Job Leveling Framework: How to Build One for a Marketing Team

A job leveling framework is a fixed set of levels plus written criteria for placing every role at one of them, so that pay attaches to the level rather than to a negotiation. For a marketing team, the framework that works is seven levels (coordinator, specialist, manager, senior manager, director, VP, CMO), eight job families, four criteria (scope of impact, autonomy, decision authority and leadership expectation), and a parallel individual contributor track from level three upward. Build it in seven steps, level the job rather than the person in it, and expect the first pass to take two to four weeks for a team of twenty.

Last updated August 2026. All figures are US dollars and reflect the most recent published data available at the time of writing.

What is job leveling?

Job leveling is the practice of defining a small number of levels for your organization and then placing each role at one of them against written criteria. It replaces the default process, which is that a hiring manager picks a title, someone in finance argues about the number, and the level is whatever the two of them settle on. That default produces two people doing identical work $30,000 apart, and it produces the marketing manager who has nothing to manage.

The distinction that matters: a level is not a title and it is not years of experience. A level is a statement about scope. Two people can both be called marketing manager and be two levels apart in what they actually control, which is why title-based pay decisions drift. Once levels exist, the title becomes a label you apply after the decision instead of the thing driving it.

Marketing needs this more than most functions, for a specific reason. Marketing job titles are the least standardized in the company. Growth marketing manager, demand generation manager and performance marketing manager are often the same job at the same level with three different names, while head of marketing means a director-level job at a 40-person company and a VP-level job at a 400-person one. A framework is what stops that ambiguity turning into a pay problem.

Why bother? What a framework actually fixes

Three things, in order of how much money they represent.

First, defensible ranges. Roughly 53.6% of US job postings now disclose a pay range, up five percentage points year over year, and a range is legally required in the posting itself in California, Colorado, Hawaii, Illinois, Maryland, Massachusetts, Minnesota, New Jersey, New York, Rhode Island, Vermont, Washington and Washington D.C. If your levels are not defined, every published range is improvised, and improvised ranges are the raw material of pay equity claims. Levels give you an answer to "why is this the range" that is the same answer for everyone.

Second, hiring accuracy. A req written from a level carries the three lines marketers actually screen on: the budget, the headcount and the metric. A req written from a title carries a list of twenty duties. The first one attracts the right people and the second one attracts everyone, which is a real cost now that US postings average 95 applicants each, roughly double the 46 they averaged in 2021.

Third, retention. People leave when they cannot see how to advance, and "work hard and we will see" is not a path. Written criteria per level turn a promotion conversation from a negotiation into an assessment, which is better for both sides even when the answer is no.

How to create job levels: the seven steps

1. Decide how many levels, and resist adding more

Seven is right for most marketing teams, and it is what the published US career progression actually looks like: coordinator, specialist, manager, senior manager, director, VP of marketing, CMO. Teams under about fifteen people can run five by collapsing senior manager and VP. What you should not do is invent eleven levels so that everyone can be promoted every eighteen months. That is title inflation with extra paperwork, and it destroys the comparability that made the framework worth building.

2. Define the job families before the levels

A family is a group of roles that share a skill set and a career path. For marketing, eight families cover almost everything: demand generation and growth, product marketing, content and SEO, brand and creative, marketing operations and technology, lifecycle and CRM, communications and PR, and marketing analytics. Families matter because they let you say that a level-three role is a level-three role whether it is a brand manager or a marketing operations manager, which is the whole point. The full grid, with 30 real US marketing titles resolved to a level code and a family, is the marketing job levels and job title hierarchy mapping.

3. Write the criteria, and make them observable

Four criteria hold up in practice, and US compensation guidance from Pave, AIHR and Ravio converges on roughly this set:

Notice what is missing: years of experience. Years are a proxy that survives because it is easy to screen on, and it filters out the people who compressed the same learning into less time. Keep years in the posting as a sanity check if you like, but never as the criterion that sets the level.

4. Level the job, not the person sitting in it

This is the step where frameworks quietly fail. You will get to a role and think, well, Sam is excellent, so this should be a senior manager slot. Write the level for the job as it exists on the org chart. If Sam is operating above the level of the job, that is a promotion conversation or a scope conversation, and it is a separate one. Mixing the two produces a framework where every level definition is a description of one specific employee, which means it cannot be used for hiring at all.

5. Attach bands from published market data, not from your current payroll

Benchmark each level against published US averages, then set a band inside the cluster. Anchoring on what you currently pay just encodes your existing inconsistencies into the framework. For marketing the practical 2026 bands run $45,000 to $65,000 at coordinator, $60,000 to $85,000 at specialist, $75,000 to $125,000 at manager, $110,000 to $150,000 at senior manager, $110,000 to $185,000 at director, $175,000 to $260,000 at VP and $180,000 to $350,000 base at CMO. Every published average behind those is named in the 2026 marketing salary guide, including where sources disagree by more than $200,000, which for a CMO they do.

