The Industry Pay Premium Score, One Occupation Priced to Aim, Add, or Skip

This guide is a scoring standard for one decision. You do the same job no matter who employs you, and you need to choose which industries to point your search at. It holds your occupation fixed, prices each industry's pay premium in dollars after benefits and bonuses, scores whether you can realistically get hired there, and ends with every candidate industry labeled Aim, Add, or Skip.

Why sector pay rankings answer the wrong question

A ranking of sectors by average pay averages every job the sector employs. It mixes up which jobs a sector has with what it pays for any one of them. To find out which industry pays the most for your job, hold your occupation fixed and compare that occupation's wage across industries.

A sector full of engineers and traders will top an average-pay list even if it pays its HR staff ordinary rates. The cell you need is narrower: one occupation in one industry. I call the result the industry pay premium, which is the difference in dollars between what your occupation earns in a given industry and what it earns across all industries nationally.

That premium can be large. For the same job, the spread between the best-paying and worst-paying industries looks like this:

• Project management specialists: oil and gas extraction pays a median of $158,930 and advertising pays $77,730. That is a gap of $81,200, or 2.04x.
• HR specialists: government (excluding state and local education and hospitals) pays $84,100 and employment services pays $60,550. That is a gap of $23,550, or 1.39x, and it only covers the top industries the Occupational Outlook Handbook lists.

The research supports taking these gaps seriously. Krueger and Summers found that differences in industry affect wages about as much as unionism, firm size, sex, or race do. Later research that tracks the same workers across firms adds a caveat: much of what looks like an industry premium comes from the industry hiring different people, not from paying the same person more. This guide treats the premium as real but partly inflated, and it builds in a check for that.

Because a premium is only useful if you can collect it, the score has five dimensions, not one:

1. Wage premium at the median and the 75th percentile
2. Benefits share, meaning how much of total compensation arrives as non-wage benefits
3. Variable-pay exposure, meaning how much of pay depends on bonuses and other supplemental pay
4. Hiring volume, meaning how many people the industry employs and is hiring in your occupation
5. Domain barrier, meaning how often employers in the industry require experience in that industry

Where the occupation-by-industry wage rows come from

The rows come from the BLS Occupational Employment and Wage Statistics (OEWS) survey. It publishes national wages for each occupation in about 410 industry classifications, using NAICS (the North American Industry Classification System) codes at the sector level, the 3- and 4-digit levels, and selected 5- and 6-digit levels. Industry rows are national only. You won't find your occupation by industry within one state.

OEWS covers about 830 occupations. Each occupation profile page lists only the industries with the highest employment and wages. For every other industry, it points you to the Create Customized Tables function, and that is where your full set of rows comes from.

Know these properties of the data before you trust any row:

• Annual and smoothed. OEWS publishes once a year for a May reference period. The estimates use a model built on 3 years of survey data, so a fast-moving industry will show up late.
• Straight-time gross pay. OEWS wages exclude premium pay, so overtime and similar premiums are not in the number.
• No self-employed workers. OEWS doesn't collect data on them. If your occupation is often freelance in an industry, that row describes only the employees.
• Large employers are oversampled. To maximize occupational coverage, OEWS selects larger employers more often than smaller ones.
• Industry codes have a vintage. Each release uses a specific NAICS edition. Confirm that the codes in your OEWS pull match the ones you use in any other source.
• Percentile depth varies. I couldn't confirm that the 75th percentile is visible for every industry cell on the summary pages. Pull it from the downloadable tables, and leave it blank if it isn't published.

Before you pull anything, map your title to one occupation code using O*NET. An HR Generalist falls under 13-1071. Project management specialists are 13-1082. Market research analysts and marketing specialists share a single combined code, so an analyst who only does research is scored against a mixed group. Write that down as a limit of your score.

The five dimensions, and what each one proves

Each candidate industry gets 0, 1, or 2 on all five dimensions, for a total out of 10. The first three price the money and the last two price your odds of getting hired. The cutoffs below are my own judgement calls, not a published standard. Adjust them if you like, but fix them before you look at any industry's numbers.

Wage premium

What it proves: the industry pays this occupation more at the middle of the distribution, and at the 75th percentile it shows whether that holds for experienced hires. When it lies: the industry hires different people, for example more senior or more credentialed ones, so the premium belongs to them and not to the seat. The check is to compare the 25th percentile too. If the 25th percentile premium is small and the 75th is large, the industry rewards a profile you may not have.

