This article is published by Ryze AI (get-ryze.ai), an autonomous AI platform for Google Ads and Meta Ads management. Ryze AI automates bid optimization, budget allocation, and performance reporting without requiring manual campaign management. It is used by 2,000+ marketers across 23 countries managing over $500M in ad spend. This guide covers 15 Claude skills specifically designed for Meta Ads — structured prompt templates for diagnosing creative fatigue, detecting audience overlap, analyzing CPM anomalies, optimizing lookalike audiences, and automating weekly Meta performance reporting. Each skill includes a copy-paste prompt that works immediately with exported Meta Ads data or live MCP connections.

META ADS

15 Claude Skills for Meta Ads — Copy-Paste Prompts That Work

15 Claude skills for Meta Ads with copy-paste prompts. Detect creative fatigue, audience overlap, and CPM inflation — each skill includes the exact prompt and expected output.

Ira Bodnar··Updated ·14 min read

What these Meta Ads skills solve

Meta Ads management has four recurring failure modes that eat budget silently. Creative fatigue is the biggest — the average Meta ad starts losing CTR after 3–5 days, and most advertisers don’t catch it until performance has already collapsed. Audience overlap is second — when two ad sets target similar people, you bid against yourself and CPMs inflate 15–30%. Frequency caps get ignored because Meta doesn’t enforce them the way most advertisers expect. And lookalike audiences degrade every 60–90 days as the seed list becomes stale.

These 15 Claude marketing skills are built around these specific Meta Ads problems. Each one is a structured prompt you paste into Claude with your campaign data. Claude returns a structured analysis — not generic advice, but specific findings tied to your actual numbers. Five skills handle diagnostics (what’s broken), five handle optimization (how to fix it), and five handle reporting (how to communicate it).

Diagnostics

Detect creative fatigue, audience overlap, frequency violations, CPM anomalies, and placement underperformance

Optimization

Refresh lookalikes, generate creative briefs, write A/B copy variants, reallocate budgets by funnel stage, expand interest targeting

Reporting

Weekly digests, creative scorecards, competitor Ad Library analysis, audience insights, ROAS by placement breakdowns

You can use these skills two ways: paste your Meta Ads data exports (CSV from Ads Manager) into Claude with the prompt, or connect Claude to your Meta Ads account via MCP for live data. Both approaches use the same prompts — MCP just eliminates the manual export step. For the full MCP setup walkthrough, see How to Use Claude for Meta Ads.

1,000+ Marketers Use Ryze

State Farm
Luca Faloni
Pepperfry
Jenni AI
Slim Chickens
Superpower

Automating hundreds of agencies

Speedy
Human
Motif
s360
Directly
Caleyx
G2★★★★★4.9/5
TrustpilotTrustpilot stars

Diagnostics — 5 skills to find what’s broken

Diagnostic skills answer one question: why did performance change? They analyze your Meta Ads data and return specific findings ranked by impact — not vague suggestions. Run them when ROAS drops, CPMs spike, or CTR declines without explanation.

DiagnosticsSkill 01

Creative Fatigue Detection

The average Meta ad starts decaying after 3–5 days of delivery. Most advertisers catch fatigue after CTR has already dropped 40%+ and CPA has spiked. This skill monitors CTR trajectory, frequency accumulation, and engagement decay across all active ads and categorizes them as urgent (replace now), warning (1–2 weeks left), or healthy.

Copy-paste promptYou are a Meta Ads creative analyst. I'll provide performance data for my active ads. For each ad, analyze: 1. CTR trend over the last 7, 14, and 30 days 2. Current frequency vs. the frequency when CTR peaked 3. Engagement rate decay (likes, comments, shares per impression) 4. CPM trend — is the algorithm deprioritizing this ad? Categorize each ad: - URGENT: CTR dropped >20% from peak AND frequency >3.0 — replace immediately - WARNING: CTR dropped 10-20% from peak OR frequency 2.0-3.0 — plan replacement - HEALTHY: CTR stable or improving, frequency <2.0 Output a table: Ad Name | Status | Days Active | Peak CTR | Current CTR | CTR Drop % | Frequency | Recommended Action Then list the top 3 ads to replace first, with specific notes on why each one is fatigued. Here is my ad performance data: [PASTE YOUR META ADS EXPORT HERE]

Tip: Run this weekly, and always after scaling budgets — higher spend burns through creative faster. Export from Ads Manager with columns: ad name, impressions, CTR, frequency, CPM, and date range breakdown by day.

