Unfairly Research · July 2026
Market signal live

Companies are racing to hire GTM engineers.

Engineering has entered revenue.

We read 201 roles to see what companies want this new category to build.

By Bob Weishar, Unfairly 29,375 postings screened 201 roles analyzed
Hiring right now
Cloudflare OpenAI Gong Clay Anthropic Snowflake Intercom Airtable Calendly Webflow Temporal Pulumi Adyen 1Password ClickHouse Buildkite Render Sanity Cloudflare OpenAI Gong Clay Anthropic Snowflake Intercom Airtable Calendly Webflow Temporal Pulumi Adyen 1Password ClickHouse Buildkite Render Sanity
The direct answer

What is a GTM engineer?

A GTM engineer brings an experimental, engineering mindset to revenue. They turn commercial hypotheses into systems, automation, data flows, and AI agents; ship them into the live revenue motion; measure what changes; and own the loop end to end.

Category decoder · 10 signals

Here’s what companies are after.

Ten signals from the hiring market, with the evidence and buyer implication behind each one.

Category decoded 0 / 10 signals
Start with category formation
  1. 01 A new revenue category is being built in real time.There are 356 open title-family roles, 51 founding the function, and 44 companies hiring more than one.Open evidence ↓ 356
  2. 02 The title misses half the function.Only 97 of 201 qualifying roles use the GTM Engineer label. Search for the work, not the newest title.Open evidence ↓ 48%
  3. 03 Engineering depth is the new table stakes.128 roles require functional or full software-engineering depth. One explicitly requires no code.Open evidence ↓ 128
  4. 04 The experiment still needs a business owner.Fifty-nine briefs name no organizational home. None of those 59 attaches a numeric business outcome.Open evidence ↓ 0 / 59
  5. 05 They build the revenue system—not just run the tools.Integrations and APIs lead at 71%, followed by CRM architecture at 65% and internal AI tools at 56%.Open evidence ↓ 71%
  6. 06 The stack is CRM + data + automation + code.Clay appears in 55% of briefs, Salesforce in 48%, and HubSpot in 43%. The mix changes with company stage.Open evidence ↓ 55%
  7. 07 AI is table stakes. Safe autonomy is scarce.Ninety-two percent mention AI, but only 19 mention evals and three require human oversight.Open evidence ↓ 3
  8. 08 Hire evidence of building—not tenure under the title.The category is roughly three years old. Seventeen current postings already require seven-plus years.Open evidence ↓ 7+ yrs
  9. 09 The market pays engineering salaries for revenue outcomes.The median range is $145K–$200K. Full-SWE depth adds $25K to the median midpoint.Open evidence ↓ +$25K
  10. 10 The missing piece is a measurable objective.Only 13 of 201 briefs attach a number to the business outcome they expect this hire to move.Open evidence ↓ 13 / 201
Takeaway 01 · Category formation

A new operating model for revenue is becoming a job category.

The first hire runs build-and-learn loops against demand. As the function matures, the same mindset expands into infrastructure, controls, and systems the whole revenue team can trust.

Founding role × established role

The first hire builds demand. The mature team builds infrastructure.

Share of founding versus established roles naming each responsibility or tool.

FoundingEstablishedDifference
Outbound 65% 37% 3.2× odds
HubSpot 57% 39% +18 pt
Clay 65% 53% +12 pt
AI agents 47% 60% −13 pt

Outbound is the strongest difference (Fisher p<0.001). HubSpot also clears the 5% threshold (p=0.033); the Clay and agent gaps are directional.

The beachhead is Clay + HubSpot + outbound. Agent infrastructure is the expansion.

unfairly research 51 founding roles vs. 150 established roles · July 2026 · unfairly.ai/research
Reporting line × numeric outcome

No organizational home, no number.

Share of postings in each reporting-line group that attach a number to the expected outcome.

Sales / revenue3 of 15
20%
Founder / executive4 of 28
14%
Operations / systems4 of 50
8%
Reporting line unstated0 of 59
0%

All 14 number-bearing postings name an organizational home. The stated-versus-unstated difference is unlikely to be sampling noise (Fisher p=0.012).

