agentleFS
gomarble-aigomarble-ai/marketing-agent/skills/gomarble-skills-creative-strategists/SKILL.md

gomarble-skills-creative-strategists

Skill15 starsChanged 39 days agoMITView on GitHub

Use for creative strategy frameworks across platforms: ad fatigue detection, hook and hold analysis, creative performance ranking, design and copy audits, A/B test interpretation, audience-to-creative mapping, creative lift forecasts, messaging theme classification, and a weekly creative brief. Pairs with analyse-creative and brief-creative for live GoMarble data.

The file

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Do I have 100+ conversions per variant for statistical significance?

    line 226

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

GoMarble Skills for Creative Strategists

9 Core Skills + Weekly Brief for analyzing and optimizing creative performance across platforms.

Read-and-recommend only. These skills never change your ad accounts. The model reads data and produces recommendations in plain language; you apply them manually.


The 9 Skills

1. Ad Fatigue Detector

What it does: Catches frequency creep and CTR decay before your cost per conversion spikes.

How it works:

  • Tracks CTR trend over 7-14 days per creative
  • Flags when CTR drops >15% from week 1 baseline
  • Correlates with frequency increase (>2.0 average reach per user)
  • Predicts conversion lift opportunity from refreshing

Minimum data needed: Ad-level metrics for last 14 days (CPM, impressions, CTR, frequency)


2. Hook & Hold Analyzer

What it does: Pinpoints which moments lose viewers and what hooks work across platforms.

How it works:

  • Calculates Hook Rate (% who watch first 3 seconds of video)
  • Calculates Hold Rate (% who finish middle section)
  • Compares against account average
  • Identifies common hook patterns in top 20% performers

Minimum data needed: Video plays, 3-second video plays, ThruPlay count per ad


3. Creative Performance Ranker

What it does: Ranks by CTR, CPC, and conversion rate. Shows which creative angles win.

How it works:

  • Pareto analysis — top 20% of creatives drive 80%+ of value
  • Calculates efficiency ratio (CTR ÷ CPC) per creative
  • Groups by angle (problem-first, social proof, urgency, scarcity)
  • Flags which angles outperform category average

Minimum data needed: Ad-level CTR, CPC, conversion count, campaign objective


4. Design & Copy Auditor

What it does: Evaluates ad copy, CTA clarity, and visual hierarchy against best practices.

How it works:

  • Scores headline length (optimal 3-8 words)
  • Checks CTA presence and specificity (vs vague CTAs like "Learn More")
  • Rates image/video visual hierarchy (primary focal point clear?)
  • Flags copy-to-visual mismatch

Minimum data needed: Ad creative (image/video + copy text), CTR, conversion rate


5. A/B Test Interpreter

What it does: Translates CTR and conversion gaps into actionable creative insights.

How it works:

  • Calculates statistical significance (need 100+ conversions each arm minimum)
  • Identifies winning variable (is it the hook? the CTA? the visual?)
  • Quantifies lift per variable
  • Recommends which elements to keep vs iterate

Minimum data needed: Two ad variants with CTR, impressions, conversions, spend each


6. Audience-to-Creative Mapper

What it does: Shows which creative resonates with which audience segment and why.

How it works:

  • Segments performance by age, geography, device (if available)
  • Identifies angle preferences per segment (e.g., younger audiences prefer urgency; older prefer authority)
  • Flags creative mismatches (creative designed for one demo, performing poorly there)
  • Recommends audience-specific creative rotation

Minimum data needed: Ad performance segmented by demographic (age/gender/location)


7. Creative Lift Forecaster

What it does: Estimates how a new creative angle could move ROAS based on peer benchmarks.

How it works:

  • Compares current account's top angle ROAS vs second-best
  • Checks peer benchmark for that angle (via industry data)
  • Models revenue impact of replacing bottom 20% with new angle
  • Accounts for learning phase loss

Minimum data needed: Current ad ROAS, spend, conversions; benchmark data for angle type


8. Messaging Theme Classifier

What it does: Breaks down ad copy themes (urgency, social proof, scarcity) and ranks by conversion.

