Guide · Market pricing

Pricing a role with no market match

Blend matches openly. Never hide a weak one.

Short answer: To price a role with no market match, start from the two or three closest survey jobs, weighted by how much of the real work each covers. Say how strong the match is and give internal comparisons real weight. Using a weak match is fine. Hiding one turns a guess into a fixed number nobody questions.

Three matches, three prices

Picture a 210-person insurance software company in Atlanta. It has created a new role: A pricing and revenue analytics lead. The job is half analysis, half commercial. The person will build the pricing models, sit with the sales leaders on large deals and set up the forecasting the company now runs with AI tools. Nobody in the company does this work today. The hiring manager wants an offer out within two weeks.

The recruiter opens the salary survey and finds no job with that name. Three jobs come close: A senior data analyst, a pricing manager and a data science manager. Their medians sit far apart. In this illustration the gap from the lowest to the highest is about $35,000. The recruiter picks the data science manager, because the candidate on the shortlist is asking for that number. Finance prefers the senior data analyst, because the role has no team to manage. HR averages all three and calls it the market rate.

All three numbers come from the same survey. Each one is presented as "the market". The offer goes out at the top figure and the candidate accepts. Six months later the company's existing pricing manager, who does part of the same work, finds out what the new hire earns and asks why the gap exists. Nobody can say what the role was matched to or why.

This is a composite of a common situation at companies this size, not a client. The numbers are illustrative. The point is that pricing a role with no market match is not rare. For a growing company it is one of the most frequent pricing calls there is and the one most often made without a record.

Why pricing a role with no market match is hard

Salary surveys are built for jobs that exist in many companies in a stable form. That is what makes the data reliable: Many participants match many people to the same description. The jobs a growing company most needs to price are often the opposite. They are new, they combine two kinds of work or they exist because the company is doing something its peers are not yet doing.

The gap is getting wider. A 2025 study drawing on more than 1,000 companies expects nearly 40% of the skills required on the job to change by 2030 [5]. In a 2026 survey of more than 3,400 compensation responses, 61% of organizations said they had added AI-related skills or competencies to existing roles. In the same survey, 55% had not yet changed pay for AI skills at all [1]. The official occupation list used for US government statistics was last revised in 2018. Its next version is due to be final in 2027 and in use from 2028 [4]. Market data, however it is collected, describes the work of the recent past.

Two good goals pull against each other here. The first is speed and hiring power: The role is open, a candidate is waiting and a number that is too low loses the hire. The second is defensibility: The number has to make sense next to the people already in the company, today and once the role is posted, reviewed and compared. A fast number tends to be a negotiated number. A defensible number takes longer, because it needs a reason that outlasts the hiring week.

The deeper problem is that a weak match produces a figure that looks exactly as precise as a strong one. A survey median for a loosely matched job is printed to the dollar, just like the median for a job that matches perfectly. Nothing on the page says "this is a guess". So the uncertainty disappears the moment the number is copied into an offer or a band.

One definition helps. A match is a survey job whose description covers most of the role's real work at the right level. One public university's published pricing rules treat a survey job as a match when 70% or more of the responsibilities are the same [3]. Rules like this are conventions, not standards. They are useful because they force a question most teams skip: How much of this role does the survey job actually describe?

Five traps when no survey job fits

Most of these traps have the same root. A judgment is made under pressure and then recorded as if it were data.

"Pick the match that gives the right number"

What it looks like: Several survey jobs could fit. The one chosen is the one that produces the figure someone already wanted. A hiring manager picks the highest to close a candidate. A finance lead picks the lowest to protect the budget. Both describe their pick as the market rate.

Why smart teams fall into it: When no match is clean, every choice feels equally arguable. The person holding the most urgent goal fills the space. It rarely feels like bias, because each candidate match really does describe part of the job.

What it costs: The price now depends on who asked, not on the role. Two similar new roles priced by two different managers in the same quarter can land far apart. That is the kind of gap that is hardest to explain later, because the reason was the person making the call. It also teaches the organization that matching is a negotiation. Once managers learn that, every pricing request comes with a preferred match attached.

"Average every job that looks a bit like it"

What it looks like: To avoid picking, the team takes every survey job with some overlap and averages them. Four or five matches go into the blend, each with equal weight, including one that covers only a small slice of the work.

Why smart teams fall into it: Averaging feels neutral and cautious. It spreads the risk of being wrong. It also avoids an argument about which match is best.

What it costs: A blend is only as good as its parts. Averaging in a job that shares one task with the role pulls the price toward work the person will not do. The result can describe no real job in the market at all. A sound blend uses two or three matches that each cover a continuing, substantial part of the role, weighted by how much of the work each one covers. If a candidate match is in the blend only because it contains one shared task, it is noise, not evidence. Our guide on which salary data to trust makes a related point about blending different data sources.

"It's a hybrid, so it pays more than either"

What it looks like: The role combines two disciplines, so it is priced at the higher of the two matches plus a premium for being rare. The reasoning is that a person who can do both is harder to find than a person who does one.