The KPI gap begins as an ownership gap.

unfairly research 201 qualifying roles · July 2026 · unfairly.ai/research
The control gap

AI fluency is common. Safe, measurable autonomy is the scarce skill.

OpenAI and Gong ask for evals, review queues, guardrails, audit trails and safe failure paths. Most hiring briefs stop at “build agents.”

92%use AI language
9%19 of 201 mention evals
1.5%3 of 201 require human oversight
22%name production reliability
unfairly research 201 qualifying roles · July 2026 · unfairly.ai/research

Cross-tabs use the 201 postings that met all three inclusion tests. Fisher’s exact test is used for small groups; p-values and the claim audit are documented in the methodology.

Background

Where the role came from, and why it is growing now.

Newly named, popularized by a vendor, and now posted by public companies. The short history behind the measured shift.

  1. 2023

    Clay names it

    The term comes out of Clay’s own go-to-market team, whose early hires solved prospects’ data problems live in under thirty minutes. The skill that mattered was not selling, it was building.

  2. 2024

    The community adopts it

    The label spreads because it describes something people were already doing without a word for it. It gains momentum well before it gains a definition.

  3. 2025

    Startups post it

    Cursor, Lovable and Webflow hire under the title. Model APIs make the research and personalisation that used to need a team affordable for one person.

  4. 2026

    It becomes a category

    Cloudflare, Snowflake, Toast and Adyen post it. Staffing firms place it as a specialty, leadership roles appear, and the live title census reaches 356 postings at 283 companies.

Origin per Clay and The GTM Engineer. Clay coined the term; it spread once the community used it.

Takeaway 02 · The title gap

The title misses half the function.

Only 97 of 201 roles that pass the functional test use a GTM Engineer label. Searching for prior title matches excludes the majority of the relevant labor market before the interview begins.

02 · The title gap

97 of 201 use the GTM Engineer label.

The other 104 roles perform the function under growth, RevOps, marketing-engineering, systems, automation and other titles.

Uses the label97
Same function, other title104
vs
Click any row to see what the title contains
GTM Engineer label
97 48%
RevOps, systems, automation
36 18%
Growth Engineer
15 7%
Marketing Engineer
12 6%
Other functional titles
41 20%

Exclusive title families. Counts sum to 201; percentages are rounded independently.

Search for the function, not prior tenure under a title that excludes half the market.
unfairly research 201 qualifying roles · July 2026 · unfairly.ai/research

The function is easier to define than the label: the person builds systems, the systems serve the employer’s own revenue motion, and the person owns the result end to end. That test finds relevant candidates across growth, operations, marketing and software backgrounds.

Takeaway 03 · The technical bar

The non-technical version of this job is gone.

Two years ago this was widely described as a Clay operator role, something a sharp ops generalist could grow into. Today the market overwhelmingly treats it as a technical function.

03 · The technical bar

128 require coding depth. One explicitly requires none.

The real hiring decision is not technical versus non-technical. It is functional coding versus production software.

01No code1 role 02Functional code76 roles 03Full software52 roles
Click any row to see what it means
Functional coding
76 38%
Full software engineering
52 26%
Light scripting
50 25%
Not stated
22 11%
No coding required
1 1%

Exclusive coding-depth categories. Counts sum to 201; rounded percentages sum to 101%.

Hire for the production depth the work requires—not a vague promise to be “technical.”
unfairly research 201 qualifying roles · July 2026 · unfairly.ai/research

Here is the part that should worry anyone planning a hire. Eighty percent of postings are mid to senior individual contributors. Fewer than one percent are junior. In a job this new, that combination is unusual: normally a young function hires cheap and trains, because the senior talent does not exist yet.

Takeaway 04 · The ownership gap

Nobody agrees who owns it.

The most common named home is operations or systems, but 29% of postings do not name a home at all. None of those 59 gives the role a numeric outcome.

04 · The ownership gap

59 roles name no organizational home.

None of those 59 gives the role a numeric outcome.

? Owner not found29% of hiring briefs leave the org chart blank.
Click any row for detail
Not stated
59 29%
Operations or systems
50 25%
Founder or executive
28 14%
Marketing
20 10%
Sales or revenue
15 7%
Other named home
12 6%
Engineering or product
11 5%
Growth
6 3%

Exclusive reporting-line groups. Named homes outside the six recurring groups are combined into “other.”

