Multi-Campaign GTM Architecture

Architecture: Campaign as First-Class Object

A campaign encapsulates everything needed to run an independent outbound motion:

Component	What It Defines
ICP	Who to target (firmographic + intent signals + scoring weights)
Scoring	How to score (criteria and weights specific to this ICP)
Sequences	What to say (touch templates, tone, CTA, cadence)
Collateral	Supporting materials (1-pagers, pitch decks, case studies)
Geography	Where to target (country/region filters)
Stages	Pipeline stages (can vary per campaign)
Timing	Cadence rules (days between touches, send windows)

1. Campaign Registry

New file at Marketing/GTM/campaigns.json:

{
  "version": "2.0",
  "campaigns": [
    {
      "id": "retirement-archive",
      "name": "Retirement Archive (EU Manufacturing)",
      "status": "active",
      "priority": 1,
      "icpFile": "campaigns/retirement-archive/icp-definition.md",
      "positioningFile": "campaigns/retirement-archive/positioning.md",
      "sequencesDir": "campaigns/retirement-archive/outreach-sequences/",
      "collateralDir": "campaigns/retirement-archive/collateral/",
      "geography": ["DE", "NL", "CH", "AT", "BE"],
      "stages": [
        "sourced",
        "recognition_qualified",
        "contacted",
        "replied",
        "interested",
        "invoiceSent",
        "paid",
        "shipped",
        "closed"
      ],
      "touchCount": 4,
      "cadence": {
        "t1_to_t2": 3,
        "t2_to_t3": 7,
        "t3_to_t4": 14
      },
      "scoring": {
        "model": "recognition-first",
        "weights": {
          "firmographic": 30,
          "recognition": 45,
          "accessibility": 25
        }
      },
      "created": "2026-02-03",
      "metrics": {
        "prospectCount": 20,
        "contacted": 14,
        "replied": 0,
        "converted": 0
      }
    },
    {
      "id": "uk-longservice",
      "name": "UK Long-Service Awards (HMRC Tax Advantage)",
      "status": "planned",
      "priority": 2,
      "icpFile": "campaigns/uk-longservice/icp-definition.md",
      "positioningFile": "campaigns/uk-longservice/positioning.md",
      "sequencesDir": "campaigns/uk-longservice/outreach-sequences/",
      "collateralDir": "campaigns/uk-longservice/collateral/",
      "geography": ["GB"],
      "stages": [
        "sourced",
        "recognition_qualified",
        "contacted",
        "replied",
        "interested",
        "invoiceSent",
        "paid",
        "shipped",
        "closed"
      ],
      "touchCount": 4,
      "cadence": {
        "t1_to_t2": 3,
        "t2_to_t3": 7,
        "t3_to_t4": 14
      },
      "scoring": {
        "model": "recognition-first",
        "weights": {
          "firmographic": 30,
          "recognition": 40,
          "accessibility": 30
        }
      },
      "created": "2026-02-07",
      "metrics": {
        "prospectCount": 0,
        "contacted": 0,
        "replied": 0,
        "converted": 0
      }
    }
  ]
}

2. Per-Campaign Folder Structure

Marketing/GTM/
├── README.md                           # Engine overview (updated)
├── campaigns.json                      # Campaign registry
├── research-process.md                 # Shared research methodology
├── log.md                              # Activity log (all campaigns)
│
├── campaigns/
│   ├── retirement-archive/             # EXISTING (migrated from root)
│   │   ├── icp-definition.md
│   │   ├── positioning.md
│   │   ├── outreach-sequences/
│   │   │   ├── t1-cold-intro.md
│   │   │   ├── t2-follow-up.md
│   │   │   ├── t3-value-add.md
│   │   │   └── t4-breakup.md
│   │   └── collateral/
│   │       └── retirement-batch-1pager.md
│   │
│   ├── uk-longservice/                 # NEW
│   │   ├── icp-definition.md           # UK manufacturing, HMRC angle
│   │   ├── positioning.md              # Tax-advantage framing
│   │   ├── outreach-sequences/
│   │   │   ├── t1-cold-intro.md        # HMRC hook
│   │   │   ├── t2-follow-up.md
│   │   │   ├── t3-value-add.md
│   │   │   └── t4-breakup.md
│   │   └── collateral/
│   │       └── hmrc-tax-guide.md
│   │
│   ├── nl-anniversaries/               # NEW
│   │   ├── icp-definition.md           # NL companies with jubileum culture
│   │   ├── positioning.md              # 12.5-year tradition angle
│   │   ├── outreach-sequences/
│   │   └── collateral/
│   │
│   └── prof-services/                  # NEW
│       ├── icp-definition.md           # UK/EU law & accounting firms
│       ├── positioning.md              # Partner gifting angle
│       ├── outreach-sequences/
│       └── collateral/
│
Sales/
├── pipeline/
│   ├── index.json                      # SINGLE pipeline (all campaigns)
│   └── SCHEMA.md                       # Updated schema
├── prospects/                          # Per-prospect folders (unchanged)
│   ├── _template/
│   ├── miele/
│   └── ...

