Gtm

Marketing / GTM

This is Marketing OS's single GTM entry point. Turn messy launch context into a governing market decision, then add only the proof, conversion, operating, calendar, channel, or social layers the launch needs. Lead with the argument and intended behavior. Add execution only after the strategy is clear.

When to invoke

• When building or revising any GTM strategy, route-to-market plan, launch plan, launch backbone, campaign sequence, conversion plan, or launch PR brief
• When producing a dated drumbeat, production plan, or concise social brief
• When applying a scoped revision to an existing GTM artifact
• When leadership needs a 10- to 15-minute approval deck for a launch content ecosystem

Load the system

Read:

1. ../../foundation/marketing-os/SKILL.md
2. ../../foundation/marketing-os/references/method.md
3. ../../foundation/marketing-os/references/collaboration.md
4. ../../positioning/every-master/SKILL.md
5. ../../marketing-science/archetyping/references/every-audiences.md
6. ../../craft/editing/SKILL.md
7. ../../craft/editing/references/compound-from-edits.md
8. references/strategy-standard.md

For an existing collaborative document, also read ../../foundation/marketing-os/references/native-document-editing.md. Apply every-style and ai-check to final prose.

For a 10- to 15-minute leadership approval deck built around a launch content ecosystem, also read references/content-system-approval-deck.md. Treat the deck as the presentation layer of the approved GTM decision, not as a replacement for the full plan. This reference owns the deck structure; do not combine it with a second executive-summary template.

For an L or XL flagship launch, also read the approved ../../launches/launch-brief/SKILL.md output and ../../launches/canonical-process.md. For a curated or application-based event, read references/curated-event-gtm.md.

Choose the launch tier and artifact mode

Classify both axes before drafting. Artifact mode never replaces launch tier.

Launch tier:

• Flagship L/XL: Use the approved launch brief and flagship process. Preserve its gates, owners, work-back, and confirmation state. An unapproved brief blocks the downstream GTM decision.
• Standard S/M: Build a right-sized GTM decision from current positioning, customer evidence, offer, proof, and constraints. Do not import flagship timelines, approval rosters, or meeting requirements.

Artifact mode:

• Master GTM plan: Build the governing market decision and only the execution layers the launch needs.
• Scoped revision: Inherit the existing strategy and edit only the authorized surface. Reopen a decision only when the source owner changed it.
• Drumbeat or social brief: Inherit the approved GTM decision. Do not reconstruct strategy inside a channel artifact.
• Content-system approval deck: Inherit the approved GTM decision and use the deck to approve the campaign system, production mandate, and remaining inputs. Do not turn the review into a post-by-post editorial meeting.

One skill owns both launch tiers and all four artifact modes. Never route the user to a second GTM skill.

Establish authority

Identify the sources that own:

• Product or event facts.
• Strategic decisions.
• Canonical copy.
• Structural precedent.
• Dates, people, pricing, and publication plans.
• Approval.

Search the supplied Notion, Slack, Drive, meeting, and editorial sources before asking for information that can be found. Cite current sources. Label conflicts and unknowns.

When this work comes from strategy/compound-marketing, inherit the approved decision record before drafting. Record its version, locked decision IDs, protected-language IDs, and open questions in the GTM artifact. Do not reinterpret those decisions from the underlying source pile.

Interview before a long draft

Confirm these decisions when the materials do not settle them:

• What is launching?
• What is the primary action?
• What should each audience think, feel, or do?
• What story or language is protected?
• What is the launch window?
• Who can lock facts and strategy?

Ask one to three high-value questions at a time. Continue when the answers already exist.

Lock the strategy

Use references/strategy-standard.md to lock the five-line value hierarchy and seven decision questions: governing decision, market belief, primary customer, offer and first value, demand and proof engine, conversion and lifecycle, and learning and measurement. Keep the first six answers to one page. If they do not fit, the strategy is still unsettled.

Then express that decision through these launch elements before building a calendar:

1. Goal: State the future condition and audience behavior.
2. Story: Name the tension, the company's position, and why the launch follows.
3. Campaign concept: Define the mechanism that creates attention and participation.
4. Launch backbone: Explain how the story moves through company voices, credible participants, the public, and the offer.

Use the user's chosen language exactly. Keep protected sections unchanged during scoped edits.

Before polishing the prose, separate the strategic jobs:

• Audience and tension: Who needs this, and what is stopping them?
• Category and outcome: What is the product, and what changes because it exists?
• Mechanism: How does the product create that outcome?
• Reason to believe: Why can Every credibly deliver it?
• Proof: What specific request, action, and result show it happening?

Make each sentence and section do one job. Do not collapse the product promise, positioning, hero UVP, supporting UVPs, mechanism, and reason to believe into one paragraph. Name the customer and the first champion separately when they differ. Lock one hero UVP; use the others as supporting stories rather than competing headlines.

Treat the one-line strategy and positioning statement as decision tools, not website copy. Do not move into pricing, metrics, calendars, or channel tactics while the product story is still being settled unless the user makes one of those the current decision.

Write proof at production specificity

Every product example must name:

1. The request.
2. The action the product took.
3. The result.

Explain causal links an unfamiliar reader could not infer. Retrieve the missing detail from Slack, Notion, Drive, or the governing source before writing. Never replace a specific field, error, file, fix, artifact, or outcome with mystery language.

