GTM Engineering

The First 90 Days of a GTM Engineer

A month by month plan for the company hiring one: what should exist at day 30, 60 and 90, what to protect them from, and the signals that the hire is not working.

Mauricio Esparza By ·Published ·2 min read
mithub.club
Short answer

By day 30 they should be able to describe your funnel end to end and name the biggest leak with evidence. By day 60 one system should be in production, documented, with an owner. By day 90 one number should have moved and they should be able to explain why. The most common way to waste the hire is filling the first month with small requests before they have seen the whole picture.

You hired the person. The next ninety days decide whether you bought a system or a dependency.

Days 1 to 30: understand before building

The temptation is a backlog on day one. Resist it. Small requests feel productive and they guarantee nobody ever looks at the whole picture, which is where the expensive leaks live.

What should happen instead:

  • Access on day one. Read access to the CRM, the ability to create a webhook, a sandbox, the reporting. Waiting three weeks for a seat is the least defensible way to lose a month.
  • Reading real records. Fifty of them, not the documentation. Documentation describes the process as designed. Records describe it as it runs.
  • Sitting on calls. Sales calls and support calls. The gap between what the CRM says and what happens on a call is usually where the first project is.
  • Mapping the funnel as it actually works, including the steps nobody wrote down.

The day 30 test: they describe your funnel end to end, name the biggest leak, and show you the evidence. If they cannot, either the access was late or the hire is wrong, and both are worth knowing now.

Days 31 to 60: one system in production

One. Not five. Narrow enough to finish, real enough to matter.

By day 60 it should be live, documented, with an owner who is not them, and an error route. The guidance from people who build these systems applies here too: start with the simplest thing that works and add autonomy only where it earns its cost.

The day 60 test: someone else runs it from the documentation, for a week, without calling them.

Days 61 to 90: the number moves

One number, explained. Plus a second thing in flight and a written backlog with reasons attached, which tells you they are thinking in priorities rather than in requests.

The day 90 test: they can walk you from the number back to what they changed, and forward to what they would do next and why.

What to protect them from

The ticket queue. The moment this role becomes the person who fixes things by hand, the systems stop being built. Route small requests elsewhere for the first sixty days.

Tool evaluations before the diagnosis. Vendors will find them. Nothing gets bought until the leak is named.

Being the CRM admin. Related and slower to notice. If they are the only person who can change a field, they will spend the year changing fields.

The signals it is not working

  • Month two ends with nothing in production
  • Only they can operate what exists
  • They cannot name something they deliberately chose not to build, which usually means they are building whatever is asked
  • Everything is described in tools rather than in outcomes

Any of these at day 60 is worth an honest conversation while it is still cheap.

The handover habit

The best version of this role produces things that survive its author. Ask for the runbook while it is being built, not at the end, and test it by having someone else follow it. A system nobody else can run is not finished, however well it works.

If you are still hiring

MitHub publishes open profiles with each person's deployments written out, including who operates them now.

See the talent.

Frequently asked questions

Should they build something in week one?

Only something small enough to be thrown away. Building before understanding produces work that has to be redone, and it costs the credibility that the first real project needs.

Who should they report to?

Whoever owns the number they were hired against. Reporting to neither sales nor marketing is the most reliable way to make the role fail, because nobody can decide what to stop.

What access do they need on day one?

Read access to the CRM, the ability to create a webhook, a sandbox, and the reporting they will be measured on. Access delays are the most common and least defensible cause of a slow start.

When should we worry?

If month two ends with nothing in production, if only they can operate what exists, or if they cannot name something they deliberately chose not to build.

Sources

  1. Building effective agents · Anthropic (accessed 2026-09-22)
  2. Why evaluate AI workflows · n8n Docs (accessed 2026-09-22)
GTM EngineeringOnboardingHiring
Mauricio Esparza
Mauricio EsparzaGTM Systems Lead · Revenue Engineer · Founder of MitHub. Designs and runs revenue systems for multi-location businesses: AI voice campaigns, enrichment, CRM automation and attribution. Founded MitHub to teach the method in the open.

Part of GTM Engineering on MitHub.

Keep going

GTM Engineering

Hiring a GTM Engineer Remotely: The Practical Guide

How to hire a GTM engineer remotely: what the role is accountable for, how to write it around one number, what to ask, and the paid trial that avoids a bad contract.

Read · 4 min →mithub.club
GTM Engineering

GTM Engineer Job Description (Template You Can Copy)

A job description for a GTM engineer written around accountability instead of tools, with the sections to keep, the ones to delete and a version you can paste and edit.

Read · 3 min →mithub.club
GTM Engineering

How to Hire a GTM Engineer for a Non-SaaS Company

Most advice about this role assumes a venture backed SaaS company. If you run branches, clinics, agencies or field teams, here is how the role changes and how to tell whether you need it.

Read · 4 min →mithub.club
Revenue Engineering

The Revenue Leak Map: A Free Template

A free Revenue Leak Map template: list every step from first contact to payment, measure the drop at each one, price the leak and rank which fix to build first.

Read · 7 min →mithub.club
Revenue Engineering

CRM Data Quality Checklist Before You Automate

Twelve checks to run on your CRM before automating anything on top of it, why each one breaks automation downstream, and the order to fix them in.

Read · 3 min →mithub.club
GTM Engineering

We Are Hiring a GTM Engineer. What Should We Look Out For?

What to look for when someone shows you their work: the four questions that separate a real deployment from a demo, and the answers that should make you pause.

Read · 2 min →mithub.club