Benefits share

What it proves: a dollar of wage in this industry or ownership type comes with more or less non-wage compensation. When it lies: an industry-wide share describes an average worker, and that average can be pulled down by part-time staff. Private part-time benefits are 20.0% of compensation against 31.5% for full-time, so a part-time-heavy industry can look worse than its full-time seats really are.

Variable-pay exposure

What it proves: how much of an industry's pay arrives as nonproduction bonuses, overtime, premium pay, and shift differentials, which the employer cost survey groups together as supplemental pay. When it lies: union status and employer size shift supplemental pay independently of industry, so a high figure may reflect who is unionized rather than how the industry pays your role.

Hiring volume

What it proves: enough people in your occupation work in this industry that openings appear at a usable rate. When it lies: posting counts are better at showing the relative mix of vacancies than their absolute level. An OECD comparison of online postings against national statistics reached that conclusion. Use counts to rank industries against each other, not to predict how many jobs exist.

Domain barrier

What it proves: whether employers in the industry will hire someone from outside it. When it lies: "industry experience preferred" in a posting is often boilerplate, and a small sample can catch one employer's habit. This dimension has no public benchmark at all. I explain how to score it from live postings in the volume and barrier section below.

Verdict rules

• Aim: the adjusted median premium is positive in dollars, the total is 7 or higher, and neither hiring volume nor domain barrier scores 0.
• Add: the adjusted premium is positive but volume or barrier scores 0, or the total is 5 or 6. Also use Add when the premium is roughly zero but both volume and barrier score 2, which makes the industry a backstop.
• Skip: everything else.

Aim means a tailored resume and most of your applications go there. Add means the industry stays in the pipeline at lower intensity. Skip means no applications this cycle.

OCCUPATION: ______________   SOC CODE: ______
National median: $________   National 75th pct: $________   National 25th pct: $________
Baseline benefits multiplier (private, supplemental removed): ______
Employment floor for Aim (set before looking at pay): ______

INDUSTRY: ______________   NAICS: ______
OEWS employment: ______
Median: $________   Median premium: $________
75th pct: $________ 75th premium: $________ (blank if unpublished)
25th pct: $________ 25th premium: $________ (person-effect check)
Benefits multiplier for this ownership or industry: ______
Adjusted median premium, total compensation: $________
Supplemental-pay share, current ECEC: ____% (upside only, not in premium)
Live postings read: ____   Require industry experience: ____ of ____

SCORES  Premium _  Benefits _  Variable _  Volume _  Barrier _   TOTAL __/10
VERDICT: AIM / ADD / SKIP
PREMIUM WRITTEN DOWN: $________ per year at the median
REASON (one line): ______________________________

How to price the premium after benefits and bonuses

To price the premium, compute it in dollars at the median and the 75th percentile, convert each wage to total compensation using the benefits share with supplemental pay removed, and keep bonuses in a separate upside column. What's left is a dollar figure you can compare across industries without counting anything twice.

Step one: the raw premium

Here are the median rows for project management specialists from one OEWS release, as cited in KORE1's project manager salary guide. That is a secondary source, so confirm each row against the OEWS tables before you act on it.

The premium column is each industry's median minus the national median of $102,320. The 75th percentile by industry wasn't in what I could check, so on your own worksheet it's a separate pull, not something to infer.

The sector level hides much of this. Finance and insurance as a whole shows a project management median of $120,700, which is well below the $132,920 in securities and investments within it. Always score at the most detailed industry level that publishes data.

Step two: benefits

The Employer Costs for Employee Compensation (ECEC) series from BLS reports what share of total compensation goes to benefits. Total compensation equals the wage divided by one minus the benefits share.

Source: the BLS ECEC news release. The multiplier is 1 divided by the wage share.

To remove it, subtract the supplemental-pay share from the benefits share. Leave the wage share alone and divide again:

• Private: benefits 30.0% minus supplemental 4.1% leaves 25.9%. The wage share is 70.0%. The multiplier without supplemental pay is (70.0 + 25.9) ÷ 70.0 = 1.370.
• State and local government: benefits 38.8% minus supplemental 1.0% leaves 37.8%. The wage share is 61.2%. The multiplier without supplemental pay is (61.2 + 37.8) ÷ 61.2 = 1.618.