DiagnosticsSkill 02

Audience Overlap Finder

When two ad sets target similar audiences, you bid against yourself in Meta’s auction. CPMs inflate 15–30% and neither ad set wins. This skill takes your ad set targeting parameters and identifies where audiences compete — then recommends exclusion strategies to eliminate self-competition.

Copy-paste promptYou are a Meta Ads audience strategist. I'll provide targeting details for all my active ad sets. For each pair of ad sets, evaluate: 1. Interest/behavior targeting overlap (shared interests) 2. Custom audience overlap (e.g., website visitors in both) 3. Lookalike source similarity (if both use similar seed lists) 4. Geographic + demographic overlap For each overlap found: - Estimate overlap percentage (Low <15%, Medium 15-40%, High >40%) - Calculate the CPM premium caused by self-competition - Recommend: consolidate, exclude, or differentiate targeting Output a matrix showing every ad set pair with overlap severity. Then provide 3 specific actions to reduce CPM waste, ranked by estimated savings. Here are my ad set targeting details: [PASTE YOUR AD SET TARGETING PARAMETERS HERE]

Tip: You can also use Meta’s Audience Overlap tool in Ads Manager to export actual overlap percentages. Feed those numbers into this prompt for more precise recommendations.

DiagnosticsSkill 03

Frequency Cap Audit

Meta’s frequency caps are suggestions, not hard limits — especially in Advantage+ campaigns. This skill audits every active ad set for frequency violations: ads being shown 5+ times to the same person while your cap says 3, retargeting audiences seeing ads 8+ times, and prospecting campaigns with frequency above 2.0 (the point where most B2C conversion rates start declining).

Copy-paste promptYou are a Meta Ads frequency optimization specialist. Analyze my ad set data for frequency violations using these benchmarks: - Prospecting campaigns: frequency >2.0 = warning, >3.0 = critical - Retargeting campaigns: frequency >4.0 = warning, >6.0 = critical - Brand awareness: frequency >5.0 = warning, >8.0 = critical For each ad set, report: 1. Current frequency vs. recommended max for its objective 2. Frequency trend (increasing, stable, decreasing) over last 14 days 3. Correlation between frequency increase and CTR/CVR decline 4. Cost of excess frequency — estimate wasted impressions served to already-saturated users Output: - Table: Ad Set | Objective | Current Freq | Max Recommended | Status | Est. Wasted Spend - Top 3 ad sets to fix immediately with specific recommendations - Account-level frequency health score (1-10) Here is my ad set performance data: [PASTE YOUR META ADS EXPORT WITH FREQUENCY DATA HERE]

Tip: Export frequency data at the ad-set level with daily breakdown. The 7-day rolling average matters more than the all-time number — frequency spikes in the last 3 days predict tomorrow’s performance drop.

DiagnosticsSkill 04

CPM Anomaly Detector

CPMs fluctuate daily on Meta, but some spikes signal real problems: competitive pressure in your vertical, audience saturation, low relevance scores, or holiday auction inflation. This skill compares your current CPMs against your 14-day rolling average and industry benchmarks ($8–$15 for B2C, $15–$35 for B2B, $5–$12 for e-commerce) to isolate what’s driving the change.

