What should a GTM engineer do in the first 90 days?
A practical first 90-day plan for a GTM engineer at a 10-person startup, with ownership, outbound, CRM, AEO, and metrics.
Short answer. Your GTM engineer should spend the first 90 days turning founder instinct into a repeatable market loop: sharper ICP, cleaner accounts, useful outbound, tidy CRM notes, proof assets, and fast buyer calls. I would judge the role by learning speed and pipeline quality, not email volume. In a 10-person startup, this hire serves a blunt goal: more real buyer calls that teach the team what to build, sell, price, and stop saying.
In short:
- Start with buyer truth from calls, closed-lost notes, founder inboxes, product logs, and support threads.
- Ship a narrow outbound lane before broad channel work.
- Measure account fit, reply quality, meetings held, stage moves, and lessons pushed back to product.
- By day 90, the founder should have a usable GTM console and a sharper US wedge.
What should the GTM engineer own?
Own a lane. At a 10-person startup, that lane should run from buyer list to booked call to CRM note to founder debrief, with clean handoffs back to product and support. Tool admin alone wastes the hire; pure SDR work shrinks the brain you paid for. I would keep the remit close to revenue.
Your founder still owns the hardest calls during this window, especially when the product needs trust or technical depth. A GTM engineer makes those calls less random by packaging founder moves into lists, snippets, proof, and follow-up paths, much like the shift in Scaling out of founder-led sales without breaking your pipeline. The job should raise the founder's hit rate.
What happens in the first month?
Start with the market map. The engineer should build a named account map with fit reason, trigger, pain, current tool, likely owner, and warm path. Founder memory matters here. A sharp map pulls from call notes, LinkedIn trails, support mail, product logs, investor intros, and every half-finished spreadsheet sitting in the company drive.
Trust breaks quietly. Google and Yahoo sender docs name table stakes such as authentication, clean domains, low spam, clear unsubscribe paths, and steady sender behavior before a founder sees the damage in replies or calendars. Email setup belongs inside the launch plan, beside CRM fields and offer notes. That plumbing shapes every outbound test.
How should outbound look by the middle month?
Pick a wedge. The GTM engineer should test a narrow outbound lane tied to a visible trigger, a painful before-state, a named buyer, and a proof asset. Spray hurts learning. A tight lane shows whether the buyer cares, whether timing helps, whether the message earns replies, and whether the founder can turn the call into next steps.
Cold email still works when the list and reason feel earned; the breakdown in Does cold outbound still work in 2026? Benchmarks and what gets replies fits this stage better than generic sequence lore. Relevance beats volume. The engineer should read every reply, log the lesson, rewrite the lead line each week, and retire angles that draw polite confusion.
- ICP slice
- Buyer trigger
- Painful before-state
- Proof asset
- Founder-owned call path
What should the data and CRM look like?
Keep the CRM boring. A founder should see account fit, source, status, next step, buyer words, risk, owner, last touch, and next call date without asking the engineer for a tour or opening another sheet. Salesforce frames CRM around shared customer data and sales work; a tiny startup can borrow that idea without buying enterprise sprawl. Fancy fields can wait.
Tag records by buyer pain and trigger, not by whatever field the tool shipped by default. Dashboards should answer live founder questions: who fits, who replied, who booked, who stalled, what changed, and what the product team needs to hear. Clean data shortens debriefs. Messy data turns every weekly review into theater.
- Accounts added
- Fit-approved accounts
- Relevant replies
- Meetings held
- Stage moves
- Learning notes shipped
When do social selling and AEO enter?
Add public proof after the first outbound lane starts to speak. Early posts, founder comments, answer pages, event notes, and teardown snippets should echo the same buyer pain heard on calls. Random thought leadership drains time. A useful post makes the next email warmer, gives the founder a sharper point of view, arms event follow-up, and leaves a breadcrumb for search or AI answers.
AEO starts with crawlable answers. OpenAI docs name crawlers for ChatGPT and search, and a founder should assume useful product answers may get read outside the website UI during deal prep, renewal checks, procurement research, or internal sharing. A GTM engineer can turn live sales doubts into pages that answer pricing fears, use-case fit, security concerns, and competitor questions. Buyer language beats keyword stuffing.
How do we measure the first 90 days?
I would score learning first. Pipeline matters, yet a 10-person startup can fool itself with meetings that never should have happened and logos that never had pain, budget, timing, urgency, or a real internal owner. The scorecard should track fit-approved accounts, relevant replies, meetings held, deal stage moves, founder follow-ups, and product notes shipped. Vanity volume earns no trust.
By day 90, the founder should see a working market console instead of a folder of campaigns. The measures in The only GTM metrics that matter before Series B belong here: stage quality, sales cycle risk, source truth, close learning, and retention hints from early users. Retention shows up early through repeat use, deeper team pull, cleaner support paths, and fewer hand-holding calls, a point founders often delay until Retention and expansion: the GTM engine founders ignore until Series B.
- ICP rejects
- Named account map
- Working message lane
- Booked buyer calls
- Clean CRM rhythm
- Founder debrief notes
What should the founder avoid?
Avoid channel theater. Tool sprawl, broad lists, generic AI copy, outsourced SDR scripts, and dashboard polish can make a tiny team feel bigger while the buyer read stays weak. A GTM engineer should kill weak angles quickly and keep the founder close to raw calls. Pretty activity can hide drift.
Busy is cheap. A founder should reward clear buyer learning, clean handoffs, better proof, honest stop signals, and fewer loose ends before rewarding more channels or prettier charts. Hiring order also matters; Your first growth and marketing hires, in order explains why a broad growth leader can arrive too early. The GTM engineer should leave behind a sharper lane and fewer loose ends.
Common questions
Should the GTM engineer report to the founder?
Yes. Founder access keeps the market read sharp, and the GTM engineer needs call notes, product tradeoffs, pricing hunches, buyer pushback, and fast edits. Distance kills signal. A sales lead can help later, after the team has a proven wedge, clean handoff, useful proof, and a repeatable call path.
Is the role closer to sales or RevOps?
I treat it as a builder role with buyer taste. Sales gives the role its scoreboard, RevOps gives it hygiene, growth gives it tests, and product gives it context. Tool work should serve live buyer learning. If the hire only sends emails or only cleans fields, the startup loses the rare mix that makes the role useful.
How many tools should the hire own?
Few. A CRM, inbox tool, enrichment source, calendar path, and shared notes space can carry the first 90 days for most tiny teams. Tool count matters less than field discipline and weekly use. The founder should see the same truth the engineer sees, without private spreadsheets or mystery dashboards.
What if outbound gets no replies in the first 90 days?
Read the silence as data. The engineer should check fit, trigger strength, pain wording, sender health, proof, and ask size before blaming the channel. Cold outbound often fails when the reason feels generic; the zero-reply failure mode starts there. A clean no can teach more than a vague maybe.
Should this hire create content?
Yes, after buyer calls shape the point of view. Founder posts, answer pages, teardown notes, event follow-up, and short proof assets should come from real objections rather than calendar pressure. Content made too early sounds like a brochure. Sales-fed content gives outbound a warmer surface and helps future buyers self-educate.