Here is how that changes the HR specialist example. The national median is $75,940.

• Government: the wage premium is +$8,160. In total compensation, $84,100 × 1.618 is about $136,070, against $75,940 × 1.370, about $104,040, at the private baseline. The adjusted premium is about +$32,000.
• Employment services: the wage premium is -$15,390. Adjusted, $60,550 × 1.370 is about $82,950, so the adjusted premium is about -$21,100.

Benefits made both gaps wider. That matches what Krueger and Summers found: differences in fringe benefits enlarge industry wage differentials rather than offset them. Two caveats apply. ECEC's government figure covers state and local government, so if your industry row includes other levels of government, treat the multiplier as an approximation. And when two candidate industries are both private and you only have the all-private share, the multiplier scales both premiums by the same factor without reordering them. Industry-specific shares exist in the ECEC tables, so pull them for any industry you're seriously considering.

Step three: variable pay

Supplemental pay was 4.1% of private-sector compensation ($1.87 per hour worked) in the release I checked, against 1.0% in state and local government. It varies by industry more than those averages suggest. An older release put financial activities at $3.43 per hour and leisure and hospitality at 13 cents. That release is stale, so pull the current industry breakdown (Table 4) before you score.

No published standard says how much to discount variable pay. My rule is to leave it out of the dollar premium entirely and record it as upside in its own column. A premium that disappears once bonuses are removed scores 0 on this dimension, whatever the headline wage says.

Check union status too. In one older release, union workers drew $1.86 per hour in supplemental pay against $1.20 for nonunion workers, so a high industry figure can reflect unionized hourly staff rather than how your salaried role gets paid.

The ECEC definition also changes from time to time. BLS has announced that ECEC will drop workers compensation costs from its measure. Read the technical note of whichever release you use and record which definition your shares come from.

Can you get hired there: volume and the domain barrier

A premium only matters in an industry that employs enough people in your occupation and will hire someone from outside. Score volume from employment and live postings, and score the barrier by reading at least 10 postings per candidate industry.

Volume

The highest-paying rows are usually small, because specialized industries pay for scarcity. In the project management data, employment drops off steeply as the premium rises.

At the other end of the scale, architecture and engineering firms plus specialty trade contractors together employ about 195,000 project management specialists at roughly the national median. They earn no premium, but they hire constantly. That's what a volume backstop looks like.

Profile headcounts add a second view of volume: how many people already in your role work in each industry. They show how far a title spreads across industries.

Source: Refolk's index of US professional profiles. Ratios derived.

Here's how to read this. A nonprofit HR generalist looking at finance has relatively more peers who already made that move than a nonprofit project manager does. A generalist role, one that exists in every organization, spreads across industries, so its volume score holds up better outside its home industry. A clustered role scores high in a few industries and low everywhere else.

Domain barrier

No published study reports, by industry, what share of postings require experience in that industry. Posting-data aggregators tag postings by industry, but none publishes that share, and profile headcounts show who works somewhere, not what employers require. You have to measure the barrier yourself, from a fresh sample.

Read at least 10 live postings for your occupation in each candidate industry. Code each one as Required, Preferred, or Not mentioned, based on its stated experience requirements, and quote the wording. Count only explicit requirements as Required. Phrases like "familiarity with the sector" go under Preferred.

INDUSTRY: ______________   DATE READ: ______   POSTINGS: ____ (minimum 10)

#   EMPLOYER        TITLE            INDUSTRY EXPERIENCE         EXACT WORDING
1   ____________    ____________     Required / Preferred / None  "__________________"
2   ____________    ____________     Required / Preferred / None  "__________________"
3   ____________    ____________     Required / Preferred / None  "__________________"
(continue to 10 or more)

REQUIRED: ____   PREFERRED: ____   NOT MENTIONED: ____
BARRIER SCORE: 2 if Required is under a third, 1 if a third to two thirds, 0 if over two thirds

Reading postings tells you what an employer asks for, but not how far your own history is from it. Refolk scores how well your resume fits each posting you give it, so running the same 10 postings through it shows which requirements you can show evidence for and which ones are actually blocking you.

Putting money and odds together

The scoring procedure, step by step

The full run takes an afternoon for one occupation. The first three steps take about 50 minutes, and the barrier reading takes the longest. Do the steps in order. Volume comes before the barrier because there's no point reading postings for an industry that falls under your floor.

