Skip to content
From the blog

Your first GTM hire: engineer or SDR?

Hire the engineer when the motion works and the repetition is what you cannot keep up with. Hire the SDR when the conversations are what you cannot keep up with. If neither is true, a hire is not the fix. What follows is the week-long count, the clock in months of runway, and the ninety days.

By Kshitij Maheshwari, co-founder · Updated August 2026 · 13 min read

The three answers

Engineer, SDR, or neither yet

Each answer has a condition attached, and the condition is about your business rather than your funding stage. The table is the short version.

The answer What it fixes What has to be true first What day 90 owes you What it costs
A GTM engineer Repetition. The same build step, by hand, every week. A motion that already gets replies, and one step you have done by hand thirty times. One step running without you, written down well enough that someone else can fix it. $127,500 median advertised, closer to $85,000 at seed.
An SDR Conversation volume. Replies and follow-ups nobody gets to. A playbook on paper someone else can run, plus your hours to coach it. Replies answered the same day, with the objection logged every time. $80,000 median on-target earnings, plus your coaching hours.
Neither, yet Nothing. Neither hire creates demand. Nothing. If no queue is overflowing, the constraint is the offer. A dated re-count, and the number that would flip the answer. Nothing in cash. It costs you a date in the calendar.

Pay figures from The Bridge Group's 2025 sales development report (February 2025, mostly North American companies) and Bloomberry's analysis of 1,000 GTM engineering job postings (updated January 2026).

The seed base is from the 2026 State of GTM Engineering survey (March 2026). Nobody publishes a ramp benchmark for the engineer seat, so day 90 there is an artifact rather than a rate.


The framework

Count both queues for a week before you read a job description

Funding stage is not a diagnosis. Keep two lists for five working days, and the one you cannot clear names the hire.

Definition

The build queue is work that repeats: rows to research, lists to rebuild, enrichment to patch. The reply queue is work that answers back: replies owed, follow-ups due, no-shows to rebook. An SDR works the second one.

The build queue
  • Rows to research, lists to rebuild by hand
  • The same manual step, again, on Tuesday
  • Enrichment to patch, exports to reconcile
  • Full means the work is repetition
The reply queue
  • Replies owed and follow-ups already late
  • Meetings to chase, no-shows to rebook
  • Conversations that went cold on Thursday
  • Full means the work is conversation

If the build queue is the one you cannot clear, the work is repetition and the answer is engineering, which may not be a hire.

If the reply queue is the one, the work is conversation, and only a person absorbs it.

Two other readings matter. Nothing in either queue means the offer is the constraint, not headcount. A reply queue full of meetings that go nowhere means you need a closer, which is a different hire.

The queue you cannot clear names the hire. If neither queue is full, a hire is an expensive way to avoid fixing the offer.

Operator note
Count it, do not guess

Founders argue this one for a month and then decide it off a podcast. Five days of tally marks settles it, because the thing you cannot keep up with is rarely the thing you complain about.

KM
Kshitij Maheshwari
Co-founder, Real Good GTM

The substitution error

What each hire multiplies, and what neither one invents

Both hires are multipliers. Point either one at a motion that does not work yet and you get the same broken thing, faster and more expensively.

The engineer multiplies
  • A working motion, made cheaper to run
  • A broken motion, made to fail faster and louder
  • One manual step, turned into a system you keep
  • Cannot tell you what to say to a stranger
The SDR multiplies
  • A written playbook, run at volume you cannot hold
  • Replies, follow-ups and reschedules, absorbed
  • Objections you have not heard yet, surfaced
  • Cannot write the playbook from nothing

Both need the same thing first: a message that has already produced replies from strangers, written down. Without one page naming who you write to, why now and what you claim, neither hire has anything to pick up.

If the role itself is still fuzzy, what a GTM engineer actually does all day is the longer answer. If the page you cannot write is the claim, the fix is an offer worth replying to.


The clock

The clock nobody prices: the search plus the ramp

Every page prices the salary. What decides this at seed is how many months of runway pass before the hire produces steadily.

  1. 1

    The search: about six weeks

    The Josh Bersin Company's 2023 hiring research with AMS put the average time to appoint someone at 44 days, across all industries globally. A scarce specialist takes longer, and both of these titles are scarce.

  2. 2

    The ramp: three months, measured for one seat only

    The Bridge Group's 2025 sales development report puts average SDR ramp to full productivity at 3.0 months. Nobody publishes a ramp benchmark for the engineer seat, so treat three months there as a floor, not a measurement.

  3. 3

    The total: about four and a half months

    Six weeks to appoint, then three months to steady output. Anything that arrives sooner is luck, and luck is not something you can staff against.

The runway test

Divide four and a half months by your remaining runway. At nine months the hire eats half of it before producing steadily; at eighteen, a quarter. Past half, this is a fundraising decision, not a hiring one.

Run it on your own runway, in your own currency. The clock does not care where you hire. The salary does.