How it works:

  • Categorizes ad copy by primary emotional trigger
  • Ranks themes by conversion rate and CPA
  • Identifies over-reliance on one theme (fatigue risk)
  • Recommends theme rotation schedule

Minimum data needed: Ad copy text, conversion count, spend per ad


9. Weekly Creative Brief

What it does: Top performers, fatigue alerts, test recommendations. One report, all platforms.

Includes:

  • Top 3 creatives by ROAS this week
  • Ads showing fatigue (CTR down >15%)
  • Hook rate / Hold rate averages
  • Theme distribution (are you over-reliant on one message?)
  • A/B test recommendations for next week
  • Platform comparison (which format + audience wins where)

Minimum data needed: Full ad account export for last 7 days


How to Use These Skills

  1. Identify what you need — fatigue check? design audit? angle testing?
  2. Gather the data — ad-level metrics + creative content
  3. Run the relevant skill — get diagnosis + recommendations
  4. Apply manually — pause underperformers, test new angles, rotate themes

Data Requirements by Source

SkillMCP (GoMarble/Meta)CSV ExportManual
Ad Fatigue Detector✅✅⚠️ spot-check only
Hook & Hold Analyzer✅✅ (if video metrics enabled)❌
Creative Performance Ranker✅✅✅
Design & Copy Auditor✅ (via creative endpoint)❌ (content not in CSV)✅ (upload image/video)
A/B Test Interpreter✅✅⚠️
Audience-to-Creative Mapper✅⚠️ (needs breakdown export)⚠️
Creative Lift Forecaster✅✅⚠️
Messaging Theme Classifier✅✅✅
Weekly Creative Brief✅✅ (7-day export)❌

✅ = works perfectly | ⚠️ = requires manual input or separate export | ❌ = not possible from this source


Key Metrics Reference

MetricDefinitionGood Benchmark
Hook Rate (video)% of viewers who watch first 3 seconds≥40%
Hold Rate (video)% of viewers who finish the middle sectionAccount average + 5%
CTRClick-through rate≥ 1.25% (depends on platform)
FrequencyAvg times same user sees ad1.5–2.0 (above = fatigue risk)
ROASRevenue per dollar spent≥3.0 for profitable
CPACost per acquisition≤ 40% of customer LTV

Output Format for Recommendations

### [Skill Name]: [Creative Name]

Finding:
[1–2 sentences describing the issue or opportunity]

Data:
| Metric | Value | Benchmark | Status |
|--------|-------|-----------|--------|
| [Metric] | X | Y | [Good/Poor] |

Recommendation:
1. [Specific action with rationale]
2. [Specific action with rationale]

When to Run Each Skill

SituationUse SkillTimeline
CTR dropping on top adAd Fatigue DetectorWeekly
Video not performingHook & Hold AnalyzerAfter 5K plays minimum
Too many similar adsCreative Performance Ranker + Messaging Theme ClassifierBiweekly
Testing new angleA/B Test InterpreterAfter 100 conversions per arm
Expanding to new regionAudience-to-Creative MapperAfter change
Planning sprintWeekly Creative BriefEvery Monday

Common Guardrails

Before recommending a pause or rotation:

  • ✅ Confirm fatigue signal is real (CTR down >15%, frequency >2.0)
  • ✅ Confirm replacement creative is ready to launch (don't create gap)
  • ✅ Keep top 20% performers running while testing new angles
  • ✅ Never pause an ad mid-learning phase (<50 conversions)

Before recommending a new angle:

  • ✅ Confirm current angle is mature (30+ days data)
  • ✅ Ensure spend budget for testing (usually 10–20% of total)
  • ✅ Have the creative ready before recommending (not hypothetical)

Questions to Ask Yourself

  • Am I testing one variable at a time? (angle, audience, placement, format)
  • Do I have 100+ conversions per variant for statistical significance?
  • Is my top performer still running while I test? (don't stop the winner)
  • Am I rotating themes or stuck on urgency?
  • What does my Hold Rate tell me vs my CTR?

Next Steps

  1. Pull your ad data (last 7–30 days)
  2. Choose 2–3 skills to run this week
  3. Apply recommendations to 10–20% of spend first
  4. Measure results, iterate

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.