If nobody owns the hire, nobody owns the number.
unfairly research 201 qualifying roles · July 2026 · unfairly.ai/research

Every number-bearing posting names an organizational home. Sales and revenue roles are the most likely to attach a figure, followed by founder-owned and growth roles. The relationship does not prove causation, but it is too large to dismiss as random noise in this sample.

Takeaway 05 · The work

What they are actually asked to build.

Set the title aside and read the responsibilities. This is the job, ranked by how often each responsibility appears across 201 qualifying postings.

05 · What they build

The job starts with systems, not sequences.

Responsibilities named across 201 qualifying postings.

Click any row for what it means in practice
Integrations and APIs
71%
CRM architecture
65%
Internal AI agents and tools
56%
Buying signals and scoring
52%
Enrichment and data quality
51%
Outbound and sequencing
50%
Attribution and reporting
45%
Stack consolidation
32%
Experimentation
32%

Responsibilities overlap. A posting can appear in multiple rows, so these percentages are not additive.

More than half the market expects the hire to build the thing that does the work.
unfairly research 201 qualifying roles · July 2026 · unfairly.ai/research

“Replacing manual research with infrastructure that runs unattended.”

Hyperbound, GTM Engineer

Read the top three together and the job description writes itself: connect the systems, own the customer record, and build software that does the repetitive work. Two of those three are engineering tasks, which is the whole explanation for the coding bar. The job did not get more technical because employers got fussier. It got more technical because the work changed underneath the title.

Takeaway 06 · The stack

The stack they are expected to know.

The tool list is not the job, but it reveals the substrate: a CRM, enrichment and automation layer, model APIs, and enough code to make the pieces reliable.

06 · The stack

Clay leads. The CRM is the substrate.

Share of qualifying postings naming each tool, counted from the job text.

Click a tool to see who is asking for it
Clay
55%
Salesforce
48%
HubSpot
43%
n8n
28%
Zapier
27%
LLM APIs
23%
Apollo
21%
AI coding assistants
18%
Gong
17%
Outreach
15%
Snowflake
13%

Tool mentions overlap. A posting can name several tools, so these percentages are not additive.

For a first hire, Clay + HubSpot + outbound is the beachhead.
unfairly research 201 qualifying roles · July 2026 · unfairly.ai/research

Clay leads overall, but it does not identify the technical version of the role: it appears at similar rates across the extracted role types. Snowflake, production integrations and agent systems are the stronger technical signals.

For the first-function role, the practical stack is clearer. HubSpot appears in 57% of founding roles versus 39% of established roles, while outbound ownership jumps from 37% to 65%.

Takeaway 07 · The control gap

AI is table stakes. Supervised execution is scarce.

92% of qualifying postings use explicit AI language. Only 19 mention evals, and only 3 explicitly require human review, oversight or a review queue.

Gong and OpenAI describe the mature standard: guardrails, audit trails, review queues, telemetry, evals and safe failure paths. Most briefs stop at “build agents.” The scarce skill is operating autonomy safely inside a revenue-critical system.

07 · The control gap
Autonomy detectedHuman controls missing

92% mention AI. Three require human oversight.

133mentions of “agent” or “agentic”
19of 201 roles mention evals
3of 201 require human review or a review queue
AI fluency is table stakes. Safe, measurable autonomy is the scarce skill.
unfairly research 201 qualifying roles · July 2026 · unfairly.ai/research
Takeaway 08 · The experience paradox

The market is hiring seniority that cannot exist under the title.

A three-year-old category already asks for seven-plus years of experience. The useful question is what somebody has built—not how long their résumé has carried the newest label.

GTM Engineer · The experience paradox

A three-year-old title. Seventeen job posts already require 7+ years.

3 years since the GTM Engineer title emerged
vs.
7+ years required by 17 current job posts
81% Of postings that state a numeric experience bar, 111 of 137 ask for at least three years.
Hire evidence of building revenue systems—not tenure under the title.
unfairly research 137 postings state a numeric experience bar · July 2026 · unfairly.ai/research

A candidate with three years in marketing operations and a candidate with three years writing backend services can both clear the same line, even though one knows the revenue system and the other knows production software.

