Key takeaways
- Anchor every plan on two numbers: the market rate for the role and total sales cost as a share of new revenue. Revenue Bench's working range is 15% to 25%.
- The federal wage table reports employer-reported pay received, with commission and incentive pay included. It does not report target earnings.
- By industry, the federal survey puts sales-manager medians within $30,480 of each other across six industries, while pay inside any one industry spreads at least $81,430 across its middle half.
- Use the company's offer-and-decline history and live candidate response to move from the published range to the figure for the role. A live search is the freshest read of what the role costs.
- Match the base-variable mix to what the person directly controls.
- Set quota the median performer can meaningfully reach, and pay accelerators above 100% of plan with no cap.
In this guide
- The two anchors
- Where the market number comes from
- What the federal wage survey shows for sales roles
- Sales pay by industry, from the same survey
- Decision artifact #1: Which compensation number settles which decision
- Decision artifact #2: The plan ledger, worked
- Decision artifact #3: The role-design table
- Setting quota
- Accelerators, draws, and clawbacks
- What the plan cannot do
- Methodology and sources
The two anchors
The market rate tells you what a qualified person can command for the role you have defined. Against that sits affordability: how much the selling model can carry for each dollar of new revenue. Revenue Bench's working range for total sales cost, meaning fully loaded selling expense across the team, is 15% to 25% of new revenue.
The tension between those anchors is the design job: pay enough to win the person the role needs, inside a cost the model can sustain. The mix, quota, and payout mechanics need to fit inside the overlap.
Where the market number comes from
Use three sources, moving from broad published direction toward evidence from the live role.
- Published compensation guides from recruiters and compensation-data platforms. They offer useful direction, but each reflects its publisher's client base, sample, and cutoff date. Figures vary widely between sources, and a table without those three things attached cannot price your role.
- The company's offer-and-decline history. What candidates accepted, declined, and countered over the last year is benchmark data nobody else has. It reveals where the company's role, story, and offer meet the market.
- The live market during a search. Qualified candidates in your geography and selling motion show what they will accept this quarter. This is the freshest benchmark there is, and it arrives as a byproduct of running the search itself. The company reads that evidence and sets its own number.
Offer history and live candidate response show what the market cleared. The role's range is still built from job-related criteria: scope, selling motion, and plan responsibility. Hold the same range for every candidate considered, and disclose it, at minimum where the jurisdiction requires disclosure. A candidate's current or prior compensation does not enter the calculation; salary-history inquiries are restricted in numerous US states and cities, and counsel should confirm the requirements that apply to the search.
The federal table below supplies broad occupational context. The company still builds the role's final range from its own plan and first-party evidence.
What the federal wage survey shows for sales roles
These pay figures carry a stated sample, a stated period, and a published collection method, and the agency that publishes them sells nothing in this market.
The federal wage survey, US sales occupations, May 2025
| Occupation, as the survey names it | Employed | 10th | 25th | Median | 75th | 90th |
|---|---|---|---|---|---|---|
| Sales representatives of services, except advertising, insurance, financial services, and travel | 1,256,010 | $37,980 | $47,670 | $69,990 | $100,480 | $148,840 |
| Sales representatives, wholesale and manufacturing, except technical and scientific products | 1,238,190 | $39,090 | $50,470 | $72,080 | $99,640 | $137,550 |
| Sales representatives, wholesale and manufacturing, technical and scientific products | 284,800 | $52,600 | $75,030 | $104,920 | $158,160 | $200,440 |
| Sales engineers | 51,790 | $73,870 | $92,890 | $124,900 | $161,230 | $195,270 |
| First-line supervisors of non-retail sales workers | 214,390 | $50,320 | $63,940 | $87,520 | $122,050 | $164,070 |
| Sales managers | 637,080 | $73,170 | $100,360 | $148,270 | $207,340 | $290,540 |
Source: the US Bureau of Labor Statistics Occupational Employment and Wage Statistics program, May 2025 estimates, released 15 May 2026 as USDL-26-0725. Annual wages, national, across all industries. Annual figures assume a full-time, year-round schedule of 2,080 hours.
