Marketing Growth Suite
v2.0 — international B2B technology marketing. The suite is being
refactored from a PLG/indie-SaaS roster to an 8-domain B2B marketing
org with a brief-driven delegation protocol. This README describes the
v2 target state. The rollout is phased — see Rollout status below.
An AI-powered marketing team for international B2B technology
companies — a CMO orchestrator coordinating 8 domain orchestrators and
~34 specialists, plus Tier-1 workflows that string them together.
Distributed as both a native Claude Code plugin and a BMAD framework module. One physical copy of every skill; pick whichever installer you prefer.
Forked from MatthiasMRC/bmad-marketing-growth — original Marketing Growth Suite by Matthias (@matthias_mrc).
Source Fidelity
Every agent enforces a Source Fidelity rule (see the dedicated section in each SKILL.md):
- No invented metrics, benchmarks, conversion rates, or audience sizes. Numbers must come from the user's input or a cited
WebSearch result.
- Missing data triggers a focused question, not a plausible default.
- Sample emails / posts / dashboards with numbers are explicitly labeled
Example — illustrative, not benchmarks.
- Persona backstory is treated as tone calibration, not authority — agents won't invent past clients, study results, or proprietary stats to fit the character.
This is the main behavioral difference from a generic AI marketing assistant.
v2 architecture
Org chart
Max Growth — CMO Orchestrator
Pixel Metrics — Measurement & Attribution Staff (scope: attribution, forecasting, dashboards, scoring model design)
├── Product Marketing
│ ├── Positioning & Messaging PMM
│ ├── Competitive Intelligence PMM
│ ├── Launch & Sales Enablement PMM
│ └── Sales Presentation Advisor (Pierce; cross-orchestrator service)
│
├── Brand
│ ├── Brand Narrative & Voice Strategist
│ └── Localization & International Lead
│
├── Content Marketing (Milo Page, elevated)
│ ├── Content SEO Strategist (Quinn, renamed)
│ └── Technical Content Writer
│
├── Digital Marketing
│ ├── Web & CRO Specialist
│ ├── Paid Search Specialist (Google/Bing, PMax)
│ ├── Paid Social & Demand Specialist (LinkedIn/Meta/Google Demand Gen/YouTube Ads)
│ ├── Marketing Automation Engineer (HubSpot/Marketo workflows, lead routing)
│ └── Technical SEO Engineer (infra, schema, Core Web Vitals, hreflang)
│
├── Growth Marketing
│ ├── Lifecycle & Activation (Ember, repositioned)
│ └── Experimentation & Funnel Lead
│
├── Field Marketing
│ ├── ABM Strategist
│ ├── Events & Webinars Producer
│ ├── Customer Advocacy & References
│ └── Social Media Strategist (Nova; orchestrator, logically nested)
│ ├── LinkedIn Creator (Ivy)
│ └── YouTube Strategist (Yuri)
│ (Twitter and Reddit are Nova's capabilities, not standalone agents)
│
├── PR & Communications
│ ├── Media Relations Specialist
│ └── Analyst Relations Specialist (Gartner, Forrester, IDC)
│
└── Channel & Partner Marketing
└── Partner & Marketplace Manager
Brief-driven delegation protocol
Orchestrators don't dispatch via chat — they write a brief.md under
_bmad-output/work/{deliverable-id}/. The specialist reads the brief,
produces v1.md, and the orchestrator reviews against the acceptance
criteria. Verdicts are APPROVED, NEEDS_REVISION (with a numbered
list of required changes), or ESCALATED (after max_revisions is
hit). On approval, state.yaml is written with frozen: true and the
folder becomes part of the company's marketing record.
Full schemas, state machine, and rules: docs/protocol.md.
Three memory tiers
| Tier | Purpose | Owned by | Lifecycle |
|---|
_bmad-output/company-context/ | Shared, durable facts (ICP, positioning, KPIs…) | Per-file owner | Read-mostly, refreshed |
_bmad-output/work/{id}/ | Per-deliverable trail (brief, versions, review) | Issuing orch. | Frozen on accept |
_bmad/_memory/{code}-sidecar/memories.md | Per-agent private notes | The agent itself | Free-form, append on save |
The sidecar pattern is unchanged from v1. See docs/company-context.md for the company-context schema and ownership table.
One-time setup: bootstrap company context