Why smart teams fall into it: Sometimes it is true. Some combinations are scarce and the market pays for them. A premium also helps close candidates, which makes it popular with hiring managers.

What it costs: Many hybrid roles are not two full jobs. They are one job at a certain level with some work borrowed from a neighbor. A role that is half data analyst and half pricing manager usually needs the depth of neither in full. Pricing it above both pays for a combination the work does not require. It also creates a quiet problem inside the company. The specialist in each discipline now earns less than the hybrid that does part of their work. The more useful question is which part of the role carries the most weight and the highest level of responsibility. That part usually sets the price, with an adjustment for the rest. A premium for scarcity should rest on evidence that the combination is scarce, such as failed searches at the base price, not on the novelty of the title.

"No match, so it's whatever it takes to hire"

What it looks like: The team decides the survey cannot help, so the price becomes the first strong candidate's ask. The offer is set, the band for the role is drawn around it and the next hire in the role is priced against that first offer.

Why smart teams fall into it: It is fast. It looks like letting the market decide, since the candidate is the market in the room. When the data is weak, a real person's number feels more real than a survey's.

What it costs: One candidate is not a market. Their ask reflects their current pay, their other offers and how they negotiate. Building the role's pay on it turns one negotiation into policy, which is the same problem as drawing bands around current pay. It also weakens a legal defense the company may need later. Under US equal pay guidance, market value counts as a reason for a pay difference only when the employer can show it actually assessed the value of the job and the person's qualifications and that the gap is not based on sex [8]. "That's what the candidate asked for" is weaker than a documented assessment. And in states that require pay ranges in job ads, the new role needs a range on its first posting, so the company has to state a price before any candidate names one [10].

"Price it once and forget it was a guess"

What it looks like: The role is priced with a weak match. The number goes into a band. A year later the band is aged forward with the rest of the structure and nobody remembers that this one was a judgment call. The match is never looked at again.

Why smart teams fall into it: Once a number is in the structure, it looks like every other number. The person who made the call may have left. There is always a more urgent role to price.

What it costs: New roles are exactly the ones whose market changes fastest. As more companies create the role, surveys add it, the data thickens and the real market rate becomes visible. A band built on last year's guess can end up well above or well below that rate without anyone noticing. Thin data also has a specific risk. Survey publishers flag results that rest on a small number of organizations and suppress results below a minimum [7]. A figure from a thin cell can move sharply from one year to the next as participants come and go. A price built on it needs a review date, not just a yearly update.

What a sound price for a no-match role looks like

A sound price for a role without a clean match is not more precise than a guess. It is more honest about what kind of number it is. Several qualities set it apart.

It can be explained in two sentences. Anyone who asks, including the person in the role, can hear which survey jobs were used, roughly how much weight each carried and why. "It's the market rate" is not an explanation. "We priced it mainly as a pricing manager, with part of the weight on senior data analyst for the modeling work, because those are the two parts of the job that take most of the time" is.

It shows its confidence. The record says whether the match is strong, partial or weak. Teams that label confidence make different decisions with the same number. They avoid stretching the top of the range for a weakly matched role. They are slower to read a gap between the role and the market as an urgent problem. They also know which prices to look at first when new data arrives.

It gives internal comparisons real weight. When the market evidence is weak, the company's own structure carries more of the decision. Where does this role sit against jobs of similar size, skill and responsibility already in the company? If the new role is priced well above an existing job that a careful review would place at the same level, that is a signal to look again. EU pay transparency law makes this explicit. Pay structures must allow work of equal value to be compared using objective, gender-neutral criteria: Skills, effort, responsibility and working conditions [9]. A market match that cannot be reconciled with that comparison is weak in more than one way.

It is reached the same way twice. Two people who price the role separately, using the same rules, land close to each other. If they land far apart, the rules are not doing the work. The pricing is being done by whoever holds the pen.

It names a review point. The record says when the price will be looked at again: When the survey adds the role, when the role has been filled a few times or at a set date. A price for an emerging role can be sound without being final. One compensation software provider gives the same advice for hybrid and emerging roles: Price on core skills, scope and impact rather than titles and review the price as the role stabilizes [1].

It is checked by someone other than the requester. The person who wants the hire should not also be the person who picks the match. That is not a matter of distrust. It is the simplest way to stop the first trap.

None of this needs a large team. It needs the habit of writing down what was matched, how strongly and why, at the moment the decision is made. That habit costs minutes. Reconstructing the reasoning later, when someone asks, costs far more and usually fails.

The calls only you can make

Some of the decisions in pricing a new role are not technical. They depend on what the company believes about its own work.

Which market does this role hire from? The pricing and revenue lead in the opening could be hired from insurance companies, from software companies or from consulting. Each market pays differently. The choice says where the company expects to find the person and whom it expects to lose them to.

Is this one job or two? Some hybrid roles are real jobs that the market is starting to recognize. Others are two part-time jobs combined because the company cannot yet afford both. The second kind is often better priced as the larger half, with a clear plan to split it later.

How much weight does internal equity carry when it disagrees with the market? Some companies will pay above their structure to fill a hard role and accept the internal tension. Others hold the structure and accept a slower hire. Both are legitimate. The choice should be made once, as a policy, not role by role.

