Meta Ads Campaign Builder Agent Skill

Meta Ads Campaign Builder

Build a complete Meta Ads campaign structure — targeting, ad sets, placements, copy framework, and budget allocation. This skill handles the strategic architecture that determines whether a campaign succeeds before a single ad runs.

Core principle: The biggest Meta Ads mistake startups make is boosting a post and calling it a campaign. Proper campaign structure — objective selection, audience layering, placement optimization, and testing framework — is what separates burning money from building pipeline.

When to Use

• "Set up Meta Ads for our product"
• "Build a Facebook/Instagram ad campaign"
• "Help me structure a Meta campaign"
• "I want to run ads on Facebook — where do I start?"
• "Create a Meta Ads plan for [product launch / lead gen / awareness]"

Phase 0: Intake

• Product name + URL — What are we advertising?
• Campaign objective:
• Awareness (brand reach)
• Traffic (website visits)
• Lead Generation (in-platform lead forms)
• Conversions (website sign-ups / purchases)
• App Installs
• ICP — Target buyer: role, company size, industry, pain points
• Monthly budget — How much for Meta specifically?
• Landing page(s) — Where will traffic go?
• Competitor names — For audience analysis
• Existing Meta Pixel / Conversions API? — Tracking setup status
• B2B or B2C? — Changes targeting strategy significantly

Phase 1: Campaign Architecture

1A: Objective Selection

| Business Goal | Meta Objective | Why |

|--------------|---------------|-----|

| "Get demos/leads" | Lead Generation OR Conversions | Lead Gen = in-app forms (higher volume, lower quality). Conversions = website (lower volume, higher quality) |

| "Drive free trial sign-ups" | Conversions | Optimize for on-site conversion event |

| "Build awareness for launch" | Awareness (Reach) | Maximize eyeballs in target audience |

| "Retarget website visitors" | Conversions | Bring warm traffic back to convert |

| "Drive traffic to content" | Traffic | Optimize for clicks to blog/resource |

1B: Campaign Structure

Campaign: [Product Name] — [Objective]
├── Ad Set 1: [Audience — Prospecting: Interest-Based]
│   ├── Ad 1: [Primary creative variant]
│   ├── Ad 2: [Secondary creative variant]
│   └── Ad 3: [Tertiary creative variant]
├── Ad Set 2: [Audience — Prospecting: Lookalike]
│   ├── Ad 1-3: [Same creative variants]
├── Ad Set 3: [Audience — Retargeting: Website Visitors]
│   ├── Ad 1-3: [Retargeting-specific creative]
└── Ad Set 4: [Audience — Retargeting: Engagement]
    └── Ad 1-3: [Engagement retargeting creative]

1C: Competitor Ad Research (Optional)

If competitors were provided, check what they're running on Meta:

web_search: site:facebook.com/ads/library "[competitor name]"
web_search: "[competitor name]" Meta Ads examples

The Meta Ad Library (facebook.com/ads/library) is publicly accessible — search for competitor pages to see their active ads, formats, and messaging angles. Use findings to inform copy angles and audience strategy.

Phase 2: Audience Strategy

2A: Prospecting Audiences

Interest-Based Targeting:

Research relevant interests, behaviors, and demographics:

web_search: [product category] Meta Ads targeting options
web_search: [ICP role] Facebook ad audience interests

Build audience layers:

| Layer | Targeting | Rationale |

|-------|----------|-----------|

| Job title / Industry | [Specific titles, industries] | Direct ICP match |

| Interests | [Tools they use, publications they read, topics they follow] | Behavioral proxy |

| Behaviors | [Business decision makers, technology early adopters, etc.] | Meta behavioral data |

| Demographics | [Age range, education, income (if B2C)] | Narrowing |

Audience size target: 500K-2M for prospecting (smaller = expensive, larger = diluted)

Lookalike Audiences:

| Source | Lookalike % | Rationale |

|--------|------------|-----------|

| Customer list (emails) | 1% | Closest match to actual buyers |

| Website converters (Pixel) | 1% | People who took action |

| Website visitors (Pixel) | 1-3% | Broader interest signal |

| Page engagers | 3-5% | Widest cold audience |

2B: Retargeting Audiences

| Audience | Window | Purpose |

|----------|--------|---------|

| All website visitors | 30 days | Broad retarget |

| Visited pricing/demo page | 14 days | High intent — push to convert |

| Engaged with ads (no click) | 30 days | Awareness → consideration |

| Video viewers (50%+) | 30 days | Warmed but not clicked |

| Lead form openers (not submitted) | 14 days | Abandoned lead capture |

2C: B2B-Specific Adjustments

B2B targeting on Meta is harder. Recommended approach:

• Upload customer email lists → Build lookalikes (best B2B signal)
• Layer interests with behaviors — "Business decision makers" + "[specific software interest]"
• Use LinkedIn for precision targeting — Meta for retargeting/awareness
• Exclude broad audiences — No one under 25, exclude students, exclude job seekers (unless relevant)

