📤 gtm — Outbound & Developer GTM Department

One head skill for outbound and developer go-to-market. Pipeline is built, not found: discover the pain via technical advisory interviews, research the accounts, score the leads, sequence the outreach, keep the list clean, arm developer champions, coach founder-led sales, and hand over context sales can close. Every mode runs handsfree — defaults are stated inline, assumptions recorded in the deliverable, zero questions asked.

Modes

Every invocation resolves to exactly one mode. Match the request, then load only the matched reference:

Only the resolved mode's reference is loaded — the rest stay on disk, saving tokens on every run.

When to Use

• Discovering customer pain and validating ICP assumptions via Technical Advisory Board (TAB) interviews.
• Researching target accounts and leads before outbound.
• Scoring leads, sizing TAM/SAM, and prioritizing prospects.
• Running a cold email sequence with follow-ups and breakup.
• Building, cleaning, and enriching a prospect list.
• Converting bottom-up developer usage into paid enterprise deals via champion enablement.
• Coaching technical founders through live sales calls and handling commercial stalls.
• Handing qualified pipeline to sales with full context.

Anti-Triggers

• Paid advertising campaigns → paidads.
• High-level brand launch strategy and narrative positioning → growth (gtm owns outbound execution and developer sales; growth owns overarching launch strategy).
• Email copy craft and wording → content (gtm owns sequencing, qualification, and timing; content owns the words).
• Social-channel outreach and viral carousels → smm.
• CRM setup and automated webhook plumbing → automation.
• Deep web traffic analytics → analytics.

Quick Reference

Routing ladder (decide before any mode)

Order for zero-to-one: discovery → research → score → list → outreach → dev-to-buyer → founder-sales → handover. Validate the pain, find the accounts, rank them, clean the list, sequence, arm champions, close, then hand over.

Metrics loop (post-handover measurement)

Source: tech-leads-club/agent-skills gtm-metrics (orig chadboyda/agent-gtm-skills), raw SKILL.md fetched per lane-a-skillshub-remainder.md #13; enrich-only, no new skill or mode.

Run this loop on the outbound the modes above produce; deep analysis stays in analytics.

1. Select 5–7 metrics max: pipeline coverage (3–4x sales-led, 2–3x PLG), CAC payback (<8 mo), Magic Number (>0.75 efficient, >1.0 excellent, <0.5 red flag), NRR (>106% median, >120% best-in-class), TTFV (<15 min self-serve, <1 day sales-led), slippage (<15% weekly), speed-to-lead (<5 min).
2. Dashboard in three tiers: board (5–7 metrics, monthly) → executive (10–12, weekly) → operator (15–25, daily). Every metric from system-of-record APIs — no hand-edited slides; every metric carries benchmark, target, or trend line.
3. Attribution: match lookback to cycle (90d SMB, 180d mid-market, 365d enterprise); pre-revenue start first-touch, $1–5M U-shaped, $5–20M W-shaped; tag AI-SDR touches with source=AI-SDR so agent-led pipeline stays visible.
4. Weekly cadence (30–45 min): scorecard → pipeline movements → leading indicators → deals at risk → 2–3 owned actions → one rotating deep-dive. Maintain ~60% leading / 40% lagging indicators.
5. Data health gate: score CRM completeness/accuracy/recency/consistency before trusting pipeline reports; below 70% = stop trusting reports, clean first (data decays ~2.1%/mo).

Verification gate (every mode)

• Ground truth loaded: ./.agents/context/product.md and ./.agents/context/roadmap.md read before any mode executes.
• The ideal customer profile is stated before any account or lead is touched.
• Every researched claim carries a source and a date; unverified fields are marked as assumptions.
• Sending volumes respect caps and deliverability rules; every sequence ends in a breakup.
• The handover packet lets sales run the first call with zero extra research.

Suite contracts

• Positioning and launch strategy behind the outbound → growth.
• Email and asset wording → content.
• Tracking replies, meetings, and pipeline → ops; measuring results → analytics.

Procedure

1. Context grounding. Read ./.agents/context/product.md and ./.agents/context/roadmap.md to establish current ICP, validated claims, open assumptions, and the active "Now" move.
2. Resolve the mode. Match the request against the mode table; exactly one mode. Ambiguous → default to discovery if assumptions unvalidated, else research.
3. Execute the mode playbook. Load references/<mode>.md and follow its Intake → Deliverable → Procedure → Quality gate in order.
4. Gate and deliver. Pass this file's Verification checklist plus the mode's quality gate, update ./.agents/context/product.md and ./.agents/context/roadmap.md if reality or assumptions changed, then deliver the artifact.

Pitfalls

• Outbound without a written ICP — every list looks good and none convert.
• Pitching in discovery calls instead of listening to past behavior.
• Asking developer champions to sell instead of arming them with executive ROI packets.
• Giving a feature tour demo instead of showing the solution to their specific pain point.
• Researching without sources and dates — stale intel kills first calls.
• Scoring on gut instead of fit plus intent signals.
• Sending without warm-up, caps, or breakup — burned domain, no replies.
• Dirty lists: duplicates, role-based addresses, and unverified emails.
• Handover without context — sales re-researches everything, momentum dies.
• Confusing sequencing (gtm) with copy craft (content) — fix timing first, words second.
• Dashboard without trust — 50-metric screens, vanity metrics with no benchmark, or reports built on sub-70% data health.

Verification

• ./.agents/context/product.md and ./.agents/context/roadmap.md loaded as ground truth.
• Ideal customer profile stated with explicit role and company profile.
• Product claims carry [validated] vs [assumption] tags.
• Researched claims sourced and dated; assumptions marked.
• Scores trace to fit plus intent signals, not gut.
• Sequence states volumes, caps, spacing, and a breakup step.
• List is deduplicated, verified, and enrichment-checked.
• Handover packet carries context, history, and suggested next step.
• Metrics loop closed: 5–7 metrics selected with targets, dashboard tiered from system-of-record data, attribution lookback stated, weekly review scheduled.