Creative Audit

Creative Audit

Categorize all active ads by format, concept, and angle. Cross-reference with performance data to identify creative gaps and build a testing roadmap. Complements creative-fatigue (which focuses on lifecycle) by focusing on creative diversity and mix.

Account Context

Read ${CLAUDE_PLUGIN_ROOT}/profile/account-profile.md at the start of every run. If it exists:

• Use known account IDs.
• Apply KPI targets for performance benchmarking.
• Note active creative tests — highlight them in the matrix.
• Check watch list for creative-related items from prior sessions. If it doesn't exist, fall back to discovery and suggest running platform-setup.

Classification Dimensions

Format

Concept

Angle

Workflow

Phase 1: Pull creative data

1. Use mcp__meta-ads__list_ads to get all active ads across campaigns.
2. Use mcp__meta-ads__list_creatives to get creative asset details.
3. Use mcp__meta-ads__preview_ad for ads where format/concept is ambiguous.
4. Use mcp__meta-ads__get_creative_performance for performance metrics.
5. Use mcp__meta-ads__analyze_account_creatives for account-level creative health.

Parallelization: If the account has more than 5 active ad sets, consider spawning the creative-analyst subagent (see agents/creative-analyst.md) for parallel analysis.

Phase 2: Categorize each ad

For each active ad:

1. Identify Format from ad metadata (image, video, carousel, etc.).
2. Classify Concept from creative content (may require preview_ad for context).
3. Classify Angle from primary text and headline.
4. Record performance metrics alongside classification.

Phase 3: Build creative matrix

Create a Format x Concept matrix showing:

• Count of active ads in each cell.
• Average performance (CTR, CPA) per cell.
• Highlight empty cells as gaps.
• Highlight high-performing cells as winning combinations.

Phase 4: Gap analysis

Identify missing creative combinations:

• Formats never tested (e.g., no video, no carousel).
• Concepts never tested (e.g., no UGC, no testimonial).
• Angles never tested (e.g., no pain-led, no curiosity).
• Winning format + losing concept combinations (format works, need new concepts).

Phase 5: Testing roadmap

Prioritize gaps by:

1. Format gaps with strong cross-industry benchmarks (e.g., video typically 20-30% lower CPA than static).
2. Concept gaps where competitors are active (if competitor-research data available).
3. Angle variety — low-risk tests to diversify messaging.

Output Format

## Creative Audit - [Date]

### Account: [Name] (act_XXXXXXXXX)

### Active Creative Summary
| Metric | Count |
|--------|------:|
| Total Active Ads | |
| Unique Formats | |
| Unique Concepts | |
| Avg Ads per Ad Set | |

### Creative Matrix: Format x Concept
| | Testimonial | UGC | Product Demo | Lifestyle | Pain Point | Social Proof | Offer | Educational |
|---|---:|---:|---:|---:|---:|---:|---:|---:|
| Static Image | | | | | | | | |
| Video (Short) | | | | | | | | |
| Video (Long) | | | | | | | | |
| Carousel | | | | | | | | |
| DPA | | | | | | | | |

*Cell = count of active ads. Green = above-avg CPA. Red = below-avg CPA. Empty = gap.*

### Performance Heatmap: Avg CPA by Format x Concept
| | Testimonial | UGC | Product Demo | Lifestyle | Pain Point | Social Proof | Offer |
|---|---:|---:|---:|---:|---:|---:|---:|
| Static Image | $ | $ | $ | $ | $ | $ | $ |
| Video (Short) | $ | $ | $ | $ | $ | $ | $ |
| Carousel | $ | $ | $ | $ | $ | $ | $ |

### Top Performing Combinations
| Rank | Format | Concept | Angle | Avg CPA | Avg CTR | Active Ads | 30d Spend |
|:-----|--------|---------|-------|--------:|--------:|-----------:|----------:|
| 1 | | | | | | | |

### Creative Gaps (Missing or Undertested)
| Gap | Type | Why Test | Priority | Benchmark CPA |
|-----|------|----------|:---------|:--------------|
| No video ads | Format | Video typically 20-30% lower CPA | HIGH | |
| No UGC | Concept | Fastest-growing format on Meta | HIGH | |
| No testimonials | Concept | Highest trust signal | MEDIUM | |
| Single angle across all ads | Angle | Message fatigue risk | MEDIUM | |

### Testing Roadmap
| Priority | Test | Format | Concept | Angle | Budget | Duration | Success Metric |
|:---------|------|--------|---------|-------|-------:|----------|:---------------|
| 1 | | | | | $ | 7-14d | CPA < $X |
| 2 | | | | | $ | 7-14d | CPA < $X |
| 3 | | | | | $ | 7-14d | CTR > X% |

### Implementation
For each test:
- Meta Ads Manager > Campaign "[Name]" > Ad Set "[Name]" > Ads > Create Ad
- Use existing ad set targeting — isolate creative as the variable.
- Run for minimum 7 days or 50 conversions before evaluating.

Guardrails

• Read-only: This skill produces analysis only. Testing roadmap actions include UI paths.
• Classification limits: AI classification of concept/angle from ad preview is approximate. Flag confidence level when ambiguous.
• Minimum data: Require 7+ days and $50+ spend per ad for performance comparisons. Newer ads go into "Insufficient Data" bucket.
• Cross-platform context: If user runs ads on Google/Bing too, note that creative learnings may transfer (video hooks, messaging angles) but format performance is platform-specific.

Profile Maintenance

After completing analysis, if ${CLAUDE_PLUGIN_ROOT}/profile/account-profile.md exists:

1. Note any creative gaps identified in Watch List.
2. If user commits to testing roadmap items, add to Active Tests.
3. Update "Last updated" date. Present proposed profile changes to the user before writing.

References

• references/benchmarks.md — 2026 Meta benchmarks by format and objective
• references/ui-paths.md — Meta Ads Manager UI paths for creating ads