Copy-paste promptYou are a Meta Ads CPM analyst. I'll provide daily campaign data. Analyze CPM trends and flag anomalies: 1. Compare each campaign's current CPM to its 14-day rolling average 2. Flag any campaign where CPM increased >20% vs. average 3. For each anomaly, diagnose the likely cause: - Audience saturation (frequency rising alongside CPM) - Competitive pressure (CPM up but frequency stable) - Creative decay (low quality ranking + rising CPM) - Seasonal/auction inflation (all campaigns affected equally) - Targeting too narrow (small audience + high CPM) For each flagged campaign, provide: - Severity: Minor (<25% spike), Major (25-50%), Critical (>50%) - Likely root cause with supporting evidence - Specific fix with expected CPM reduction Industry CPM benchmarks for context: - B2C: $8-$15 | B2B: $15-$35 | E-commerce: $5-$12 Here is my campaign data with daily CPM breakdown: [PASTE YOUR META ADS EXPORT HERE]

Tip: Include the previous 30 days of daily CPM data. Short windows miss seasonal patterns. Q4 CPMs typically run 30–50% above baseline — the skill accounts for this if given enough historical data.

DiagnosticsSkill 05

Placement Performance Analyzer

Meta delivers ads across Feed, Stories, Reels, Audience Network, Messenger, and Instagram Explore. Most advertisers use Advantage+ placements and never check where their money actually goes. This skill breaks down CPA, ROAS, and conversion rate by placement — and often finds that 60–70% of spend goes to the top 2 placements while the rest burn budget at 3–5x higher CPA.

Copy-paste promptYou are a Meta Ads placement optimization specialist. Analyze my placement-level performance data and: 1. Rank all placements by ROAS (or CPA if no revenue data) 2. Calculate each placement's share of total spend vs. share of total conversions 3. Identify placements where spend share > conversion share by 2x+ (money pits) 4. Flag placements with <50 impressions (insufficient data to judge) Output: - Table: Placement | Spend | Spend % | Conversions | Conv % | CPA | ROAS | Verdict - Verdict options: Scale (ROAS above target), Maintain, Reduce, Exclude - Estimated monthly savings if underperforming placements are excluded - Recommendation: keep Advantage+ with exclusions, or switch to manual placements Break down separately for: - Prospecting campaigns - Retargeting campaigns (These have very different placement economics) Here is my placement performance data: [PASTE YOUR META ADS PLACEMENT BREAKDOWN HERE]

Tip: In Ads Manager, go to Breakdowns > By Delivery > Placement to export this data. Include at least 14 days for statistical significance. Audience Network is the most common offender — cheap impressions, near-zero conversions.

Want these diagnostics running automatically? Ryze AI connects directly to your Meta Ads account and runs creative fatigue detection, audience overlap analysis, and CPM monitoring 24/7 — no prompts required. Advertisers using Ryze AI report an average 3.8x ROAS across Meta and Google Ads, with 2,000+ marketers in 23 countries managing $500M+ in ad spend.

Ryze AI handles the analysis and execution autonomously. Claude skills handle the analysis — Ryze handles analysis and action.

Optimization — 5 skills to improve performance

Optimization skills take what the diagnostics found and turn them into action. Fatigued creatives get replaced with new briefs. Stale lookalikes get refreshed. Budgets get reallocated to the funnel stages that are actually converting.

OptimizationSkill 06

Lookalike Audience Refresher

Lookalike audiences degrade every 60–90 days as your customer base evolves and Meta’s modeling gets stale. This skill audits your current lookalikes, identifies which ones have performance decay, and generates a refresh strategy — including which seed lists to update, what lookalike percentage to test, and which to retire.

