What an Average Salary Actually Tells You

“The average salary for this role is…” is one of the least useful sentences in careers advice. Not because it is false, but because an average is a summary of other people, and you are making a decision about yourself.

Mean and median are not the same claim

Both get called the average, and the gap between them is information.

The mean adds every salary and divides by the number of people. A small number of very high earners pulls it up.

The median is the midpoint: half the people in the occupation earn more, half less. It describes a typical person.

When the mean sits well above the median, the field has a long upper tail, meaning a minority earn a great deal more than the rest. That tells you something real: the occupation rewards specialisation, location, ownership or seniority very unevenly. When the two are close, pay is predictable and the ceiling is nearer.

So the first thing to do with any average is find out which one it is, and find the other.

What the single number hides

Four dimensions collapse into one figure, and each of them moves pay more than the occupation itself sometimes does.

Experience. An occupational average pools graduates with people who have done the job for twenty years. Nobody is paid the average, they are paid their point on a curve.

Location. The same occupation pays materially differently between metro areas, and the differences do not always track the cost of living in the direction you expect.

Industry. The same job title in two industries can sit in different pay bands entirely. An accountant in a small nonprofit and an accountant at an investment firm share an occupation and not a market.

Employer size and sector. Public, private, nonprofit, large, small. Each has its own norm, and each trades pay against something else.

An average is the shadow of all four. Reading it as your number is like reading a national weather average before deciding what to wear.

Look at the distribution instead

The BLS Occupational Employment and Wage Statistics programme publishes percentile estimates by occupation, industry and metro area. That is the same data most salary articles summarise, so going direct removes the summarising.

Read four numbers, not one.

Figure	What it tells you
10th percentile	Roughly the entry or low end of the field
25th percentile	A realistic first offer in many cases
Median	The typical practitioner
90th percentile	The realistic ceiling without being an outlier

Then repeat for your own metro area. The movement between the national and local figures is often larger than people expect and it is the single most decision-relevant comparison available.

We deliberately do not print figures here. Wage estimates are revised, and a number pasted into an article is wrong within a year while the source stays current.

Be careful with self-reported salary sites

They are useful for a sense of a specific employer, and they have a structural problem: the sample is whoever chose to submit.

That is not random. People submit when they are pleased with a number, when they are angry about one, or when a site nags them at a particular career stage. Job titles are also self-selected, so a self-described Senior Engineer at a ten-person startup and one at a large firm land in the same bucket.

Use them as a supplement to survey data, not as a replacement, and prefer them for questions like “what does this specific company pay” rather than “what does this occupation pay”.

Using the range in a negotiation

This is where the distinction earns its keep, because citing an average argues for the middle.

The structure that works: name the occupation’s upper range for your metro area, then give the reason you sit there rather than at the median. The reason has to be specific and about you, not about the market.

Based on the wage data for this occupation in this metro area, the upper quartile runs to around X. Given that I have run this function end to end for four years and delivered the vendor consolidation last year, I am looking at Y.

Two properties make that work. It cites a source the other person can check, and it attaches a specific piece of evidence to a specific number. An average with no evidence invites the answer that you are an average candidate.

What to do before you quote any number

Three checks, quickly.

Find the right occupation rather than your job title, since wage data is organised by standardised occupation. Read the O*NET profile for it to confirm the tasks actually match your work, because a title match with a task mismatch will be argued with. And note the release year of the figures you are citing.

Then get the CV to the point where the number is credible. If your document reads as a list of duties rather than outcomes, the upper quartile argument has nothing to stand on. See resume achievements and run it through the resume checker.

Related

For finding the top of the distribution, see highest paying jobs, and for how much geography moves it, see highest paying jobs in the USA. For pay against everything else you trade for it, see best paying jobs, and for how pay changes over a career, high income careers.

Find the median and the mean, read four percentiles, localise, then argue with evidence. You can make the resume support the number.