Phase 3: Ad Copy Framework

Per Placement, Generate Copy Structure

| Placement | Headline Limit | Primary Text Limit | Notes |

|-----------|---------------|-------------------|-------|

| Feed (FB + IG) | 40 chars | 125 chars visible (500 max) | Most versatile |

| Stories (FB + IG) | 40 chars | Minimal text — visual-first | CTA button matters most |

| Reels | Overlay text only | Hook in first 3 seconds | Video required |

| Right column (FB) | 40 chars | Short | Desktop only — cheap impressions |

| Audience Network | 40 chars | 90 chars | Lower quality traffic |

Copy Framework Per Ad

| Element | Framework | Example |

|---------|-----------|---------|

| Primary text | [Hook → Pain/Outcome → Proof → CTA] | "Still prospecting manually? [Product] finds qualified leads while you sleep. 500+ teams use it. Try free →" |

| Headline | [Direct benefit or action] | "Automate Your Outbound in 5 Min" |

| Description | [Support or specificity] | "No credit card required. Cancel anytime." |

| CTA button | [Match to objective] | Learn More / Sign Up / Get Started / Book Demo |

Generate 3-5 Ad Copy Variants Per Ad Set

| Variant | Angle | Primary Text Hook | Headline |

|---------|-------|------------------|----------|

| 1 | Pain | "[Pain point question]" | "[Solution benefit]" |

| 2 | Outcome | "[Result promise]" | "[Specific metric]" |

| 3 | Social proof | "[Customer result + name]" | "[Join X teams]" |

| 4 | Contrarian | "[Myth busting]" | "[Unexpected claim]" |

| 5 | Product-led | "[Feature highlight]" | "[Feature → benefit]" |

Phase 4: Budget & Bidding

Budget Allocation

| Budget Tier | Prospecting | Retargeting | Testing |

|------------|-------------|-------------|---------|

| < $1K/mo | 60% | 30% | 10% |

| $1K-5K/mo | 50% | 30% | 20% |

| $5K+/mo | 45% | 25% | 30% |

Bidding Strategy

| Objective | Recommended Bid Strategy | When to Switch |

|-----------|------------------------|----------------|

| Conversions | Lowest Cost (start) → Cost Cap (after 50 conversions) | Once you have conversion data |

| Lead Gen | Lowest Cost | Usually sufficient for lead forms |

| Traffic | Lowest Cost per Click | Keep it simple |

| Awareness | Lowest Cost per 1K Impressions | Maximize reach |

Learning Phase

Critical: Each ad set needs ~50 conversions/week to exit learning phase. If your budget can't support that:

• Consolidate ad sets (fewer, larger audiences)
• Use an earlier funnel event as the optimization target (e.g., "Add to Cart" instead of "Purchase")
• Start with Traffic or Landing Page Views objective, switch to Conversions later

Phase 5: Output Format

# Meta Ads Campaign Plan — [Product Name] — [DATE]

## Campaign Overview
- **Objective:** [Selected objective]
- **Monthly budget:** $[X]
- **Target audience:** [ICP summary]
- **Geographic targeting:** [Countries/regions]
- **Placements:** [Automatic / Manual selection]
- **Conversion event:** [What we're optimizing for]

---

## Campaign Structure

[Visual tree]

---

## Audience Targeting

### Ad Set 1: [Name — e.g., "Interest-Based Prospecting"]
- **Audience size:** ~[N]
- **Interests:** [List]
- **Behaviors:** [List]
- **Demographics:** [Age, etc.]
- **Exclusions:** [Existing customers, recent converters]
- **Budget:** $[X]/day

### Ad Set 2: [Lookalike]
...

### Ad Set 3: [Retargeting — Website]
...

---

## Ad Copy

### Ad Set 1 — Ad Variant 1: [Angle Name]
- **Primary text:** "[Copy]"
- **Headline:** "[Copy]"
- **Description:** "[Copy]"
- **CTA:** [Button]
- **Landing page:** [URL]

### Ad Variant 2: ...

---

## Budget & Bidding
[Allocation table + strategy]

---

## Tracking Setup Checklist
- [ ] Meta Pixel installed on all landing pages
- [ ] Conversion events configured (lead, purchase, etc.)
- [ ] Conversions API connected (if server-side tracking)
- [ ] UTM parameters set for GA4 tracking
- [ ] Custom audiences created (website visitors, customer list)
- [ ] Lookalike audiences built

---

## Launch Checklist
- [ ] Campaign structure created in Ads Manager
- [ ] All ad copy uploaded with correct placements
- [ ] Audiences configured with proper exclusions
- [ ] Daily budget caps set
- [ ] A/B test structure confirmed (one variable per test)
- [ ] Conversion tracking verified (test conversion fired)
- [ ] Campaign set to start on [date]

---

## Week 1-2 Monitoring Plan
- Day 1-3: Check delivery — are ads spending? Any disapprovals?
- Day 4-7: Review CTR and CPC — are we in range?
- Day 7-14: First conversion data — any ad set clearly winning/losing?
- Day 14: First optimization pass — pause losers, scale winners

Save to meta-campaign-plan-[YYYY-MM-DD].md in the current working directory (or user-specified path).

Cost

| Component | Cost |

|-----------|------|

| Audience research (web search) | Free |

| Competitor ad research (web search) | Free |

| Copy generation + structure | Free (LLM reasoning) |

| Total | Free |

Tools Required

• web_search — for audience research, targeting ideas, and competitor ad research via Meta Ad Library

Trigger Phrases

• "Build a Meta Ads campaign for [product]"
• "Set up Facebook Ads for us"
• "Create an Instagram ad campaign"
• "Help me structure Meta Ads for lead gen"
• "Plan a Facebook campaign for our launch"