How to read the table
Which row fits which hire. Technical and scientific products is the closest survey category to a complex-cycle B2B seller. Services covers everything from a renewal desk to a consultative sale, so the category cannot serve as a screen. Sales managers is a broad management category that runs from a car dealership floor to a software VP. Those are the limits of occupational data. The same survey's figures for one metro area appear in the Miami sales recruiting guide.
The spread is the useful part, and the median is the least useful number in the row. The distance between the 25th and the 75th percentile inside one occupation is wider than the distance between two occupations' medians. Pricing a role means choosing a position in that spread and being able to say why.
The 90th percentile is not a stretch target. It marks what the top tenth of people already in that occupation were paid. A plan that pays at the 90th only at full attainment is not paying at the 90th.
Sales pay by industry, from the same survey
The federal wage survey also publishes its sales occupations by industry, free, at the national level only. For sales managers, the May 2025 median runs from $149,980 in wholesale trade to $180,460 in computer systems design and related services, against $148,270 across all industries.
The industry cut, US sales occupations, median annual wage and middle half, May 2025
| Industry, as the survey names it | Sales managers | Sales representatives, technical and scientific products | Sales representatives of services |
|---|---|---|---|
| All industries, the national table above | Median $148,270 Middle half $100,360 to $207,340 |
Median $104,920 Middle half $75,030 to $158,160 |
Median $69,990 Middle half $47,670 to $100,480 |
| Manufacturing, sectors 31 to 33 | Median $156,150 Middle half $113,290 to $204,070 |
Median $106,650 Middle half $76,610 to $142,050 |
Median $76,290 Middle half $51,230 to $108,220 |
| Wholesale trade, sector 42 | Median $149,980 Middle half $102,080 to $206,860 |
Median $105,910 Middle half $75,790 to $160,390 |
Median $78,280 Middle half $56,410 to $115,850 |
| Finance and insurance, sector 52 | Median $174,440 Middle half $128,460 to $279,290 |
No estimate published for this industry | Median $75,770 Middle half $50,550 to $104,500 |
| Professional, scientific, and technical services, sector 54 | Median $173,700 Middle half $130,020 to $222,840 |
Median $108,100 Middle half $76,930 to $161,080 |
Median $83,880 Middle half $60,590 to $126,970 |
| Computer systems design and related services, part of sector 54 | Median $180,460 Middle half $134,650 to $232,520 |
Median $103,070 Middle half $74,990 to $152,190 |
Median $97,780 Middle half $66,930 to $140,260 |
| Software publishers | Median $168,370 Middle half $134,990 to $216,420 |
Median $120,620 Middle half $79,920 to $166,380 |
Median $98,540 Middle half $73,110 to $132,820 |
Source: the US Bureau of Labor Statistics national industry-specific occupational estimates, May 2025, released 15 May 2026 with the national estimates above. Annual wages. The middle half runs from the 25th to the 75th percentile. The agency publishes relative standard errors for the mean wage and for employment, and none for the median. In the six industry rows they run from 0.5% to 5.8% on the mean wage and from 1.2% to 9.3% on employment, against 0.4% to 0.8% and 0.7% to 1.8% in the all-industries row, because each industry sample is smaller. The industry is the industry of the employing establishment, which can differ from the product a person sells. Computer systems design sits inside professional, scientific, and technical services, so that sector row includes it. The services occupation excludes insurance and financial services sales agents, whom the survey counts as separate occupations, so the finance and insurance row does not describe them.
How to read the industry cut
Inside any one industry, the spread of pay is wider than the gap between industries. Across the six industries, sales-manager medians sit within $30,480 of each other, while the middle half of sales-manager pay inside any one of those industries is at least $81,430 wide. The same holds for both representative occupations: technical and scientific sellers' medians sit within $17,550 across the five industries with an estimate, against a middle half at least $65,440 wide, and services sellers' medians sit within $22,770, against a middle half at least $53,950 wide.
