Meta Ads Diagnostics

Comprehensive diagnostic toolkit for Meta (Facebook/Instagram) advertising campaigns. Covers Learning Phase management, creative fatigue detection, audience overlap resolution, Advantage+ Shopping optimization, and platform-specific metric analysis.

Capabilities

1. Learning Phase Diagnosis - Identify why campaigns are stuck, how to exit faster, and when resets occur
2. Breakdown Effect Analysis - Detect when granular breakdowns distort performance signals
3. Audience Overlap Detection - Find and resolve overlapping audiences cannibalizing performance
4. Creative Fatigue Signals - Measure frequency vs CTR decay and predict creative lifespan
5. Advantage+ vs Manual Comparison - Determine when ASC outperforms manual campaign structures
6. Meta-Specific Metrics Analysis - Interpret CPM, ThruPlay, Hook Rate, Hold Rate, and Quality Ranking

Meta-Specific Metrics Reference

Workflows

Workflow 1: Learning Phase Diagnosis

What is Learning Phase? Meta's algorithm needs approximately 50 optimization events per ad set per week to stabilize delivery. During Learning Phase, performance is volatile and CPA is typically 20-50% higher.

Check current status:

1. Navigate to Ads Manager → Columns → Delivery
2. Look for "Learning" or "Learning Limited" in the Delivery column

Learning Phase resets when you:

How to exit Learning Phase faster:

1. Consolidate ad sets — Fewer ad sets = more events per ad set
2. Use broader audiences — 2M+ audience size recommended
3. Set appropriate budgets — Daily budget ≥ 10× target CPA
4. Use the right optimization event — If Purchase events are too rare, optimize for Add to Cart first
5. Don't touch it — No edits for 7 days after launch

Learning Limited diagnosis:

• Ad set got < 50 optimization events in 7 days
• Actions: Broaden audience, increase budget, switch to higher-volume event, consolidate

Workflow 2: Creative Fatigue Detection

Step 1: Pull frequency and CTR data over time

Export daily metrics for the past 30 days:

• Date, Impressions, Reach, Frequency, CTR, CPC, CPM, Conversions

Step 2: Plot the decay curve

Fatigue typically follows this pattern:

Day 1-3:   CTR peaks (novelty effect)
Day 4-10:  CTR stable (performing period)
Day 11-20: CTR declining 10-30% (early fatigue)
Day 21-30: CTR declining 30-60% (severe fatigue)

Step 3: Calculate fatigue score

Fatigue Score = (CTR_week1 - CTR_current) / CTR_week1 × 100

Interpretation:
  0-15%:  Healthy — no action needed
  15-30%: Early fatigue — prepare refresh
  30-50%: Active fatigue — rotate creatives now
  50%+:   Severe fatigue — pause and replace

Step 4: Check frequency correlation

Fatigue Frequency Threshold:
  Prospecting campaigns: 2.5-3.0 frequency
  Retargeting campaigns: 4.0-6.0 frequency
  High AOV products: 5.0-8.0 frequency (more consideration needed)

Step 5: Creative refresh strategy

Workflow 3: Audience Overlap Analysis

Step 1: Check overlap in Ads Manager

1. Go to Audiences → Select 2-5 audiences
2. Click "…" → Show Audience Overlap
3. Review overlap percentage

Step 2: Interpret overlap levels

Step 3: Symptoms of audience overlap cannibalization

• Rising CPMs across both ad sets (auction self-competition)
• Frequency increasing faster than expected
• Delivery becoming "Learning Limited" on both
• One ad set consistently outspending the other

Step 4: Resolution strategies

1. Exclusion approach: Exclude Audience B from Ad Set A and vice versa
2. Consolidation: Merge overlapping audiences into one broader ad set
3. Advantage+ Audience: Let Meta's algorithm find the best segments
4. Campaign Budget Optimization: Use CBO to let Meta allocate between overlapping ad sets

Workflow 4: Advantage+ Shopping (ASC) vs Manual Comparison

When ASC outperforms manual:

• E-commerce with product catalog
• 100+ conversions per week
• Broad targeting (minimal exclusions needed)
• Simple conversion goals (Purchase, Add to Cart)

When manual outperforms ASC:

• B2B lead generation
• Narrow/niche audiences required
• Complex exclusion lists needed
• Low conversion volume (< 50/week)
• Need for creative-level budget control

ASC Optimization checklist:

