You are looking at two requisitions with the same title. One posts $130k. The other posts $190k. Both say “Senior GTM Engineer,” both list Clay and HubSpot, and neither tells you whether the person can do anything harder than click into a table someone else built. Title inflation is now the default in this function — the role is three years old, nobody agrees on the bands, and candidates write “senior” because the market rewards it. So stop reading titles. Score the posting, and then score the candidate, against one question: what can this person actually ship without supervision?
Posted pay is a scope signal, not a seniority signal
Start with the honest baseline. Across GTM engineering postings that actually published a range, the median posted pay is $127,500 — that is the number to score shippable scope against, not a scraped aggregator estimate. And the aggregators are where most hiring managers get their anchor wrong: they show $132k–$241k while company-specific pages post lower bands because scope and location differ. A wide posted band is telling you the employer hasn’t decided what the job is, not that the job is senior.
Stage does the rest of the distortion. Early-stage companies pay $110K-$140K base with meaningful equity, while public and growth-stage firms pay $150K-$200K base with smaller grants — roughly $40k of base movement on the same band. Which means two “senior” reqs can post wildly different numbers while asking for identical work, and one “mid” req at a funded company can outbid your senior offer. If you are calibrating comp before you have calibrated capability, you are negotiating against noise. Our salary benchmarks exist for exactly this reason.
Pay bands describe the employer’s budget. Capability bands describe the hire. Only one of those predicts whether the pipeline works in ninety days.
The three bands, by what each can actually ship
Here is the split that survives contact with a real build. Read down the “Ships unsupervised” column — that is the whole test.
| Band | Ships unsupervised | Owns | Breaks down when | Comp reality |
|---|---|---|---|---|
| Junior | Wires an existing Clay table, adds enrichment columns, runs a waterfall someone else designed, QA’s a list | Execution inside a system built for them | The source schema changes, or a provider silently degrades | Below the $127,500 posted median |
| Mid | Builds the table from a written brief, stands up a new signal-to-sequence flow, writes SQL against the warehouse, debugs a sync | Delivery of a defined outcome | Asked to define the outcome, or to trade off cost vs. coverage vs. compliance | At or just above median |
| Senior | Designs the system, sets the data contract, owns inbox placement and domain policy, writes code-level automation and monitors it | End-to-end design, data contracts and SLOs — $120k–$160k total comp | Rarely — they escalate with a proposal, not a question | Top of band, or repriced by stage |
The useful detail: the market’s centre of gravity sits mid-band, not junior. Postings ask for just over four years of experience on average — 4.11 years. Employers are effectively all fishing in the same mid-band pool while writing “senior” on the door. If you want a true senior, you are competing on scope and autonomy, not on the word.
Junior: wires what someone else designed
A junior GTM engineer is genuinely valuable and genuinely bounded. Give them a working Clay table, a defined ICP, an enrichment waterfall and a sequence, and they will keep it fed, clean the output, and raise their hand when a column goes null. They know the tools. They do not know the failure modes.
The tell in an interview is how they describe a broken run. A junior says “the enrichment stopped returning emails, so I switched providers.” They fixed the symptom. They cannot tell you what the switch did to match rate, cost per verified contact, or the bounce profile you inherited two weeks later. Hire this band when there is already a senior owning the architecture — otherwise you are buying a tool operator and calling it a system.
Mid: builds from a brief, does not write the brief
Mid-band is where most of the market actually is, and where most of the value gets delivered. Give a mid a one-paragraph brief — “we want job-change signals from the warehouse routed to AEs within 24 hours, deduped against open opps” — and they ship it. They write the SQL. They handle the API. They test the routing.
What they don’t do is challenge the brief. They will build a 24-hour signal flow without asking whether the signal predicts anything, whether the volume will torch your sending domain, or who owns the field when marketing renames it next quarter. That gap is not a character flaw. It is the band. And it’s why a mid hired into a senior req looks great for sixty days and then stalls — the work that was supposed to be theirs was never defined, because defining it was the job. See the GTM engineer hiring guide for how to scope a req that doesn’t set a mid up to fail.
Senior: the data contract, deliverability, and code
A senior is identifiable by three things they refuse to let someone else own.