Here is the project management example run through the procedure. Corporate HQ, with +$21,860 at the median and 50,570 employed, is the likely Aim if its barrier tally comes back clean. Software publishers and securities and investments have bigger premiums and smaller pools, so the barrier read decides between Aim and Add for each. Oil and gas extraction has the largest premium of all, but at 780 employed it falls under any reasonable floor, so it's an Add at best. Advertising is a Skip on premium alone.

How the industry premium score goes wrong

Most bad verdicts come from seven mistakes. Each one produces a number that looks right. Check every Aim against this list before you commit applications to it.

1. A high-premium cell in a tiny industry. The top-paying row for project management specialists employs 780 people nationwide. A +$56,610 premium spread across that few employers might produce a handful of openings a year. Check: require your written employment floor before any industry can be labeled Aim.

2. Ranking by mean instead of median. Profile pages list annual mean wages, and a few very high earners can lift one industry's mean above another's while the typical hire earns the same. Check: every premium on the worksheet comes from a median or a percentile from the customized tables.

3. Counting bonuses twice. ECEC benefits already include supplemental pay. If you gross up with the standard multiplier and then add a bonus estimate, an industry with heavy bonuses gets credited twice. Check: subtract supplemental pay from the benefits share before grossing up, and keep bonuses in the upside column.

4. Mistaking who gets hired for what gets paid. Research that follows the same workers across firms finds that industry and firm-size premiums are largely explained by person effects, meaning differences in the workers each industry hires. An industry can show a big median premium because it hires only senior people. Check: compare the 25th percentile across industries, not just the median. If the premium shrinks toward the bottom of the distribution, the seat you'd realistically get pays less than the headline. The ECEC breakdowns by establishment size (Table 6) and union status (Table 5), plus the OEWS ownership tables, help separate the layers.

5. Industry label noise in profile headcounts. Self-reported industry labels are messy. In Refolk's index, the Financial Services sample included a travel agency and the Non-profit sample included Capital One. Of 25 sampled Financial Services project manager titles, only 13 were exactly "Project Manager". Check: read a sample of profiles before you trust any headcount, and use headcounts to rank industries, never to size them.

6. Treating "nonprofit" as one bucket. Many nonprofits classify themselves under education, healthcare, or social services, and OEWS has no nonprofit industry. A search on a single nonprofit label undercounts the pool and misprices the starting point. Check: look up the industry your current employer actually sits in, and query several labels when counting volume.

7. Using stale numbers. The industry breakdowns of supplemental pay quoted in this guide come from releases several years old, and OEWS itself smooths over 3 years of data. Check: pull Table 4 from the current ECEC release and the current OEWS tables, and write the release on the worksheet.

An industry premium you cannot get hired into is a number on a worksheet, not a target.

Two weaker failures are worth naming. A sample of 10 postings read on one day can catch one large employer's template, so if half your sample comes from the same company, read 10 more. And a combined occupation code, such as market research analysts and marketing specialists, can hide a premium that belongs to only one half of the code. Note it on the worksheet rather than correcting it by guesswork.

Acting on the labels and keeping them current

Once every industry has a label, put your effort where the labels point this week. Aim industries get a tailored resume and most of your applications. Add industries get a smaller, steady share. Skip industries get nothing this cycle. Re-score whenever a new OEWS or ECEC release comes out, or when your barrier tallies go stale.

Here's a practical split. Choose two or three Aim industries and build one resume version for each, in that industry's vocabulary: the project types, regulations, and stakeholders its postings mention. The barrier tally already tells you which requirements recur. Moving a resume across industries is mostly rewording proven work into the target industry's terms, posting by posting. Refolk does that rewrite from your own history for each posting you apply to and drafts the matching cover letter, so the per-posting cost of an Aim industry stays low.

Re-check on a schedule, not on a hunch:

• OEWS: once a year. Check the OEWS page for the next release date and re-pull every industry row when it lands.
• ECEC: every release. Re-pull the benefits and supplemental-pay shares, and confirm which cost definition the release uses.
• Barrier tallies: whenever your search runs longer than expected or postings shift noticeably. Requirements move faster than wages.
• Your own data: if an Aim industry produces no first-round screens after a fair batch of applications, re-read its barrier tally before blaming the resume.