Meta Ads Creative Fatigue

[Didoo AI] Analyzes creative fatigue signals across Meta Ads campaigns. Use when reviewing declining CTR or ROAS, planning creative rotation schedules, or ma...

Security Scan

The skill's requested data and runtime instructions are coherent with its stated purpose (analyzing Meta Ads creative fatigue) and it only asks for the expected Meta credentials; nothing appears disproportionate or unrelated.

Purpose & Capability

Name and description match the requirements: it needs access to Meta Ads data and declares META_ACCESS_TOKEN and META_AD_ACCOUNT_ID. Both env vars are appropriate and sufficient for read-only ad performance queries.

Instruction Scope

SKILL.md confines actions to fetching ads and insights, computing fatigue metrics, classifying creatives, and storing a small set of session-context keys. It does not instruct reading arbitrary files or unrelated credentials. Note: the doc uses abstract calls (get_ads_list, get_ads_insights) rather than concrete Graph API endpoints or error/rate-limit handling; that's an operational gap but not a security incoherence.

Install Mechanism

No install spec or code files are present (instruction-only), so nothing will be written to disk or automatically installed by the skill.

Credentials

Only two env vars are required (META_ACCESS_TOKEN and META_AD_ACCOUNT_ID), which are expected for this task. The token is sensitive — the SKILL.md correctly references ads_read scope; users should avoid providing overly-broad or long-lived tokens.

Persistence & Privilege

always: false and default autonomous invocation are used (expected). The skill writes a small set of keys into session context (rotation_pipeline, fatigue_level, days_until_death) so downstream skills (meta-ads-recommendation, meta-ads-builder) can consume them. This is standard but you should confirm those downstream skills are trusted and that session context will not be logged or exported to third parties.

Assessment

This skill looks coherent for analyzing Meta Ads creative fatigue. Before installing: 1) Provide a least-privilege Meta token (ads_read) and prefer short-lived or app-scoped credentials rather than long-lived user tokens. 2) Confirm that the agent environment will keep session context private (the skill stores per-creative fatigue keys) and that any downstream skills that read those keys are trustworthy. 3) Be aware the SKILL.md uses abstract API actions instead of explicit Graph API calls — verify the agent implementation will respect rate limits and error handling. If you need stronger assurance, ask the publisher for details on the exact API calls performed and how tokens are used/rotated.

Required Credentials

Credential	Where to Get	Used For	OAuth Scope
META_ACCESS_TOKEN	Meta Developer Console → Graph API Explorer → Generate Token	Fetching ad performance and frequency data	ads_read (read-only)
META_AD_ACCOUNT_ID	Ads Manager URL: adsmanager.facebook.com/act_XXXXXXXXX	Identifying which account to query	—

When to Use

When creative has been running for 7+ days and performance is declining — or proactively when managing a portfolio of creatives. Use before scheduling creative refresh.

---

Requirements

• account_id (required)
• campaign_id or ad_ids (optional — defaults to all active ads in account)
• date_range: last_30d (default)

---

Step 1: Fetch Active Ads

Get list of all active ads with performance data:

• get_ads_list → ad names, IDs, campaign association
• get_ads_insights per ad: frequency, CTR, ROAS, spend, impressions

---

Step 2: Calculate Fatigue Signals

For each ad, compute:

• Days active: from ad creation date
• CTR drop %: (Week 1 CTR − Current CTR) / Week 1 CTR × 100
• Frequency decay: CTR at freq 1-2x vs CTR at freq 5x+
• Decay rate: CTR drop % per day since launch
• Days until death: estimate based on current decay rate and "Dead" threshold
• Lifespan status: see Classification table below

---

Step 3: Fatigue Classification

Level	Criteria	Action
Fresh	< 7 days OR CTR drop < 10%	Monitor
Mild	7-14 days, CTR drop 10-20%	Monitor
Moderate	14-21 days, CTR drop 20-35%	Rotate Soon
Severe	21-28 days, CTR drop 35-50%	Refresh Now
Dead	> 28 days OR CTR drop > 50% OR Freq > 5x with CTR < 50% of baseline	Kill Immediately

---

Step 4: Lifespan Status

Status	Criteria
Growing	ROAS within 10% of peak, < 14 days
Peak	ROAS at maximum, 7-21 days
Declining	ROAS decay > 5%/day, > 21 days
Expired	ROAS < 50% of peak OR > 35 days

---

Step 5: Build Rotation Pipeline

Classify all active creatives into pipeline buckets:

Status	Count
Active Winners
Active Decaying
Recently Killed
In Pipeline

Calculate:

• Creatives dying this week
• Creatives dying next week
• New creatives needed
• Testing velocity (new creatives launched per week)

---

Step 6: Output Format

SECTION 1: Fatigue Status Overview

Fatigue Health Table (top 15-20 creatives):

Creative	Days Active	Current Freq	CTR Week 1	CTR Now	Drop %	Fatigue Level	Action

SECTION 2: Frequency Decay Analysis

Frequency Impact Table (top 10-12 highest frequency):

Creative	Freq 1-2x CTR	Freq 3-4x CTR	Freq 5x+ CTR	Current Freq	Decay Rate

SECTION 3: Performance Degradation Timeline

ROAS Decay Table (top 12-15):

Creative	Launch Date	Peak ROAS	Current ROAS	Days to Peak	Days Since Peak	Decay Rate	Lifespan Status

SECTION 4: Rotation Pipeline Status

Creative Inventory Table:

Status	Count
Active Winners
Active Decaying
Recently Killed
In Pipeline

Refresh Needs: Dying this week, dying next week, new needed, testing velocity.

SECTION 5: Refresh Recommendations

Action Priority Table (top 10-12):

Creative	Current Status	Days Until Death	Action	Priority	New Direction

Priority: Urgent (24h), High (3d), Medium (7d), Low (monitor).

---

Rules

• Prioritize by: Severity > Days Until Death > Spend
• New Direction must be specific angles, not generic advice ("Test urgency angle" not "new creative")
• No executive summary
• This is analysis only — recommendations route to meta-ads-recommendation

---

Session Context — What This Skill Writes

After completing analysis, store the following in session context:

Key	Description	Example
rotation_pipeline	Creative inventory by status bucket	"2 growing, 3 declining, 1 dead, 2 in pipeline"
fatigue_level	Per-creative fatigue classification	"Blue Banner v2: Severe; CTR drop 52%"
days_until_death	Estimated days before each creative expires	"Banner v3: 8 days; Banner v4: 22 days"

meta-ads-recommendation reads these keys to produce the creative rotation action plan. For new creative generation, route to meta-ads-builder.