1. ☐ Set existing customer budget cap (recommended: 20-30%)
2. ☐ Upload customer list for existing customer identification
3. ☐ Include 5-10 diverse creatives (mix of formats)
4. ☐ Set appropriate daily budget (≥ 10× target CPA)
5. ☐ Allow 7-day learning period before evaluation
6. ☐ Use Advantage+ Creative features (text, media enhancements)

Key ASC metrics to watch:

• New vs returning customer split (keep new > 70%)
• Incremental ROAS vs blended ROAS
• CPM trends over time (rising CPM in ASC = audience saturation)

Workflow 5: Breakdown Effect Analysis

What is the Breakdown Effect? When you segment campaign data by placement, platform, age, gender, or other dimensions, the per-segment results can be misleading because Meta optimizes holistically.

Example:

• Campaign total CPA: $25 ✓
• CPA breakdown by placement:
  • Feed: $22
  • Stories: $35
  • Reels: $45
  • Audience Network: $80

Wrong conclusion: "Audience Network is terrible, turn it off" Right analysis: Audience Network may be serving low-cost impressions that assist Feed conversions

How to properly evaluate breakdowns:

1. Never optimize based on breakdown data alone
2. If you must test placement exclusion, use a holdout test (see incrementality-testing skill)
3. Use lift studies to measure true incremental impact
4. Allow Meta's algorithm 14+ days with all placements before drawing conclusions

Diagnostic Decision Tree

Campaign underperforming?
├── Check: Is it in Learning Phase?
│   ├── Yes → Don't touch for 7 days. Budget ≥ 10× CPA?
│   └── No → Continue diagnosis
├── Check: Frequency > 3.0?
│   ├── Yes → Creative fatigue. Check CTR decay curve.
│   └── No → Continue diagnosis
├── Check: CPM rising while CTR stable?
│   ├── Yes → Audience saturation or competition spike. Broaden audience.
│   └── No → Continue diagnosis
├── Check: CTR good but CVR dropping?
│   ├── Yes → Landing page issue or audience quality degradation.
│   └── No → Continue diagnosis
├── Check: Quality Ranking below average?
│   ├── Yes → Creative quality issue. Test new formats.
│   └── No → Continue diagnosis
└── Check: Conversion Ranking below average?
    ├── Yes → Post-click experience issue. Audit landing page.
    └── No → May be attribution/tracking issue. Check pixel health.

Reference Data

CPM Benchmarks by Industry (Meta Ads, 2025-2026)

Video Performance Benchmarks

Ad Format Performance Comparison

Examples

Example 1: Diagnosing a CPM Spike

Scenario: CPMs jumped from $12 to $22 over 2 weeks.

Diagnosis steps:

1. Check auction competition: Is it seasonal (Q4, Black Friday)?
2. Check frequency: If > 3.5, audience is saturated
3. Check audience overlap: Multiple ad sets competing?
4. Check quality rankings: Did they drop?
5. Check bid strategy: Bid cap too aggressive?

Finding: Frequency hit 4.2, quality ranking dropped to "Below Average" Fix: Rotate 3 new creatives, expand lookalike from 1% to 3%, and consolidate 4 ad sets into 2.

Example 2: Learning Limited Resolution

Scenario: 3 of 5 ad sets stuck in "Learning Limited" for 2+ weeks.

Analysis:

• Daily budget: $30/ad set
• Target CPA: $45
• Weekly optimization events per ad set: 8-12 (need 50)

Fix:

1. Consolidate 5 ad sets → 2 ad sets (each gets 5× the events)
2. Increase daily budget to $100/ad set (allows 50 events at $14 CPA margin)
3. Switch from Purchase to Add to Cart optimization (3× more events)
4. Wait 7 days without changes

Example 3: ASC vs Manual A/B Test

Setup:

• ASC campaign: $200/day, 8 creatives, broad targeting
• Manual campaign: $200/day, 3 ad sets (LAL 1%, Interest, Broad), 4 creatives each

Results after 14 days:

Conclusion: ASC wins on efficiency. Manual wins slightly on new customer acquisition. Recommendation: Run ASC as primary, keep one manual ad set for prospecting control.

About this skill

Part of the Synter free skills collection — open-source agent skills for advertising, PPC, and marketing automation.

Want this skill (and 40+ more) running on autopilot against your live ad accounts? Try Synter — AI Agent Media Buyers that connect to Google, Meta, LinkedIn, TikTok, Reddit, Amazon, and 7+ more platforms.

• 🌐 syntermedia.ai
• 📚 All free skills
• 💬 Built by @syntermedia — questions? Open an issue.