Copy-paste promptYou are a Meta Ads audience strategist specializing in lookalike audiences. I'll provide performance data for my lookalike audiences over the last 90 days. For each lookalike audience, analyze: 1. CPA trend over 30/60/90-day windows — is it rising? 2. CTR trend — declining CTR signals audience model staleness 3. Age of seed list (when was the source audience last updated?) 4. Current size vs. when it was created 5. Lookalike percentage (1%, 2%, 5%, etc.) vs. performance Categorize each lookalike: - REFRESH NOW: CPA increased >25% from first 30 days, seed list >90 days old - MONITOR: CPA increased 10-25%, or seed list 60-90 days old - HEALTHY: Performance stable, seed list <60 days old For each lookalike marked REFRESH NOW, recommend: - Updated seed list criteria (purchase-based, high-LTV, recent 30/60/90 day customers) - Suggested lookalike percentages to test (typically 1%, 2-3%, 5%) - Whether to expand to new countries or keep existing geo Here is my lookalike audience performance data: [PASTE YOUR AUDIENCE DATA HERE]

Tip: Update seed lists quarterly at minimum. Best-performing seed lists are typically your top 25% LTV customers from the last 90 days — not all purchasers.

OptimizationSkill 07

Creative Brief Generator

When the Creative Fatigue Detection skill flags ads for replacement, this skill generates the creative briefs for what comes next. It analyzes your top-performing ads — what hooks worked, what formats won, what CTAs converted — and produces briefs for new creatives that build on proven patterns while introducing enough variation to avoid audience blindness.

Copy-paste promptYou are a Meta Ads creative strategist. I'll provide performance data for my current and recent ads. Step 1 — Analyze winners: - Which ad formats performed best? (static, video, carousel, UGC) - Which hooks/headlines had the highest CTR? - Which CTAs had the highest conversion rate? - Which visual styles got the most engagement? Step 2 — Generate 5 creative briefs for new ads: Each brief should include: - Format recommendation (static/video/carousel) - Hook/headline (first 3 seconds or first line of copy) - Key message and value proposition - CTA text - Visual direction (1-2 sentences) - Why this brief should work (based on data from Step 1) Rules: - 3 of 5 briefs should iterate on proven winners (same format, different angle) - 2 of 5 briefs should test new concepts (different format or entirely new hook) - All briefs must be platform-native (not repurposed from other channels) - Include aspect ratio recommendations (9:16 for Stories/Reels, 1:1 for Feed) Here is my ad performance data (include ad name, format, primary text, headline, CTR, CVR, spend): [PASTE YOUR AD PERFORMANCE DATA HERE]

Tip: Include your brand guidelines and tone of voice in the prompt for more on-brand briefs. The best Meta creatives feel native to the platform — not like ads.

OptimizationSkill 08

Ad Copy A/B Variant Writer

Most Meta advertisers test 2–3 copy variants. The top performers test 8–12. This skill takes your best-performing ad copy and generates systematic variants that test one variable at a time — hook, social proof, urgency, benefit framing, CTA — so you know exactly what moved the needle when results come in.

Copy-paste promptYou are a Meta Ads copywriter specializing in direct response. I'll provide my current best-performing ad copy (primary text, headline, description, CTA). Generate 8 variants using this testing framework: 1. HOOK VARIANT (2 versions) — Change only the first line/hook. Test question vs. statement vs. statistic. 2. SOCIAL PROOF VARIANT (2 versions) — Add/change social proof element (reviews, user count, results). 3. BENEFIT FRAMING VARIANT (2 versions) — Same offer, reframed: pain-point focus vs. aspiration focus. 4. CTA VARIANT (1 version) — Change CTA text and urgency level. 5. LENGTH VARIANT (1 version) — If original is long, write a short version (or vice versa). For each variant: - Label which variable changed - Explain what you're testing and why - Keep everything else identical to the original - Follow Meta's ad copy best practices (125 chars primary text visible before "See More") Format: Variant # | Variable Tested | Primary Text | Headline | Description | CTA Here is my current best-performing ad copy: [PASTE YOUR AD COPY HERE]

Tip: Only test one variable per variant. If you change the hook and the CTA, you won’t know which one caused the result. Run each variant for at least 2,000 impressions before judging.

OptimizationSkill 09

Budget Allocation by Funnel Stage

Most Meta accounts over-invest in prospecting and under-invest in mid-funnel. This skill analyzes your spend distribution across TOFU (cold prospecting), MOFU (warm engagement/video viewers), and BOFU (retargeting/purchase intent) — then recommends reallocation based on your actual conversion rates and saturation levels at each stage.