3. Pipeline: Single Index, Campaign Field Per Prospect

Design decision: One index.json with a campaign field on each prospect — NOT separate files per campaign.

Why single file:

Cross-campaign deduplication: Instantly check if a company is already in another campaign
Global search: Find any prospect by company name regardless of campaign
Simpler tooling: One file to read/write, one schema to maintain
Aggregate reporting: Roll up stats across all campaigns

Updated prospect object (additions in bold):

{
  "id": "SKY-2026-021",
  "campaign": "uk-longservice",          // ← NEW: which campaign
  "company": {
    "name": "Rolls-Royce plc",
    "website": "https://rolls-royce.com",
    "industry": "Manufacturing",
    "size": "5000+",
    "country": "GB",
    "city": "Derby"
  },
  "contact": { ... },
  "scoring": {
    "total": 88,
    "firmographic": 28,
    "recognition": 35,
    "accessibility": 25,
    "notes": "HMRC long-service eligible. 25+ year award program."
  },
  "tier": "A",
  "stage": "sourced",
  "recognitionSignal": "...",
  "source": "firecrawl",
  "personalization": { ... },
  "dates": { ... },
  "outreach": { ... },
  "notes": [],
  "folderPath": "prospects/rolls-royce/"
}

4. Updated Stats Object

{
  "version": "2.0",
  "lastUpdated": "2026-02-07T12:00:00Z",
  "stats": {
    "total": 35,
    "byCampaign": {
      "retirement-archive": {
        "total": 20,
        "byTier": { "A": 5, "B": 15 },
        "byStage": { "sourced": 1, "contacted": 14, "recognition_qualified": 5 },
        "byCountry": { "DE": 14, "NL": 1, "CH": 1, "AT": 2, "BE": 2 }
      },
      "uk-longservice": {
        "total": 15,
        "byTier": { "A": 5, "B": 10 },
        "byStage": { "sourced": 15 },
        "byCountry": { "GB": 15 }
      }
    }
  },
  "campaigns": ["retirement-archive", "uk-longservice"],
  "prospects": [ ... ]
}

Cross-Campaign Rules

Deduplication

Before adding ANY prospect to ANY campaign:

Search index.json for matching company.website or company.name

If match found in SAME campaign → skip (already exists)

If match found in DIFFERENT campaign → flag for review:

- Same contact? → Don't send from two campaigns simultaneously - Different contact? → OK if different department/decision-maker - Log cross-campaign overlap in prospect notes

Contact Collision Prevention

A single contact should NEVER receive emails from two campaigns at the same time. Rules:

Scenario	Action
Same person, same company, different campaign	Choose highest-priority campaign. Pause the other.
Different person, same company, different campaign	OK — proceed with both (different decision-makers)
Same person moved to new company	OK — new company context

Campaign Priority

When conflicts arise, defer to campaign priority in campaigns.json. Lower number = higher priority.

Workflow Commands (Updated)

Discovery

"Source SKYMOD prospects for [campaign]"
→ Load ICP from campaigns/{campaign}/icp-definition.md
→ Run Firecrawl Agent with campaign-specific query
→ Dedupe against FULL pipeline (all campaigns)
→ Score using campaign-specific weights
→ Add to index.json with campaign field

Outreach

"Draft T1 for [company] in [campaign]"
→ Load T1 template from campaigns/{campaign}/outreach-sequences/t1-cold-intro.md
→ Read prospect research + persona
→ Generate personalized email
→ Save to prospects/{company}/outreach/t1.md