The money

The two pay numbers, and what they are not

Two real numbers, each measured a different way. Anyone who divides one by the other is selling you a side.

Six rows, sourced
SDR pay

$80,000 on target, and $55,000 of it is base

The Bridge Group's 2025 sales development report puts median on-target earnings at $80,000, split as $55,000 base and $25,000 variable. The companies in it are mostly North American, so read it as a benchmark, not your local rate.

Engineer pay

$127,500 advertised, four years asked

Bloomberry's analysis of 1,000 GTM engineering job postings, updated January 2026, puts the median advertised salary at $127,500 across postings that published a range, with just over four years of experience asked. Not a junior seat.

What seed pays

Closer to $85,000 for the same title

The 2026 State of GTM Engineering survey, 228 practitioners across more than 30 countries, has pre-seed and seed companies closer to $85,000 base against roughly $145,000 at Series B, which in practice usually means a more junior person.

Where, not which

The country moves the number more than the role

That same survey puts US median base at $135,000 against roughly $75,000 for peers outside the US. Take the method rather than the figure: get two or three real local ranges, then run the runway arithmetic on those.

The hidden manager

One rep quietly hires a manager, and it is you

The Bridge Group's report counts one first-line leader per 6.4 reps, and one per 3.6 under $5 million in revenue. At one rep there is nobody to spread, so you are the coach, and those hours have a price you can work out.

Why these do not divide

An advertised salary is not an on-target figure

One is what a job ad offers. The other is a base plus a variable half that must be earned. Medians from different surveys do not divide, which is the rule we keep in the outbound metrics guide. Read the gap instead.

The trade founders reach for, less cash for more equity, often is not honored: nearly 68% of respondents to that same survey report little or no meaningful equity. Say what it is worth and when it vests.

Operator note
Name the variable half

At seed there is no quota history, so the variable half of an offer is either unearnable or a bonus in disguise. Decide which, out loud, at offer stage. The unnamed version ends the relationship in month five.

RB
Rahul Bageria
Co-founder, Real Good GTM

The receipts

The versus is really a sequence

The two seats are usually the same career, a few years apart. That changes what a good first hire looks like.

How the choice gets framed

"Two different species. Engineers build systems, reps talk to people, so pick a side and write the ad."

  • Treats the two seats as opposite talents
  • Ignores where working engineers came from
  • Sends you hunting a unicorn or settling for a rookie
What the profiles show

"Carry a quota, get sick of manual prospecting, teach yourself the automation, build things the team adopts, get the title."

  • The most common route in starts in an SDR seat
  • Read across 100 profiles of people doing the job
  • Quota experience is the judgment half of the build
One route, two uses
1
The route

Quota first, automation second

Bloomberry read 100 profiles of working GTM engineers and found the most reliable route into the job starts in an SDR seat: carry a number, get frustrated with manual prospecting, teach yourself the automation, then get the title.

2
If you hire the rep

A curious rep may be both hires, in order

If you like a rep who automates their own work, you may be buying both seats in sequence. Put that in the offer: it is a real career story, not a perk, and that is why they stay.

3
If you hire the engineer

Prefer the one who has carried a number

Someone who has missed a quota knows what a broken list costs and what a weak claim does to a week. That judgment is the half of the job no tool supplies, and the harder half to teach.

You are not choosing between two kinds of people. You are choosing which half of one job you need first.

Want a second pair of eyes on your queue count?

Book a Fit Check

The first ninety days

What the first ninety days should actually produce

Write the scorecard before the job description, and write it in things you can look at. Outcomes belong to the motion. Artifacts belong to the hire.

An illustrative walkthrough of the method, not a specific client result. We report real numbers only when they are real.

Don't

Borrow a scorecard from a company with a sales team

Own pipeline generation. Hit 12 qualified meetings a month by day 90.

  • Asks for an outcome the motion cannot produce yet
  • Tells you nothing in week six, when it still matters
  • Punishes the person for your offer
Do

Write it as artifacts with dates on them

By day 30: one manual step runs without you, with a page saying what to check when it breaks.

  • Visible in week four, not month six
  • Fails cheaply, and the fix is usually the brief
  • Reads the same to both of you
Three checkpoints
Weeks 1 to 2

Can they say the motion back to you?

The engineer rebuilds one list by hand and writes down its filters. The rep answers live replies with you watching and logs each objection. Either way you are checking comprehension, not output.

Weeks 3 to 6

One thing runs without you

For the engineer, a step you did by hand now fires on time, and there is a page for when it breaks. For the rep, replies get answered the same day and the objection log repeats.

Weeks 7 to 13

A second person could run it

One motion is documented well enough that somebody else could pick it up. Meetings are the motion's job, not the hire's. By day 90 you want something that keeps working while the person is away.

If neither artifact exists at week six, the brief was missing, not the person. That is a cheap thing to find out in week six and an expensive one to find out in month five.