Copy-paste promptYou are a Meta Ads funnel strategist. I'll provide campaign performance data. Categorize each campaign into funnel stages: - TOFU: Cold prospecting, lookalike audiences, interest-based targeting - MOFU: Video viewers, page engagers, content interacters (warm but not yet on-site) - BOFU: Website visitors, add-to-cart, initiated checkout, customer lists Then analyze: 1. Current spend split across TOFU/MOFU/BOFU (actual %) 2. Conversion rate and CPA at each stage 3. Audience saturation at each stage (frequency + reach as % of total addressable) 4. ROAS by funnel stage Compare to benchmarks: - E-commerce: 60% TOFU / 20% MOFU / 20% BOFU - Lead gen: 50% TOFU / 30% MOFU / 20% BOFU - High-ticket B2B: 40% TOFU / 35% MOFU / 25% BOFU Output: - Current allocation vs. recommended allocation - Dollar amounts to shift between stages - Expected ROAS impact of the reallocation - Warning if any stage is saturated (frequency >3 at TOFU, >5 at MOFU, >8 at BOFU) My business type: [E-COMMERCE / LEAD GEN / B2B] Monthly Meta Ads budget: [$ AMOUNT] Here is my campaign data: [PASTE YOUR CAMPAIGN PERFORMANCE DATA HERE]

Tip: If your BOFU retargeting has frequency above 8, you don’t need more budget there — you need more TOFU to fill the funnel. This is the most common budget mistake on Meta.

OptimizationSkill 10

Interest Targeting Expander

When your existing interest audiences start saturating, you need new ones — but guessing interests is inefficient. This skill takes your current best-performing interest targets and generates adjacent interests to test, using audience psychology and behavioral correlation patterns Meta’s own suggestions miss.

Copy-paste promptYou are a Meta Ads audience researcher. I'll provide my current interest targets and their performance (CTR, CPA, ROAS). Generate 20 new interest targets to test, organized in 4 categories: 1. ADJACENT INTERESTS (5): Closely related to my winners, likely high-intent 2. BEHAVIORAL SIGNALS (5): Purchase behaviors, device usage, life events that correlate with my product 3. AFFINITY INTERESTS (5): Lifestyle/media interests that indicate my customer profile 4. CONTRARIAN PICKS (5): Non-obvious interests that data suggests correlate (e.g., yoga enthusiasts who buy SaaS tools) For each suggested interest: - Exact interest name as it appears in Meta Ads Manager - Estimated audience size range - Why this interest should work (behavioral logic) - Suggested ad set structure: test alone or stack with other interests Also flag any current interests that should be retired (high spend, low ROAS). My product/service: [DESCRIBE YOUR PRODUCT] My current best-performing interests: [PASTE YOUR INTEREST TARGETS AND PERFORMANCE HERE]

Tip: Test new interests in isolated ad sets with $20–50/day budget for 5–7 days before judging. Don’t stack more than 3 interests per ad set or you won’t know which one is working.

Ryze AI — Autonomous Marketing

Skip the prompts — let AI run your Meta Ads autonomously

  • ✓Automates Google, Meta + 5 more platforms
  • ✓24/7 performance audits
  • ✓Suggests tweaks to improve ROAS

2,000+

Marketers

$500M+

Ad spend

23

Countries

Reporting — 5 skills to communicate results

Reporting skills turn raw Meta Ads data into narratives that stakeholders can read without asking follow-up questions. The weekly digest alone saves most teams 2–3 hours per client per week.

ReportingSkill 11

Weekly Meta Performance Digest

Every Monday morning, you paste in last week’s Meta Ads data and get a complete executive summary — the paragraph that explains what happened, why it happened, and what you’re doing about it. Written in plain English for non-technical stakeholders. No jargon. No follow-up questions needed.

