LinkedIn Campaign Development

Purpose

End-to-end LinkedIn campaign development: from lead enrichment through to campaign-ready personalized messages. This skill embeds Identity-First messaging as the core methodology, not an afterthought.

The Personalized Observation Philosophy

Most outreach messages focus on WHAT people did (achievements, awards, milestones). These prove research was done. But messages that resonate deeply focus on WHO they are (values, mission, philosophy, the thread connecting their choices). These create CONNECTION.

The Pattern:

[Observation] + [Reaction/Closer]

Two sentences maximum. Show you've done your homework without pitching anything. This observation follows "Thanks for connecting, {{firstName}}." and precedes a question about AI plans.

5-Phase Workflow Overview

Critical Rules

File Storage (Mandatory)

All campaign files live in ./workspace/campaign-{campaign-name}/:

workspace/
└── campaign-{campaign-name}/
    ├── campaign-brief.md               ← Campaign configuration
    ├── campaign_research_audit.csv     ← Final merged output
    ├── all_leads_master.csv            ← Status tracking
    └── batches/
        ├── batch_001/
        │   ├── leads_extracted.json
        │   ├── leads_enriched.json
        │   └── messages_draft.json
        └── batch_002/
            └── ...

Naming Convention: Use kebab-case for campaign folder names (e.g., campaign-charlotte-jerome, campaign-faie-accelerator).

Incremental Save Protocol

After processing each lead, immediately save to prevent data loss:

1. Read existing batch file
2. Append new lead data
3. Write back to file

This ensures no data loss if context runs out mid-batch.

Subagent Delegation

• Use general-purpose agents for batch enrichment research
• Use haiku model for validation tasks (faster, cheaper)
• Never process more than 5 leads without saving

Phase Navigation

Load the appropriate phase reference file when working on that phase:

Phase 1: Enrichment

When: Starting a new campaign or processing new leads Load: references/phases/phase-1-enrichment.md Focus: Gather 4 dimensions (Individual, Activity, Company, Opportunities) and validate

Phase 2: Message Crafting

When: Enrichment complete, ready to write messages Load: references/phases/phase-2-message-crafting.md Focus: Write personalized observations (2 sentences: observation + reaction closer)

Phase 3: Quality Assurance

When: Draft messages ready for validation Load: references/phases/phase-3-quality-assurance.md Focus: Perplexity sense-check with full context, pilot user review (~5 leads)

Phase 4: Scale

When: Pilot approved, ready to process remaining leads Load: references/phases/phase-4-scale.md Focus: Apply approved approach to all remaining leads in batches

Phase 5: Finalize

When: All leads processed, ready for final review Load: references/phases/phase-5-finalize.md Focus: Merge batches, final user review partnership, export

Required Inputs

Before starting any campaign:

1. Campaign Directory Setup

   • Create campaign folder: ./workspace/campaign-{campaign-name}/
   • Create batches subfolder: ./workspace/campaign-{campaign-name}/batches/
   • Example: ./workspace/campaign-charlotte-jerome/

2. Campaign Brief (Mandatory)

   • Located in campaign folder: ./workspace/campaign-{campaign-name}/campaign-brief.md
   • Contains: offer, target persona, value proposition, message templates
   • Example: ./workspace/campaign-faie-accelerator/campaign-brief.md

3. Lead List

   • From LinkedIn Sales Navigator or provided CSV
   • Minimum fields: firstName, lastName, company, title, linkedInUrl

4. PFC Criteria

   • Perfect Future Customer definition
   • See references/guides.md for exclusion/inclusion criteria

Output Files

All outputs are stored within the campaign directory:

Reference Files

Phase Guides

• references/phases/phase-1-enrichment.md - 4-dimension data gathering
• references/phases/phase-2-message-crafting.md - Personalized observations (2 sentences)
• references/phases/phase-3-quality-assurance.md - Perplexity + pilot review
• references/phases/phase-4-scale.md - Batch processing remaining leads
• references/phases/phase-5-finalize.md - Merge and final review

Supporting Guides

• references/guides.md - PFC criteria + CSV schema
• references/troubleshooting.md - Common issues and solutions

Scripts

• scripts/batch_update_linkedin_urls.py - Update CSV with LinkedIn URLs
• scripts/csv_to_multilead_json.py - Export to Multilead format
• scripts/merge_csv_batches.sh - Merge batch files

Examples

• examples/campaign-brief-template.md - Campaign brief structure
• examples/enrichment-example.json - Sample enrichment data
• examples/message-examples.md - Personalized observation examples

Quick Start

1. Create campaign directory: mkdir -p ./workspace/campaign-{name}/batches
2. Read campaign brief from ./workspace/campaign-{name}/campaign-brief.md
3. Load Phase 1 (references/phases/phase-1-enrichment.md)
4. Progress sequentially through phases
5. Save incrementally after each lead to ./workspace/campaign-{name}/batches/
6. User reviews at Phase 3 (pilot) and Phase 5 (final)

Resuming Mid-Campaign

If context was lost mid-campaign:

1. Check ./workspace/campaign-{name}/batches/ for latest batch files
2. Review ./workspace/campaign-{name}/all_leads_master.csv for status
3. Load appropriate phase reference
4. Continue from last saved state