Operator note
Artifacts, not outcomes

The cheapest correction to a bad first hire is a rewritten scorecard, and it is available early. I have never seen a founder regret asking for one boring artifact in week four. Waiting until month five costs everybody.

KM
Kshitij Maheshwari
Co-founder, Real Good GTM

Our recommendation

When the answer is neither, and how to make waiting a decision

At two people the honest default is usually neither. What this page adds is the part that stops waiting from turning into drift.

Waiting, written down
Three lines in your calendar
The date
Pick the week you will re-count both queues. Put it in the calendar today, with the tally sheet attached to the invite.
The threshold
Write the number that would flip the answer: a queue you failed to clear three weeks running, or a step done by hand thirty times.
The meanwhile
Name who does the work until then, in hours a week, and name what gets dropped to pay for those hours.
Key takeaways
4 points
  • 1 Count both queues for five days before writing any job description.
  • 2 Price the hire in months of runway, not in salary.
  • 3 Neither hire invents a motion. Both multiply whatever exists.
  • 4 A decision to wait needs a date on it, or it is drift.

If the constraint is your calendar rather than your knowledge, renting execution is the third option, and when renting the motion is the right call is its own question.

Otherwise, keep running it yourself for now and re-read the count on the date you just set.


Failure modes

Five ways a first GTM hire fails

Four of these are recoverable in a week. The fifth spends money you cannot get back.

A borrowed scorecard

A scorecard lifted from a company that already has a sales team asks the engineer for meetings and the rep for systems. Both people then fail at a job you did not hire them for.

Budgeting the salary, not the clock

The offer is signed and the runway model still assumes output next month. Six weeks to appoint plus three months of ramp is the real line item, and it lands before any pipeline does.

Hiring out of a raise

New money makes a headcount plan feel obligatory. The queue count is what says a seat should exist, and a bank balance has never cleared a queue.

One job description, both jobs

An ad asking for Python and cold-call resilience gets you the overlap of two small pools, which is usually nobody. Pick the half your queue says is binding.

!
Caution

Hiring against a motion that has never produced a reply

You cannot unwind this one. Months of runway go on proving that automation makes an unproven message fail faster, and you burn somebody's quarter doing it. They get judged on a number your offer cannot produce, and they leave.

Do this instead
Get replies from real strangers by hand first, then hire to multiply the thing that worked.

FAQ

Questions founders ask

Should my first GTM hire be a GTM engineer or an SDR?
Read which queue you cannot clear for a week. If repetition is what buries you, the answer is engineering, and it may not even be a hire. If replies and follow-ups are what buries you, only a person absorbs that. If neither queue is full, the constraint is your offer, and a hire will not fix it.
What does an SDR cost?
The Bridge Group's 2025 sales development report puts median on-target earnings at $80,000, split as $55,000 base and $25,000 variable, across mostly North American companies. Then add the part nobody budgets. At one rep you are the manager, because that same report counts one first-line leader per 3.6 reps at companies under $5 million in revenue.
What does a GTM engineer cost?
Bloomberry's analysis of 1,000 GTM engineering job postings, updated January 2026, puts the median advertised salary at $127,500 across postings that published a range. Early-stage reality is lower. The 2026 State of GTM Engineering survey, 228 practitioners across more than 30 countries, has pre-seed and seed companies closer to $85,000 base, which usually buys a more junior person.
How long until a first GTM hire produces anything?
Budget about four and a half months to steady output: roughly six weeks to appoint someone, from the Josh Bersin Company's 2023 hiring research, plus three months of ramp, from The Bridge Group's 2025 report. That ramp figure is measured for reps only. Nobody publishes one for the engineer seat, so treat three months there as a floor.
Can one person do both jobs?
You can write that job description. The market answers it by asking just over four years of experience for the build half alone, and the two halves get judged on different artifacts. At seed you are better off hiring the half your queue says is binding and keeping the other half yourself for a quarter.
Do these salary numbers apply outside the US?
No, and one survey splits it. The 2026 State of GTM Engineering survey puts US median base at $135,000 against roughly $75,000 for peers outside the US. Use the published figures as a benchmark, then get two or three real local ranges and run the runway arithmetic on those.
Kshitij Maheshwari, co-founder of Real Good GTM
About the author
Kshitij Maheshwari

Co-founder of Real Good GTM. He has been the first business hire and Chief of Staff at seed-stage B2B startups, building outbound pipeline before any playbook existed. This post is the decision seen from the other chair: he was the first hire, on a motion that was still being invented.

Connect on LinkedIn

Keep going

Before you write the job ad

The role in full, the motion your first hire would inherit, and the version you run yourself.

Not sure which seat you actually need?

Book a fit check. We will look at which queue is actually full, what your motion already proves, and tell you straight if the honest answer is neither hire yet.

Book a Fit Check

No hard sell. No fake numbers. Real good work speaks for itself.