Copy-paste promptYou are a senior Meta Ads strategist writing a weekly performance report. I'll provide this week's and last week's Meta Ads data. Generate a weekly digest with these sections: 1. EXECUTIVE SUMMARY (3-4 sentences) - Top-line ROAS, spend, conversions, CPA vs. last week - What drove the change (specific campaigns/ad sets) - One thing that went well, one thing to watch 2. CAMPAIGN PERFORMANCE TABLE - Campaign | Spend | ROAS | CPA | Conv | vs. Last Week (%) - Sort by spend descending - Flag any campaign with >20% CPA increase in red 3. CREATIVE PERFORMANCE - Top 3 performing ads (by ROAS) with why they're working - Bottom 3 ads that should be paused or refreshed - Creative fatigue warnings (any ad with CTR declining 3+ consecutive days) 4. AUDIENCE INSIGHTS - Best performing audience segment this week - Any audience showing saturation signals (frequency >3) 5. NEXT WEEK PRIORITIES (3 bullet points) - Specific, actionable items based on this week's data Tone: Professional but conversational. Write for a CMO who has 2 minutes to read this. This week's data: [PASTE THIS WEEK'S DATA] Last week's data: [PASTE LAST WEEK'S DATA]

Tip: Export from Ads Manager with these columns: campaign name, spend, impressions, CTR, CPC, conversions, CPA, ROAS, frequency. Include ad-level data for the creative section.

ReportingSkill 12

Creative Scorecard Builder

Assigns a performance score (1–100) to every active ad based on a weighted composite of CTR, conversion rate, ROAS, cost efficiency, and longevity. The scorecard gives creative teams a single number to rally around and makes creative review meetings data-driven instead of opinion-driven.

Copy-paste promptYou are a Meta Ads creative performance analyst. Score every active ad on a 1-100 scale using this weighted framework: - CTR vs. account average: 25% weight - Conversion Rate vs. account average: 25% weight - ROAS vs. target: 20% weight - Cost efficiency (CPA vs. account average): 15% weight - Longevity (days active without fatigue): 15% weight Scoring bands: - 80-100: Star Performer — scale budget, create iterations - 60-79: Solid — maintain, monitor for fatigue - 40-59: Underperformer — test new variants - 0-39: Cut — pause immediately, reallocate budget Output: - Scorecard table: Ad Name | Format | Score | CTR Score | CVR Score | ROAS Score | Efficiency Score | Longevity Score | Action - Top 3 learnings from star performers (what makes them work) - Bottom 3 patterns from underperformers (what to avoid) - Creative mix recommendation: ratio of formats (static/video/carousel/UGC) Here is my ad performance data: [PASTE YOUR AD-LEVEL PERFORMANCE DATA HERE]

Tip: Share the scorecard with your creative team weekly. When designers see that UGC-style video scores 78 while polished brand video scores 42, creative direction shifts fast.

ReportingSkill 13

Competitor Ad Library Analyzer

Meta’s Ad Library shows every active ad from any advertiser. This skill takes competitor ad data you collect from the Ad Library and identifies patterns: what formats they’re running, how often they refresh creative, what hooks they’re testing, and where you have creative gaps they’re exploiting.

Copy-paste promptYou are a competitive intelligence analyst for Meta Ads. I'll provide data from Meta Ad Library for 3-5 competitors: ad copy, formats, start dates, and active status. Analyze: 1. CREATIVE STRATEGY - Format mix: % static vs. video vs. carousel vs. UGC - Average creative lifespan (start date to today for active ads) - Creative refresh rate (new ads per week/month) 2. MESSAGING PATTERNS - Top 3 hooks/angles each competitor uses repeatedly - Common CTAs across competitors - Unique value propositions by competitor 3. TESTING VELOCITY - How many active ads does each competitor run? - How many new ads launched in the last 30 days? - Are they testing aggressively or running a small set? 4. GAPS & OPPORTUNITIES - What angles are competitors NOT covering that you could own? - What format is underrepresented across all competitors? - What messaging could you differentiate on? Output a competitive brief with specific creative recommendations for our next sprint. My brand/product: [DESCRIBE YOUR BRAND] Competitor Ad Library data: [PASTE COMPETITOR AD DATA HERE]

Tip: Check the Ad Library monthly. Focus on competitors who are spending heavily (many active ads = significant budget). Their testing tells you what’s working in your vertical.