Read the row for the industry your own establishment is classified in. A software business may be classified as a software publisher or in computer systems design and related services, and the two sales-manager medians differ by $12,090. The row still carries no company stage, deal size, selling motion, or equity, and an occupation is not a job title.
No row describes a trend. The industry estimates pool the same six semiannual panels as the national table, so the agency's warning against comparison over time applies to them too.
The national limits in Methodology and sources apply to every row.
Decision artifact #1: Which compensation number settles which decision
A buyer pricing a sales role meets five kinds of compensation number. They measure different things, and using one where another belongs produces a mis-set plan.
The benchmark-translation table
| The number | What it measures | Who reports it | The decision it can settle | Where it misleads |
|---|---|---|---|---|
| The federal wage survey, the table above | Straight-time gross pay, exclusive of premium pay. The survey's technical note defines the measure this way: base rate, cost-of-living allowances, guaranteed pay, hazardous-duty pay, incentive pay including commissions and production bonuses, and tips are included. Overtime pay, severance pay, shift differentials, non-production bonuses, employer cost for supplementary benefits, and tuition reimbursements are excluded. A draw is excluded because the survey defines it as payment based on expected future wages. The figure is what people were paid at whatever attainment they reached. Base salary and on-target earnings are different measures. | Employers, through a survey drawn from state unemployment-insurance records under a published method. | Whether a proposed range sits inside, above, or below what employers across the country are paying people doing that kind of work. | An occupation is not a job title, and the figure carries no company stage, deal size, selling motion, or equity. It also cannot be read as a trend. The statistical agency says its data are not built for comparison over time because each release pools six semiannual panels across three years. |
| A published on-target-earnings benchmark | Target earnings at 100% of plan. It records a design intention. No payment is measured. | Companies selling compensation software, recruiting services, or market data. | What a competitive offer is likely to look like on paper, if the sample behind it fits the role. | Most publish no sample size, collection period, or method, so the figure cannot be checked. Target earnings also overstate what was paid wherever attainment ran below plan. |
| A self-reported earnings panel | What people who chose to submit their own numbers say they earn. | The people doing the jobs, through a platform that collects the submissions. | What candidates in a segment believe the market pays, which is the figure that appears in a negotiation. | Respondents select themselves, so the sample skews toward the engaged and the successful. Self-reported pay is usually reported at target. |
| The company's own offer-and-decline history | What this company offered, and what was accepted, countered or refused. | The company, from its own records. | Whether this company's package clears the market for this company's role, which no external table can answer. | The sample is small, it is a year or more old by the time it forms a pattern, and it records the market's answer to the role as it was written then. |
| Live candidate response during an open search | What qualified people will accept this quarter, for this role, in this geography. | The candidates, through the search. | The final number, because it is the only evidence produced by the actual role. | It arrives late, after the range has already been set, and the pool is whoever the search reached. |
The first three sources are external and get you a range. The last two are first-party and get you the figure for the role. Revenue Bench is paid to run searches, so publishing that the freshest benchmark arrives through a search aligns with the firm's commercial interest. The evidence in row 5 is the candidate's own answer. A company that reads it can set its own number without engaging anyone.
Decision artifact #2: The plan ledger, worked
The plan ledger builds the affordability anchor from the company's own numbers. Start with planned new revenue, choose a point inside the working cost range, and test whether headcount, pay, quota, margin, and sales-cycle assumptions hold together.
