Gtm Engine

GTM Engine — Custom Plan Generator

Design a revenue-generating GTM machine for a specific brand, the way a senior GTM engineer at a $100M company would: signals over static lists, systems over headcount, feedback loops over gut feel.

What you produce

A single, complete GTM Engine Plan markdown document following references/plan-template.md exactly. It covers: ICP/TAM, signal map, lead collection, enrichment, segmentation & scoring, nurture, outbound engine (infra + cadences + example copy), outbound management, tool stack with budget, metrics/feedback loops, 90-day roadmap, and brand-specific risks.

Reference files — read before writing

1. references/playbook.md — the full GTM engineering system (architecture, all 6 stages, benchmarks, automation blueprints, roadmap, anti-patterns). Read this first, always. It is the source of truth for methodology, numbers, and patterns; do not improvise benchmarks or deliverability rules that contradict it.
2. references/demand-layers.md — the modules beyond the core engine: social listening, community-led growth, referrals/advocacy/reviews (the warm layer), partnerships/nearbound, paid/ABM, PLG/PQL motion. Read whenever any module applies (its selection cheat-sheet tells you which do — for most brands at least 2 apply; an outbound-only plan is usually incomplete).
3. references/tool-stack.md — tool choices by budget tier. Read when writing the Tool Stack section.
4. references/plan-template.md — the exact output structure and quality bar. Read before drafting.
5. references/example-plan.md — a condensed worked example (fictional brand) showing the quality bar per section: derivation from the Step-2 forks, vertical-specific signals, consistent volume math, passing email copy. Skim before drafting your first plan in a session; match its density, not its content.
6. references/b2c-dtc.md — the full stage re-mapping for brands whose end buyer is a consumer (DTC/ecommerce, apps, consumer services): list growth, zero-party data, RFM/lifecycle segmentation, lifecycle flows, creator/wholesale outbound, retention ops. Mandatory read for any B2C or hybrid brand — it replaces playbook §4–9 for the consumer side.

Step 1 — Gather brand inputs

Required inputs (collect from the user's message, an attached brand doc, brand memory, or the current project context):

If the user provided a brand profile (e.g., via the brand-context/brand-memory skills or a pasted doc), extract these from it. Ask for missing inputs only if 3+ of them are unknown and can't be reasonably inferred — otherwise proceed and state assumptions inline in the plan. A plan with clearly-labeled assumptions beats an interrogation.

Step 2 — Derive the shape of the motion

Before writing, decide (and let these decisions drive every section):

1. TAM size class → volume strategy (playbook §3.2): <2k accounts = low-volume high-touch; 2k–20k = hybrid; >20k = programmatic. This is the single most important fork — a small-TAM brand given a volume plan is malpractice.
2. Deal-size class → touch budget: ACV <$3k = automation-only + PLG-ish; $3k–25k = hybrid; >$25k = human-heavy fast lanes, phone layer.
3. Buyer geography → channel + compliance posture (EU-heavy → LinkedIn-first consideration, Cognism).
4. Budget tier → tool stack (tool-stack.md Tier 1/2/3).
5. The brand's 5–8 strongest signals from playbook §4.3, plus any vertical-specific signals you can identify for their market.
6. Which demand-layer modules apply — use the selection cheat-sheet at the end of demand-layers.md (e.g., PLG brand → PQL motion mandatory; DTC tool vendor → listening + platform partnerships; founder-led → build-in-public community + manual warm layer).

Step 3 — Write the plan

Follow plan-template.md section by section. Anchor everything in the playbook's patterns and benchmarks. Brand-specificity checks:

• Volume math must chain correctly: goal → prospects → sends/day → inboxes → domains → cost (playbook §8.1).
• Example emails: write them in the brand's plausible voice, each anchored to a specific signal from the signal map. 50–90 words, per playbook §8.3.
• Suggest actual lookalike domain names for the brand.
• Tool stack: every pick gets a brand-tied reason. Prices and tool capabilities in tool-stack.md decay — if web search is available, spot-check current pricing for the 3–4 most expensive picks before quoting a budget; if not, label the budget "verify current pricing" and date it. If a listed tool seems to have been discontinued or renamed, substitute the current category leader and say so.
• Risks section: name the failure modes this brand is most exposed to, not generic ones.

Step 4 — Deliver

• Save the plan as a markdown file named GTM-ENGINE-PLAN-{brand}.md in the working directory (or where the user asks).
• In chat, give a tight summary: the motion in 3 sentences, the volume math, total budget, and the first 3 actions for this week.
• Offer natural next steps: draft the full sequences, build the Clay/n8n blueprint specs, or generate the ICP query for a specific database.

Scope notes

• If the user only wants ONE stage (e.g., "just design my cold email infrastructure"), read the playbook and produce only that section at full depth — don't force the whole template.
• If the user asks to execute (actually send emails, build Clay tables, launch campaigns), this skill produces the design + specs; hand off execution to the relevant tools/skills (e.g., Apollo MCP, the email skill for copy variants) and say so.
• B2C/DTC brands: read references/b2c-dtc.md and use its stage re-mapping — do not force the B2B template. Hybrid brands (DTC + wholesale/creator motion) get both: the B2C file for the consumer side, the core playbook for the business side.
• Not this skill's job (defer to more specific skills/tools when the user wants only that): writing a single piece of content or one email sequence (content/email skills), post-meeting sales process — discovery, proposals, closing (sales skills), pure content/SEO strategy. This skill owns the system design that connects them.