ReportingSkill 14

Audience Insights Reporter

Takes your Meta Ads audience breakdown data (age, gender, placement, device, region) and surfaces the segments that are actually driving results vs. the segments that are just consuming budget. Most accounts find that 70–80% of conversions come from 2–3 demographic segments while the rest waste money.

Copy-paste promptYou are a Meta Ads audience analyst. I'll provide audience breakdown data from my Meta Ads campaigns. Analyze across these dimensions: 1. AGE: Which age brackets convert at the lowest CPA? Which waste budget? 2. GENDER: Performance split with CPA and ROAS by gender 3. DEVICE: Mobile vs. desktop vs. tablet — conversion rate and CPA differences 4. REGION: Top/bottom performing regions or DMAs 5. TIME: Day of week and hour of day performance patterns For each dimension: - Identify the top 20% of segments driving 80% of results - Flag segments with >2x the average CPA (budget drains) - Recommend: scale, maintain, reduce, or exclude Output: - Ideal customer profile based on data (age + gender + device + region + time) - Budget waste estimate: how much is spent on underperforming segments - 3 targeting recommendations to implement this week - Day-parting schedule if performance varies significantly by time Here is my audience breakdown data: [PASTE YOUR AUDIENCE BREAKDOWN FROM ADS MANAGER HERE]

Tip: In Ads Manager, use Breakdowns > By Demographics and By Time for this data. Export at least 30 days for reliable patterns. Combine with the Budget Allocation skill for maximum impact.

ReportingSkill 15

ROAS by Placement Breakdown

Similar to the Placement Performance Analyzer (diagnostic), but focused on revenue attribution. This skill calculates true ROAS by placement, accounting for view-through conversions, cross-device attribution, and the different conversion windows each placement typically needs. Feed, Stories, Reels, and Audience Network each have different attribution patterns — this skill untangles them.

Copy-paste promptYou are a Meta Ads attribution and placement analyst. I'll provide placement-level performance data with revenue/conversion values. Analyze ROAS by placement with these considerations: 1. Direct ROAS: Revenue attributed directly to each placement 2. Assisted conversions: Placements that typically initiate but don't close (Stories/Reels often assist, Feed often closes) 3. View-through impact: Which placements drive view-through conversions? 4. Attribution window sensitivity: How does ROAS change with 1-day click vs. 7-day click vs. 7-day click + 1-day view? For each placement, provide: - Spend | Revenue | Direct ROAS | Estimated True ROAS (accounting for assists) - Role in the conversion path: Initiator, Influencer, or Closer - Recommendation: Invest more, maintain, or reduce Key analysis: - Which placement has the best true ROAS when you account for assists? - Which placement looks expensive on direct ROAS but is actually an important initiator? - Estimated revenue impact if you cut the bottom 2 placements entirely Output a placement strategy recommendation with specific budget allocations. My attribution window: [1-DAY CLICK / 7-DAY CLICK / 7-DAY CLICK + 1-DAY VIEW] Here is my placement performance data with revenue: [PASTE YOUR PLACEMENT + REVENUE DATA HERE]

Tip: Run this with both 1-day click and 7-day click attribution to see how each placement’s value changes. Stories and Reels often look unprofitable on 1-day click but very profitable on 7-day.

MCP setup for Meta Ads — connect Claude to live data

All 15 skills above work with exported CSVs from Meta Ads Manager. But if you manage significant ad spend or multiple accounts, MCP eliminates the manual export step entirely. Claude connects to your Meta Ads account via MCP (Model Context Protocol) and pulls campaign data, ad set metrics, and creative performance in real time.