Worked illustration on stated assumptions
| Step | The question | The arithmetic | The result |
|---|---|---|---|
| 1 | What new revenue is planned for the year? | The company's own plan | $3,000,000 |
| 2 | What can selling cost? | $3,000,000 x 20%, the midpoint of the 15% to 25% working range | $600,000 fully loaded |
| 3 | How does the envelope split? | Three sellers at $160,000 OTE, the rest to the management share of selling cost, tools, and recruiting | $480,000 + $120,000 |
| 4 | What quota does that imply? | $3,000,000 across three sellers | $1,000,000 per seller |
| 5 | Does the plan hold together? | Quota at $1,000,000 against $160,000 OTE, a 6.25x ratio; check it against your gross margin and sales cycle | The plan fits the envelope; gross margin and cycle length decide whether it holds |
The ledger is one worked illustration on stated assumptions. Every dollar figure in it exists to show the arithmetic, and none of them is a market benchmark. Most companies also assign total quota above the revenue target so the plan absorbs misses; the illustration keeps the two equal for clarity.
Decision artifact #3: The role-design table
The base-variable mix is a design decision that follows what the person directly controls. The decision requires no dollar figures.
| Role | What the person directly controls | Where the mix leans | The failure pattern when it is set wrong |
|---|---|---|---|
| SDR / BDR | Activity volume and meeting quality, rather than deal outcomes | Base-heavy, with variable pay on qualified meetings held | Paying on booked-but-not-held meetings buys calendar spam. |
| AE, new-logo | The deal from first meeting to close | The loosest mix on the team, approaching an even split as deal control rises | A base-heavy plan pays for effort while the pipeline stalls. |
| AE / AM, retention-heavy | Shared outcomes with product, service, and success teams | Base-heavy, with variable pay on retention plus expansion | A hunter mix on a farmer role pays for churn nobody caused and rewards discounting to hold logos. |
| Front-line sales manager | Coaching, pipeline discipline, and hiring, rather than individual deals | Base-heavy, with variable pay on team attainment | A manager paid on personal deals competes with their own reps. |
| VP / Head of Sales | The system: process, hiring, and forecast | Base-heavy, with variable pay on the operating plan the board reviews | An all-commission leader optimizes this quarter and starves next year's foundation. |
For the added structure around equity, exit terms, and sponsor timing, sales-leader packages in sponsor-owned companies are covered in their own guide.
Setting quota
The common failure is a number only the star reaches. It looks ambitious in a plan and demoralizes everyone else when sustained effort has little chance of producing meaningful payout.
Revenue Bench's working test is that the median performer can meaningfully reach 70% to 80% of plan. That standard protects motivation because effort visibly pays. It also improves forecast honesty, since a plan most of the team will miss overstates the pipeline.
Accelerators, draws, and clawbacks
Above 100% of plan, the rate on the overage rises and is never capped. A cap prices the marginal effort of the exact people who can carry the number at zero and gives them a stopping point.
A draw can bridge a new hire through ramp while the person builds a working pipeline. State in writing whether the draw is recoverable, and have counsel review the wage terms of a recoverable draw before offering it.
A clawback recovers paid commission when revenue reverses. Define the trigger events and the look-back window in the plan document, then have counsel review those terms as well. Clear recovery rules connect payout to revenue quality without leaving either party to interpret the plan after a reversal.
What the plan cannot do
One of Carlos Garrido's operating philosophies reads, "Your compensation model shapes your sales behavior and outcome." The plan steers the people you have; every plan encourages some behaviors and discourages others, so the design question is which.
OMG's data shows most salespeople are intrinsically motivated, and only about a quarter are extrinsically motivated. A richer plan alone will not move most of a team, and pay cannot install drive that did not arrive with the person. Treat drive as a hiring decision.
Methodology and sources
The federal wage survey is a semiannual mail survey of non-farm establishments. Its sampling frame comes from the establishment lists that state workforce agencies maintain for unemployment-insurance purposes.
Each release combines six semiannual panels collected over three years. The May 2025 estimates use panels with reference dates of May 2025, November 2024, May 2024, November 2023, May 2023, and November 2022. Across those three years, the sample covers 1.1 million establishments, and sampled employment accounts for about 55% of total national employment for this release.