6. Build the individual contributor track at the same time

From level three upward, run a parallel track: lead, staff, principal sitting at the same levels and the same bands as manager, senior manager and director. If you skip this, your best practitioners have to become managers to get paid, and you lose twice. This matters most in marketing operations and product marketing, where the deepest person on the team is frequently the wrong person to run it.

7. Test it against three real reqs before you publish it

Take three roles you are about to post and level them with the framework. If two of the three come out at a level nobody in the room agrees with, the criteria are not observable enough yet. This is a twenty-minute test that saves a rollout.

Job leveling examples: three roles levelled properly

A concrete pass, because abstract criteria are hard to trust.

"Marketing manager" running the newsletter and social, no budget, no reports. Scope: one to two channels. Autonomy: told the outcome. Authority: none, spend goes through the founder. Leadership: none. That is a level two specialist, and posting it at $75,000 to $125,000 as a manager overpays for the work while attracting people who will be disappointed by it. Level it as a marketing specialist at $60,000 to $85,000 and the applications improve.

"Growth marketer" owning a $90,000 monthly ad budget and two contractors. Scope: several channels. Autonomy: given a target, brings back the plan. Authority: real budget with a ceiling. Leadership: manages a contractor bench. That is a clean level three, growth marketing manager, $95,000 to $150,000 given the budget size.

"Head of marketing" at a 45-person company, reporting to the CEO, three reports. Scope: the whole function. Autonomy: sets strategy. Authority: owns the functional budget, hires. Leadership: manages three, none of them managers. That is a director, $110,000 to $185,000, not a VP, because nobody underneath is managing anyone yet. Titling it VP costs you the ability to hire an actual VP later without a demotion conversation. Where exactly that line falls is worked through in marketing director versus marketing manager.

What is a job leveling matrix?

A job leveling matrix is the framework as a grid: job families across the top, levels down the side, and in each cell the title and the expectations for that combination. It is the artifact people actually use, because it answers the two questions a hiring manager has at once, which level and what do we call it. A full worked example for marketing, with 30 titles mapped across eight families and seven levels, is the marketing job levels matrix. If you only need one title settled rather than the whole framework, what level a marketing manager is works through the most commonly mislevelled role on the list.

Keep the cells short. A matrix where each cell is a paragraph does not get read, and the detail belongs in the level definitions rather than in the grid. Two or three lines per cell is the working limit.

How long does job leveling take, and who should be in the room?

For a marketing team of twenty, budget two to four weeks of part-time work. The steps that take real time are writing the criteria (a day of argument, well spent) and levelling the existing roles (an hour per five roles, plus the conversations that follow). Company-wide rollouts across every function run one to two quarters, which is why starting with one function is sensible.

In the room: whoever owns marketing, whoever owns compensation, and one finance person. Not the whole team, and not a committee of nine. Published guidance consistently recommends a cross-functional group, which is right for a company-wide framework, but for a single function three people who can decide is faster and produces a cleaner document.

One thing worth planning for before you announce anything: the framework will immediately generate development conversations, because the first question everyone asks is what they need to do to reach the next level. Having something concrete to point at, whether that is a competency list per level or a way to train and certify people against the skills each level requires, is the difference between a framework that motivates and one that reads as a ceiling.

What are the criteria for job leveling?

Scope of impact, autonomy, decision authority and leadership expectation, scored the same way for every role in the company. Some frameworks add technical depth and cross-functional influence as separate dimensions, which is worth doing above level four where the individual contributor track needs its own definition. Anything beyond six criteria stops being usable in a real conversation.

The four mistakes that make a framework collapse

Too many levels, so that comparability disappears and every level boundary is arguable. Levelling people instead of jobs, so the document describes your current team rather than your org. No individual contributor track, so senior specialists have to become bad managers to get paid. And publishing the levels without publishing the bands, which is the one that generates the most cynicism, because a framework that tells people there is a level above them without telling them what it pays reads as a management tool rather than a career one.

A fifth, less obvious one: never revisiting it. Levels drift as the company grows, and the head of marketing job that was genuinely a director role at 45 people is a VP role at 200. Review the framework once a year against what you are actually posting. If your marketing team is growing fast, marketing team structure by company size covers which levels you should expect to staff at each stage, and who marketing reports to covers the reporting line that changes underneath them.

The short version

Seven levels, eight families, four criteria, one parallel individual contributor track. Level the job and not the person, benchmark against published market data rather than your own payroll, and test the framework on three live reqs before you roll it out. Publish the bands alongside the levels, because in thirteen states plus D.C. you are going to publish them in the job ad anyway. Then use it: the level writes the req, the req sets the range, and the range is the thing that decides whether the right marketers apply.