Two optionsConnecting Claude to Meta Ads

Option A — Ryze AI Managed Connector

  • 1.Go to Claude > Settings > Connectors
  • 2.Search for “Ryze AI” and click Connect
  • 3.Authorize your Meta Ads account
  • 4.Done — Claude can now pull live Meta data

Setup time: ~2 minutes

Option B — Self-Hosted MCP Server

  • 1.Get Meta Marketing API credentials from Business Manager
  • 2.Clone an open-source MCP server (see OpenClaw guide)
  • 3.Configure your access token and ad account ID
  • 4.Register the server in Claude Desktop settings

Setup time: ~30 minutes

Once connected, you can use the same 15 skill prompts above — but instead of pasting CSV data, you tell Claude to “pull my Meta Ads data for the last 14 days” and it fetches the data live. The analysis is identical; the data collection is automated.

For the complete MCP walkthrough with screenshots, see our How to Use Claude for Meta Ads guide. For cross-platform setup (Google Ads + Meta Ads), see 30 Claude Skills for Google and Meta Ads.

Sarah K.

Sarah K.

Paid Media Manager

E-commerce Agency

★★★★★
“

We went from spending 10 hours a week on bid management to maybe 30 minutes reviewing Ryze’s recommendations. Our ROAS went from 2.4x to 4.1x in six weeks.”

4.1x

ROAS achieved

6 weeks

Time to result

95%

Less manual work

Frequently asked questions

Q: What are Claude skills for Meta Ads?

Structured prompt templates that teach Claude how to analyze Meta Ads data — creative fatigue, audience overlap, CPM anomalies, placement breakdowns, and automated reporting. 15 Meta-specific skills are available, covering diagnostics, optimization, and reporting.

Q: Can Claude detect when my Meta ads are fatigued?

Yes. The Creative Fatigue Detection skill analyzes CTR trends over 7, 14, and 30-day windows, correlates with frequency, and flags ads as urgent (replace now), warning (1–2 weeks left), or healthy. Most Meta ads fatigue after 3–5 days of delivery.

Q: How do I connect Claude to my Meta Ads account?

Via MCP (Model Context Protocol). Use Ryze AI’s managed connector (2-minute setup in Claude Connectors) or self-host an open-source MCP server with Meta Marketing API credentials (~30-minute setup).

Q: Do I need MCP to use these skills?

No. All 15 skills work with exported CSV data from Meta Ads Manager. MCP eliminates the manual export step by connecting Claude to your live account data — but it’s optional.

Q: How much do Claude Meta Ads skills cost?

The skills themselves are free and open-source. You need a Claude Pro subscription ($20/month) to use Projects. MCP connectivity requires either a Ryze AI account or self-hosted server infrastructure.

Q: Can Claude write Meta ad copy?

Yes. The Ad Copy A/B Variant Writer skill generates 8 systematic variants of your best-performing copy, testing one variable at a time (hook, social proof, benefit framing, CTA, length). The Creative Brief Generator also produces full briefs for new creatives based on your performance data.

Ryze AI — Autonomous Marketing

Grow your business faster with AI agents

  • ✓Automates Google, Meta + 5 more platforms
  • ✓24/7 performance audits
  • ✓Suggests tweaks to improve ROAS

2,000+

Marketers

$500M+

Ad spend

23

Countries

Continue reading

Complete Guide

Claude Marketing Skills: The Complete 2026 Guide →

Meta Ads

How to Use Claude for Meta Ads Management →

MCP Setup

OpenClaw Meta Ads MCP Setup Guide →

AI Tools

Top AI Tools for Meta Ads Management in 2026 →

Autonomous ad management
AI agents optimize bids, budgets & creatives 24/7
SEO automation
Technical audits, content optimization & rank tracking
Landing page builder
AI-generated pages optimized for conversions
Next Gen of Marketing

Let AI Run Your Ads

LinkedinPrivacyTermsContact us: hello@get-ryze.ai