Do you pay for the skills the job needs or the skills the person brings? A candidate may bring a scarce skill the role does not use. Paying for it is a choice to invest in future work. It should be named as that, not folded silently into the role's price.

How much risk are you willing to carry on a guess? A company may decide that weakly matched roles get offers nearer the middle of the range until the match improves. Another may accept more risk to win scarce talent. That tolerance belongs to leadership, not to the recruiter on a deadline.

The data, the law and the numbers

  • Roles are changing faster than data. A study drawing on more than 1,000 companies, published in January 2025, expects job disruption equal to 22% of jobs by 2030 and nearly 40% of the skills required on the job to change. It lists technology, data and AI roles among the fastest growing [5].
  • Pay has not caught up. A compensation software provider's 2026 survey (3,413 responses, October to December 2025) found 61% of organizations had added AI-related skills to existing roles, 55% had not yet adjusted pay for AI skills and 65% were confident in their market pricing approach [1]. A separate 2026 study of job ads reports an average wage premium of 62% for jobs asking for AI skills [6]. That figure is an average across job ads, not a price for any one role.
  • Many companies lack the level structure that matching needs. A compensation data provider's 2026 survey of 525 organizations reports that 51.4% have a formal job architecture [2]. Without settled levels, a match is a guess at both the job and the level.
  • Official classification moves slowly. The US Standard Occupational Classification in use today dates from 2018. The revision now under way is planned to be final in 2027 and used from reference year 2028 [4].
  • Thin cells are flagged. A major salary budget survey marks results that may rest on fewer than 30 organizations and does not show results below five [7]. Thin data is one reason a new role can show no usable figure at all.
  • Common matching conventions. One public university's published rules treat a survey job as a match at 70% or more of shared responsibilities, ask for more than one survey source where possible and limit adjustments to a match to about 5% to 15% [3].
  • Market value as a defense is narrow. US equal pay guidance accepts market value as a factor other than sex only where the employer proves it assessed the marketplace value of the job and the person's qualifications and that the gap is not based on sex [8].
  • EU work of equal value. The pay transparency directive requires pay structures that allow equal value to be assessed on skills, effort, responsibility and working conditions. For employers that must report pay gaps, a gap of 5% or more within a category of workers that cannot be justified on objective, gender-neutral criteria can trigger a joint pay assessment [9].
  • Posting laws need a price up front. In states that require ranges in job ads, such as Washington, a covered employer's posting must show the wage scale or salary range. That includes the first posting for a new role [10].

This is general information, not legal advice. Rules and thresholds change. Check the rules where you hire.

Our view

We think a role without a clean market match should be priced from the two or three closest survey jobs, blended openly and weighted by how much of the real work each one covers. The record should say how strong the match is and when it will be checked again. When the market evidence is weak, the company's own comparisons with similar jobs should carry more weight, not less.

Using a weak match is not the mistake. Every growing company has roles the surveys have not caught up with. The mistake is hiding a weak match: Presenting a blend or a guess as "the market rate", so a judgment made in a hiring week becomes a fixed number nobody questions.

We would argue hardest against pricing a new role at whatever the first candidate asks. It is the fastest route to a band built on one negotiation. It also leaves the company without the assessment it would need if the gap were ever challenged.

Sources

  1. Payscale, 2026 Compensation Best Practices Report (3,413 responses, October to December 2025), checked 2026-10-26
  2. Salary.com, Beyond the Job Title: Pricing Hybrid and Emerging Roles in a Skills-Based World (31 July 2026; cites its 2026 State of Pay and Compensation Practices report, 525 organizations), checked 2026-10-26
  3. Oregon State University Office of Human Resources, Market pricing steps, checked 2026-10-26
  4. US Bureau of Labor Statistics, 2028 SOC revision (page last modified 30 September 2026), checked 2026-10-26
  5. World Economic Forum, Future of Jobs Report 2025, press release (7 January 2025), checked 2026-10-26
  6. PwC, 2026 Global AI Jobs Barometer, press release (15 June 2026), checked 2026-10-25
  7. WorldatWork, 2025-2026 Salary Budget Survey: Top-level results (published 15 July 2025), checked 2026-10-25
  8. US EEOC, Compliance Manual Section 10: Compensation Discrimination, 10-IV.F.2.g (market factors), checked 2026-10-26
  9. Directive (EU) 2023/970 on pay transparency, Articles 4 and 10 (EUR-Lex), checked 2026-10-25
  10. Washington L&I, Administrative Policy ES.E.2, Job Posting Requirements (issued 3 December 2025), checked 2026-10-24

Which salary data should you trust?

Salary surveys, live pay data and job ads answer different questions. Five ways teams misread their market data and what a sound source policy looks like.

Can market data set your salary bands?

Benchmark data shows what others pay. The midpoint is still your choice. Five ways teams let the data decide and what a band you can explain rests on.

Added AI skills to a job? Price them

Job ads that ask for AI skills offer 28% to 62% more. What that means for your own roles and when a premium belongs in the range.

See Resources for every guide and the self-check.