The response rate for this release was 66.2% of establishments and 67.2% of weighted sampled employment. The relative standard error on the wage estimates in the national table runs from 0.4% to 1.2%, and the relative standard error on the employment estimates runs from 0.6% to 2.8%. Annual wages assume a full-time, year-round schedule of 2,080 hours. The release's technical note carries the wage definition, sample, and response rate.
A percentile is a boundary inside the occupation. The median is the boundary between the highest-paid 50% and the lowest-paid 50% of people in that occupation, and it says nothing about a career path. The survey does not cover the self-employed, owners and partners in unincorporated firms, household workers, or unpaid family workers. Wage estimates exclude benefits entirely. Industry-level estimates are published at the national level only.
The industry cut uses the same release's national industry-specific estimates. Each industry row is a subset of the national sample, so its margin of error is wider, and the caption under that table states the range.
The working ranges on this page, the 15% to 25% envelope and the quota test, are Revenue Bench's own working guidance from sales hiring and selection work. They are not survey findings. The OMG figure is OMG's research. The Revenue Bench sources page lists every figure on the site with its producer, sample, and period.
Candidate assessment runs through Swanston Growth Advisors, a Certified Partner of Objective Management Group. Revenue Bench does not design, set, benchmark, or approve compensation; the company and its advisers own those decisions.
How much should I pay a salesperson in 2026?
The closest federal survey category for a complex-cycle B2B seller is sales representatives, wholesale and manufacturing, technical and scientific products. Its May 2025 median is $104,920, with a 25th-to-75th percentile spread of $75,030 to $158,160. The figure is pay people received, including commission and incentive pay. Target earnings measure a different thing. Anchor the plan on the market rate for the role, using sourced compensation data, the company's offer history, and live candidate response, plus affordability, using Revenue Bench's working range of 15% to 25% of new revenue as total sales cost.
What is a good base-to-variable mix for sales roles?
The mix follows what the person directly controls. New-logo roles run the loosest mix on the team; retention, management, and leadership roles carry a higher base share.
How should I set sales quota?
Revenue Bench's working test is that the median performer can meaningfully reach 70% to 80% of plan. That level keeps effort visibly connected to pay and keeps the forecast honest, because a plan most of the team will miss overstates the pipeline.
Should I cap sales commission?
No. Pay accelerators above 100% of plan. A cap prices the marginal effort of your best performers at zero and tells them when to stop selling.
Which sources should set a sales pay range?
Read them in sequence. The federal wage survey sets the outside range for the occupation. A published earnings benchmark shows what a competitive offer looks like on paper, carrying its sample and date caveats. The company's own offer-and-decline history shows whether this company's package clears the market. Live candidate response during a search gives the freshest read of what the role costs this quarter.
Where can I find sales compensation data with a published sample and method?
The federal wage survey provides it. The May 2025 estimates combine six semiannual panels collected over three years, with a sample of 1.1 million establishments covering about 55% of total national employment, and a response rate of 66.2% of establishments. The estimates are free and public.
Is sales compensation benchmark data available by industry?
Yes. The federal wage survey publishes its sales occupations by industry, free, at the national level only, so there is no industry cut for a single city. In the May 2025 estimates, the median for sales managers runs from $149,980 in wholesale trade to $180,460 in computer systems design and related services, against $148,270 across all industries. One occupation code still covers unlike businesses, and each industry's smaller sample carries a wider margin of error than the national figure. The industry cut on this page sets six industries side by side with those limits attached.
How much do sales managers earn at software companies?
In the federal wage survey's May 2025 national estimates, sales managers employed by software publishers had a median annual wage of $168,370, with the middle half between $134,990 and $216,420. Sales managers in computer systems design and related services, where some software businesses are classified, had a median of $180,460. The survey counts commissions and incentive pay and excludes equity and year-end bonuses, so the figure is neither base salary nor on-target earnings. It carries no company stage, deal size, or selling motion.