Status

"Show SKYMOD pipeline status" → Read campaigns.json → list all campaigns → For each: show prospect count, stage breakdown, next actions "Show SKYMOD pipeline for [campaign]" → Filter index.json by campaign → Show that campaign's metrics and upcoming actions

"Show all SKYMOD next actions" → Scan ALL campaigns for nextActionDate ≤ today → Group by campaign, sorted by date

Campaign Lifecycle

"Create SKYMOD campaign [name]" → Add entry to campaigns.json → Create campaign folder structure → Generate ICP definition template → Generate outreach sequence templates → Status: "planned" "Activate SKYMOD campaign [name]" → Set status: "active" → Run first prospecting batch "Pause SKYMOD campaign [name]" → Set status: "paused" → Skip all scheduled outreach for that campaign

"Archive SKYMOD campaign [name]" → Set status: "archived" → Move stats to archive, keep data for reference

Migration Plan (Existing Data)

SKYMOD Migration (retirement-archive)

Move existing files to campaign subfolder:

- Marketing/GTM/icp-definition.md → campaigns/retirement-archive/icp-definition.md - Marketing/GTM/positioning.md → campaigns/retirement-archive/positioning.md - Marketing/GTM/outreach-sequences/ → campaigns/retirement-archive/outreach-sequences/ - Sales/collateral/retirement-batch-1pager.md → campaigns/retirement-archive/collateral/

Add campaign field to each existing prospect in index.json:

   "campaign": "retirement-archive"

Update stats to nested byCampaign structure

Create campaigns.json with retirement-archive as first entry

Update README.md to reference new structure

Maad House Migration (hospitality)

Create campaigns.json with hospitality as first entry

Add campaign: "hospitality" field to each existing prospect

Create campaign folder: campaigns/hospitality/

Move/create ICP and sequence files into campaign subfolder

Maad House Adaptation

Maad House follows the same architecture with different campaign definitions:

Maad House/Marketing/GTM/
├── campaigns.json
├── campaigns/
│   ├── hospitality/              # EXISTING (migrated)
│   │   ├── icp-definition.md
│   │   ├── outreach-sequences/
│   │   └── collateral/
│   ├── legal-lawtech/            # NEW (Priority #1 expansion)
│   │   ├── icp-definition.md     # Linkilaw as proof point
│   │   ├── outreach-sequences/
│   │   └── collateral/
│   │       └── linkilaw-case-study.md
│   ├── ecommerce-dtc/            # NEW (Priority #2)
│   │   ├── icp-definition.md     # SKYMOD as proof point
│   │   ├── outreach-sequences/
│   │   └── collateral/
│   │       └── skymod-case-study.md
│   └── lifestyle-wellness/       # NEW (Priority #3)
│       ├── icp-definition.md
│       ├── outreach-sequences/
│       └── collateral/

Same pipeline rules: single index.json, campaign field per prospect, cross-campaign deduplication.

Dashboard Commands

After migration, the portfolio dashboard supports:

"Show SKYMOD pipeline status" → All campaigns summary: retirement-archive: 20 prospects (14 contacted, 0 replies) uk-longservice: 0 prospects (planned) nl-anniversaries: 0 prospects (planned) prof-services: 0 prospects (planned)

"Execute SKYMOD ICP expansion (UK → NL → Prof Services)" → Creates campaign entries → Generates ICP + sequence templates → Sets priority order → Ready for prospecting

Key Design Principles

Single source of truth: One pipeline file per business, not per campaign

Campaign isolation: Each campaign has its own ICP, messaging, and metrics

Shared infrastructure: Same Firecrawl, Hunter, Outlook tools across campaigns

Cross-campaign intelligence: A reply from Campaign A informs Campaign B for the same company

Incremental adoption: Migrate existing first, add new campaigns one at a time

No code required: Everything is file-based (JSON + Markdown). Claude reads the campaign config and executes accordingly.

Implementation Order

Step	Action	Effort
1	Create `campaigns.json` for both businesses	10 min
2	Add `campaign` field to existing prospects	5 min
3	Restructure folders (move existing files to campaign subfolders)	15 min
4	Update README files	10 min
5	Update SCHEMA.md with campaign field	5 min
6	Create first new campaign (UK Long-Service)	30 min
7	Run first prospecting batch for new campaign	15 min

Total: ~90 minutes to go from single-campaign to multi-campaign for both businesses.