Do not use “real” as a credibility word. Replace it with the exact claim: shipped, published, corrected, approved, customer-facing, live, verified, or another concrete fact. If the evidence is incomplete, mark the gap instead of inventing intrigue.

Build the master GTM plan

Use the relevant parts of this order:

1. Decision-record version and inherited decision IDs, when supplied.
2. Launch tier, confirmation state, facts, and primary action.
3. The seven-section decision core from references/strategy-standard.md.
4. Goal, story, campaign concept, and launch backbone.
5. The chosen proof, seeding, editorial, PR, event, video, or evidence modules.
6. Launch production and conversion ownership.
7. Launch sequence and, when required, the dated drumbeat.
8. Social brief link and copy, design, build, and QA surfaces.
9. Scorecard, decision date, open questions, and sources.

Right-size the document. Use one primary decision metric, a small diagnostic set, and guardrails. Use the fuller business, conversion, proof, and brand scorecard when the launch has targets, capacity, revenue, or lifecycle decisions. Do not invent baselines or targets.

For a dated drumbeat:

• Use calendar days, including weekends.
• Distinguish confirmed dates from proposed workback dates. Label derived dates as proposed.
• Run strategic lanes simultaneously when the campaign requires it.
• Give each day a new argument, participant, proof point, program detail, or reason to act.
• Match editorial moments only when the actual article, speaker, or contributor argument supports the connection.
• Use N/A when no credible connection exists.
• Name owners and dependencies only when they help the team ship.
• Keep undecided pricing, talent, channel, and publishing choices under open questions.

Read references/document-architecture.md for the execution-layer format.

For a content-system approval deck, use references/content-system-approval-deck.md. Lead with the exact decision, organize the ecosystem by strategic job, prove every format with a literal example, translate the approved cadence into a credible calendar, and close with the work to lock, make, publish, and ship.

Build the social brief separately

Default to three jobs:

1. State the story social must tell.
2. Name the actions and behavior social should elicit.
3. List concrete ideas to discuss.

Keep it short. Do not repeat the GTM plan or turn the brief into a channel manual unless the user asks for one. Include specific examples when they sharpen an idea. For events, make desire, participation, and application behavior explicit.

Verify before delivery

• Check every date, name, title, quote, speaker status, contributor statement, editorial item, price, and link against an authoritative source.
• Distinguish a locked decision from a missing production value. For example, pricing can be approved while the exact display price is still absent from the source packet.
• Quote the actual statement when claiming a thematic match.
• Mark approval dependencies explicitly.
• Preserve unresolved questions.
• Compare the final artifact with every inherited locked decision and protected phrase. Restore drift or record the human-approved replacement in the decision ledger.
• Remove generic language, repeated strategy, and unsupported specificity.
• Avoid “room,” “who belongs in the room,” and related event-marketing shorthand.

Use references/edit-deltas.md for the durable lessons behind these rules.

Revise with a scoped diff

When the user identifies a section boundary:

1. Fetch the latest document.
2. Record the protected prefix and suffix.
3. Replace only the authorized section.
4. Re-fetch the document.
5. Compare protected content exactly.

Treat the user's edited version as canonical. Learn from the delta before drafting again. When the user rejects an edit, restore their wording immediately and carry that decision through the authorized scope. Preserve parentheticals, fragments, provocations, and rough one-two-punch rhythms when they carry the idea or voice; tighten the connective prose around them.

Compress only after the strategy and canonical language are stable. Cut repetition, throat-clearing, and process detail. Do not cut a unique idea, mechanism, proof point, owner, creative detail, or unresolved decision.

Evaluate

Use evals/cases.json and evals/rubric.md. Run scripts/check_output.py for deterministic checks, using --mode approval-deck for a content-system approval deck. The deterministic check covers structure and prohibited approval behavior; source fidelity, invention, and strategic quality still require rubric review against the governing sources. The skill passes only when it clears the score threshold with zero hard failures.

Dependencies

• foundation/marketing-os
• positioning/every-master
• marketing-science/archetyping
• marketing-science/measurement
• marketing-science/research
• craft/editing
• launches/launch-brief and launches/canonical-process.md (L/XL only)
• launches/claims-clearance

Quick checklist

• Authority and source roles are explicit
• The work has a launch tier and an artifact mode: master plan, scoped revision, drumbeat/social brief, or content-system approval deck
• S/M work did not inherit flagship gates, rosters, or timelines
• Ambiguous strategic questions were resolved before the long draft
• Governing decision, market belief, customer, offer, proof, journey, and scorecard are locked
• Goal, story, concept, and launch backbone precede the calendar
• The artifact names its inherited decision-record version and preserves its locked IDs
• Dates, people, claims, prices, and links are verified or clearly marked open
• Positioning separates audience, category, outcome, mechanism, and reason to believe
• Every proof point names the request, action, and result
• User-edited language and voice-bearing details remain canonical
• Measurement names a primary decision metric and decision date
• Scoped revisions preserved protected content exactly
• The relevant evaluation and deterministic checks passed
• A content-system approval deck states the review decision, preserves the approved cadence and any supplied content-role mix, uses literal calendar copy, and ends with executable asks whose open-input count matches the governing sources