The data contract
They write down what each field means, where it comes from, who may change it, what happens downstream when it’s null, and what the freshness guarantee is. Then they instrument it. When a CRM field gets repurposed by an ops admin, the senior’s monitoring fires before the campaign does. This is warehouse-adjacent work and it is why strong GTM engineers increasingly look like the profiles we place through data warehouse recruiting rather than like sales ops with a Clay login.
Deliverability as engineering, not a setting
This is the single fastest way to separate bands, because juniors think it’s a toggle. It isn’t. Microsoft mailboxes accepted only 75.6% of legitimate marketing email to the inbox — meaning roughly a quarter of mail to your best enterprise accounts never lands, regardless of how good the copy is. A senior treats that as a measured system with levers: domain segmentation, volume ramps, seed testing, engagement-based suppression.
Authentication shows the same gap. 78% of domains now publish a DMARC record, but only 42% reach enforcement — a 36-point gap between looking compliant and being protected. Closing that gap is precisely the work a senior GTM engineer is hired to do, and it is invisible in every job title ever written. Ask any candidate what policy their sending domains are on today. The answer sorts the bands in one question.
Code-level automation
SQL and Python each appear in 38% of GTM engineer postings — the first hard gate between tool-wirers and builders. Not because every workflow needs a script, but because a person who can write one is not trapped inside what a vendor’s UI permits. They can reconcile two systems no integration covers, backfill a bad month of data, and monitor their own pipeline. This is the overlap with our technical recruiting work, and it’s where capability evidence beats résumé claims every time.
The three capabilities that move someone up a band
Everything above collapses into three movers. Score each candidate 0–2 and you have a band, not a guess.
| Capability | Junior looks like | Senior looks like | Evidence to demand |
|---|---|---|---|
| Code fluency | Uses formulas and no-code steps | Writes SQL against the warehouse and Python for reconciliation — the 38% gate | A query they wrote and why it was hard |
| System ownership | Executes in a built system | Sets the data contract and an SLO on freshness — the $120k–$160k senior definition | A schema doc, a monitor, an incident they caught |
| Provable AI capability | Prompts a tool | Builds evaluated, monitored AI steps into a pipeline; priced by the market at a 28% wage premium, nearly $18,000 a year | An output quality check, not a demo |
That third one matters for your budget as much as your scorecard. Proven AI capability is already priced at a 28% salary premium — roughly $18,000 more per year across job postings generally. If you want it, plan to pay above the $127,500 median and be explicit about why. If you post median comp and ask for all three movers, you will run a spray-and-pray search for six weeks and lose your finalists to a funded competitor paying $150K-$200K base.
How to score twenty postings — and then your own
Run this on the competitive set before you write your req. Pull twenty live GTM engineering postings, ignore the titles, and mark each one:
- Does it name a language? SQL or Python mentioned — builder req. Only tool names — operator req, regardless of the title.
- Does it name an owner-level artifact? Data contract, schema, SLO, monitoring, deliverability, domain policy. If yes, the scope is genuinely senior.
- Does it name years? If it says 5+ but the average ask is 4.11 years, they are padding to filter, not to specify.
- Does the range exceed $40k of spread? Undecided scope. Expect the interview loop to move the target.
- Does the range sit below $127,500 with all three movers listed? Mispriced. That req will sit open.
Then score your own posting the same way and fix the mismatch before it hits a job board. Most “we can’t find GTM engineers” problems are actually a mid-band budget attached to a senior-band artifact list — and that’s diagnosable in an afternoon.
Hire against evidence, in days not weeks
The practical shift is small. Replace “5+ years as a GTM Engineer” with three asks: show me a query you wrote and the decision it changed; show me a field definition you own and how you know when it breaks; tell me what DMARC policy your sending domains are on and what your inbox placement looked like last quarter. Candidates who have shipped senior work answer all three in ten minutes. Candidates who have wired someone else’s table answer the first one halfway.
That is also how a search gets fast. When the scorecard is capability evidence rather than title matching, you can shortlist in days, not weeks, because you are no longer interviewing to discover what the job is. If you want the band definitions applied to a live req, GTM engineer recruiting is where we do it — and candidates reading this should look at GTM engineering roles instead. For adjacent scoping on the ops side, the marketing operations recruiting page covers where the two functions overlap.