Hiring for years under the title shrinks an already thin pool. The useful screen is evidence of building, revenue fluency, ownership after launch, and the coding depth the actual work requires.

Category formation

A role becomes a market when companies hire in batches and recruiters specialize.

44 companies are hiring more than one

117 of the roles sit at companies opening several at once. Clay has six live. Gong and Apollo have five and four. When a company hires a second and third, the function has stopped being a role and become a team.

5 staffing firms now place this role as a specialty

Huzzle, Jobgether, Talent Pluto, Manifest OS and Flosum post multiple variants each, including listings titled GTM Engineer, Clay Specialist. When recruiters build a specialism around a title, a labour market has formed around it.

This is no longer one startup’s title. It is a labor market.
unfairly research 356 live title-family roles · July 2026 · unfairly.ai/research
Takeaway 09 · The compensation

It costs about $170,000, and the market pays a premium for building.

From the 96 postings that published a dollar range.

The compensation range

The median posting spans $145K to $200K.

The middle half runs from $130K to $230K. The top of market reaches $400K.

Click a band to see what sits inside it
$80K
$160K
$240K
$320K
$400K
Median range, $145K to $200K. Half of all published ranges sit inside this band. If you are writing a req and want to be competitive without overpaying, this is the number to beat. 96 postings contained a parseable US dollar figure; the rest disclosed nothing or used another currency.
Middle half of the market, $130K to $230K. The interquartile range. A quarter of roles pay below $130K, usually light-scripting operator roles or non-US postings. A quarter pay above $230K, almost all of them full software engineering inside an engineering org.
Top of market, to $400K. OpenAI tops the set at $230K to $385K for Product Engineer, GTM Growth Engineering. Cloudflare reaches $303K. These are not GTM salaries, they are engineering salaries attached to a revenue mandate.

The plotted axis begins at $80K. Band positions use their actual endpoints on the $80K–$400K scale.

unfairly research 96 published US-dollar ranges · July 2026 · unfairly.ai/research
The coding premium

Software-engineering depth carries a $25,000 median pay premium.

Median midpoint by the coding depth named in the job description.

Skill bonus+$25Kfor production software depth
Full software engineering
$190K
Functional coding
$165K
Light scripting
$160K

Bar lengths are indexed to the $190K full-SWE median, the highest value in this comparison.

The premium is for software that survives contact with the revenue system.
unfairly research 96 postings with a published US-dollar range · July 2026 · unfairly.ai/research

Across 96 parseable US-dollar ranges, full-SWE roles have a $190,000 median midpoint, versus $165,000 for functional coding and $160,000 for light scripting. The average is more distorted by senior and director roles, so the median is the safer comparison.

Takeaway 10 · The measurement gap

Companies know what to measure. They will not say what they expect.

177 postings name an outcome: pipeline, conversion, retention or seller productivity. Thirteen attach a figure. The rest describe the direction of travel and leave the destination blank.

The thirteen that commit are worth reading. Riot wants to take pipeline from €20M to €40M in a year. AthenaHQ wants to build the engine that reaches $10M ARR. Ashby wants to scale past $100M ARR. Level AI wants to forecast accuracy inside five points. Perplexity wants to replace two vendors with internal builds within twelve months.

Everyone else is buying a $170,000 hope.

10 · The measurement gap
13 of 201 postings attach a number to the expected outcome

Only 13 of 201 attach a number.

Hiring brief completion13 / 201

Objective found. Success condition mostly missing.

13name an outcome and a number
164name an outcome without a number
24name no outcome
Everyone else is buying a $170,000 hope.
unfairly research 201 qualifying roles · July 2026 · unfairly.ai/research
11

What to do with this.

If you are hiring one: define the function before the title. Name the executive owner, the business number, the actions the system may take without review, and the evidence that proves it worked.

If this is your first hire: the market says to start with the demand engine. Founding roles are 3.2 times more likely to own outbound and twice as likely to name HubSpot. Clay + HubSpot + outbound is the beachhead; production agents and data infrastructure come later.

If you are becoming one: AI fluency is table stakes. The scarce layer is safe execution: production integrations, evals, observability, approval policy, audit trails and ownership after launch.

If you are deciding whether to hire at all: do not buy a title. Buy a measurable unit of leverage, then decide whether it requires a person, software, or both.

