rather than letting somebody discover it.

Do not adjust a published figure for inflation and present the result as data. Quote the published number with its date, and separately note that prices have risen. Two verifiable statements beat one invented one.

Pair the market figure with something current

Published surveys are authoritative and a year old. Advertised ranges are current and less authoritative. Together they are much stronger than either alone, because the obvious objection to each is answered by the other.

“The survey says X for this metro, and three comparable roles advertised this month range from Y to Z” is close to unanswerable without simply declining.

What to do when the data does not support you

Find out early, in private. If the offer already sits at the seventy-fifth percentile, the market argument is not available and pressing it damages your credibility for the rest of the conversation.

Then switch ground: to scope, to what the role actually requires, or to the non-cash terms — start date against vesting, retirement contribution, premium share, time off. Those are frequently more movable than base and almost nobody asks for them.

Common questions

What data should I use in a negotiation?

The published employer survey for your occupation code in your own metropolitan area. It is checkable by the person you are asking, which is what makes it work.

Why not use a salary website?

Where a site will not say what its figure measures or when, the person you are negotiating with can dismiss it in one sentence. A published table cannot be dismissed that way.

Should I name a number or a range?

A number, with the sentence that produced it. A range invites the bottom of it, and 'more' invites 'no'.

What if the data shows I am being paid well?

You have learned something valuable at no cost. An offer above the local 75th percentile is a good offer, and knowing that is worth as much as an argument.

Is personal need a valid argument?

It is true and it is not an argument about the market, so it does not move a compensation decision. Keep the case on what the work is worth.

What data should I bring to a pay conversation?

The published median and full spread for your occupation in your metro, plus advertised ranges for comparable roles — including your employer's own.

How do I cite a figure properly?

Name the source, the reference period, the geography, and whether it is median or mean. Never adjust a published figure yourself and present it as data.

What if the data does not support my case?

Find out in private first. Then switch to scope or the non-cash terms — start date against vesting, retirement contribution, premium share, time off.

CS

Charles Slocs

Data and research

Charles Slocs builds the data side of this site — pulling the federal wage and employment series, matching job titles to occupation codes, and working out what the numbers do and do not support. He writes the pages that are mostly a question about evidence: what a survey measured, how wide the spread really is, and which published figure is out of date.

All articles by Charles Slocs →