12

Who is hiring their first one.

Fifty-one roles are classified as founding, first-function or greenfield hires. This field is being manually audited; the sample below contains the clearest cases.

CompanyRolePay
Confido Founding GTM Engineer $170K – $210K First hire
Clay Head of GTM Architecture $220K – $275K First hire
Assured AI Marketing Engineer $180K – $280K First hire
Harper GTM Engineer $130K – $250K First hire
Hyperbound GTM Engineer $150K – $180K First hire
Harvey Growth Engineer $136K – $204K First hire
AthenaHQ Growth Engineer $110K – $165K First hire
Assembly Growth and GTM Engineer $139K – $180K First hire
Maniac Labs GTM Engineer $120K – $140K First hire
1Password Director, GTM Engineering First hire
Hightouch Go-to-Market Engineer First hire
Linkup Founding GTM Engineer First hire
Clera Founding GTM Engineer First hire
Netic GTM Engineer First hire
OpenAI Product Engineer, GTM Growth Engineering $230K – $385K Established
Cloudflare GTM Engineer $161K – $303K Established
Fireworks AI Head of GTM Engineering and Systems $250K – $270K Established
Ashby Director, GTM Engineering and Systems $240K – $262K Established
Vanta Sr. AI GTM Engineer $122K – $183K Established
Gong GTM Engineer, GTM Automations $100K – $145K Established
Put the research to work

Design your GTM-engineering function with the data.

Request access to the private, read-only research server. Approved teams can benchmark a job description, design a 90-day brief, and query the evidence behind this report.

Request private access
What brings you here?

We review every request. Access is prioritized for teams designing the function and practitioners doing the work.

get_summary the ten headline findings
get_finding any topic in depth
get_tool_demand how often a tool is named
search_open_roles query the underlying hiring market with approved access
benchmark_job_description audit your draft against the 201-role benchmark
design_90_day_brief build a stage-specific operating plan for your company

How this was built

  1. We enumerated companies and read their entire job boards rather than searching for a job title. The title is unstable, so title-search under-counts the role and over-samples companies that adopted the vocabulary early.
  2. Boards were pulled through the public APIs of Greenhouse, Ashby, Lever, Workable, SmartRecruiters, across three independent discovery methods: a hand-built list, the full Y Combinator company directory, and 190 keyword permutations. Using one method alone would inherit its bias.
  3. Postings were classified by function, not title. A role qualifies only if the person builds systems rather than configures tools, the systems serve the employer's own revenue motion, and the person owns them end to end. 201 roles at 158 companies passed all three tests. Every classification carries a written reason so the boundary can be checked.
  4. Percentages are of the 201 postings meeting that bar. The 356 live-title figure comes from a separate title census at 283 companies; it is not the denominator for the functional sample.
  5. Cross-tabs use two-sided Fisher exact tests where groups are small. Associations are not presented as causal, and small groups retain their denominators.
  6. What this cannot show: boards on Workday and iCIMS are not covered, and postings that exist only on LinkedIn are invisible. This is a single snapshot, so it cannot measure whether any responsibility is rising. Monthly snapshots are required for a trend line.
  7. 149 roles found through keyword search had already closed by the time we fetched them, which is itself a finding about how fast these reqs turn over.
Or skip the search

Hire Elana instead.

You just read the requirements from 201 job descriptions. Elana is built to meet them. She is an AI GTM engineer who works inside your context, connects your stack, and does the work this report describes.

The job asks for
Integrations and APIs71% of postings
CRM architecture65% of postings
Internal AI agents and tools56% of postings
Buying signals and scoring52% of postings
Enrichment and data quality51% of postings
Elana does it
Connects Clay, HubSpot, Salesforce and your warehousethe stack this report ranked
Owns routing, scoring and record hygienewith approval gates on every write
Builds the internal agents rather than being one more toolreasons from a spec, not a fixed workflow
Watches signals and drafts the outreachyou approve before anything sends
Reports in pipeline, not ticketsthe number 94% of postings never wrote down
$170,000
median midpoint, plus equity, for a function whose exact title misses half the labor market
vs
$1,500
per month. Working the day you connect her, across the whole stack, with a receipt for what she did
Meet Elana  →