Creative Fatigue Detector

Overview

This skill diagnoses creative fatigue by analyzing performance trend data against statistical thresholds, identifies which specific creatives are fatigued versus underperforming for other reasons, and prescribes targeted refresh strategies. It distinguishes between true creative fatigue, audience saturation, seasonal effects, and market-level shifts to prevent premature creative kills.

When to Use

• Ad performance metrics are declining over consecutive days or weeks
• CPA is rising while CTR and engagement are dropping simultaneously
• Frequency metrics are exceeding platform-recommended thresholds
• Planning proactive creative refresh schedules for evergreen campaigns
• Auditing a creative library to identify stale vs. still-performing assets
• Investigating whether poor performance is creative-driven or systemic

Required Inputs

Input	Required	Description
Performance Data	Yes	Daily or weekly metrics per creative: impressions, clicks, CTR, CPA, ROAS, conversions
Time Range	Yes	Minimum 14 days of data; 30–90 days preferred for trend analysis
Creative Identifiers	Yes	Creative/ad IDs, names, or descriptions to differentiate assets
Frequency Data	Recommended	Average frequency per creative or ad set over the analysis period
Platform	Recommended	Meta, Google, TikTok, etc. — fatigue thresholds vary by platform
Audience Size	No	Estimated audience pool size to calculate saturation metrics
Benchmark KPIs	No	Historical or target benchmarks for comparison

Methodology

Step 1 — Data Normalization & Baseline Establishment

1. Calculate the baseline performance window: first 7 days of delivery (or first 25% of data)
2. Compute baseline averages for CTR, CPA, ROAS, conversion rate, and CPM
3. Normalize data for spend fluctuations — use rate metrics (CTR, CPA) not absolute volumes
4. Flag and exclude anomalous days (>2σ deviations from mean due to external factors)
5. Adjust for day-of-week effects using 7-day rolling averages

Step 2 — Fatigue Signal Detection

Apply the following diagnostic tests to each creative:

Primary Fatigue Indicators (must show 2+ to confirm fatigue):

Metric	Fatigue Signal	Threshold
CTR	Sustained decline	>20% drop from baseline over 7+ days
CPA	Sustained increase	>25% increase from baseline over 7+ days
Frequency	Over-saturation	>3.0 (prospecting) or >8.0 (retargeting) on Meta
CPM	Auction competitiveness drop	>30% increase without seasonal cause
Conversion Rate	Declining efficiency	>15% drop from baseline, sustained 5+ days

Secondary Fatigue Indicators (supporting evidence):

Metric	Signal	Threshold
Video completion rate	Attention decline	>25% drop in ThruPlay or 75% VCR
Engagement rate	Audience disinterest	>30% drop in likes, comments, shares
Negative feedback	Active rejection	>2× increase in "Hide Ad" or negative reactions
Impression share decay	Algorithmic deprioritization	>20% drop in delivery share within ad set

Step 3 — Fatigue Classification

Classify each creative into a fatigue category:

Category	Criteria	Recommended Action
Healthy	All metrics within ±10% of baseline	Continue running; monitor weekly
Early Fatigue	1 primary indicator triggered	Alert; prepare replacement creative
Confirmed Fatigue	2+ primary indicators triggered	Reduce spend; begin creative swap
Exhausted	3+ primary + secondary indicators	Pause immediately; replace creative
Audience Saturated	High frequency + stable CTR but rising CPM	Expand audience; creative may be fine
Systemic Decline	All creatives declining uniformly	Investigate external factors, not creative

Step 4 — Root Cause Differentiation

Distinguish creative fatigue from confounding factors:

• Seasonal Effects: Compare YoY data; check industry CPM indices (Q4 CPMs rise 30–60%)
• Audience Exhaustion: If frequency is high but creative is fresh, the audience pool is too small
• Market Saturation: If all competitors show similar trends (use auction insights if available)
• Landing Page Issues: If CTR is stable but CPA rises, investigate post-click experience
• Algorithm Changes: If CPM spikes without frequency change, suspect platform auction dynamics
• Budget Scaling: If budget increased >20% recently, expect efficiency decline (not fatigue)

Step 5 — Fatigue Velocity Scoring

Calculate the rate of performance decay to prioritize action:

Fatigue Velocity = (Baseline CPA - Current CPA) / (Baseline CPA × Days Since Baseline)

Interpretation:
  FV < 0.5%/day  → Slow burn; 2-3 weeks before critical
  FV 0.5-1.0%/day → Moderate; prepare refresh within 1-2 weeks
  FV > 1.0%/day  → Rapid; refresh immediately

Estimate Days to Exhaustion (DTE):

DTE = (CPA Ceiling - Current CPA) / (Daily CPA Increase Rate)

Where CPA Ceiling = Baseline CPA × 1.5 (or max acceptable CPA)

Step 6 — Refresh Strategy Prescription

Based on fatigue classification, recommend specific actions:

For Early Fatigue:

• Introduce 2–3 new variants of the same winning angle
• Adjust visual treatment (color, layout, talent) while keeping message
• Test new format (static → video, or carousel → Reel)

For Confirmed Fatigue:

• Launch entirely new messaging angle
• Rotate to different creative framework (AIDA → PAS)
• Introduce new audience segments to reduce frequency pressure
• Test UGC or creator content if currently using branded

For Exhausted Creatives:

• Full creative replacement — new angle, format, and visual treatment
• Archive learnings: document what worked in baseline period
• Consider concept evolution rather than complete departure from brand

Proactive Refresh Cadence (CPG/retail benchmarks):

• Prospecting: Refresh every 2–3 weeks
• Retargeting: Refresh every 1–2 weeks (smaller audiences fatigue faster)
• Seasonal campaigns: Build 3× creative volume upfront; rotate weekly
• Always-on: Maintain 5–8 active variants per ad set; rotate bottom 2 every 2 weeks

Output Specification

1. Fatigue Dashboard: Per-creative summary with classification, fatigue velocity, and DTE
2. Trend Visualizations: Described data visualization specs for CTR, CPA, and frequency over time
3. Root Cause Analysis: For each fatigued creative, the diagnosed cause and confidence level
4. Action Plan: Prioritized list of creatives to pause, refresh, or continue with timelines
5. Refresh Recommendations: Specific creative direction for replacement assets
6. Monitoring Cadence: Ongoing fatigue-check schedule and alert thresholds

Examples

Input: "Our Meta prospecting campaign has 6 creatives running for 4 weeks. CTR dropped from 1.8% to 1.1%, CPA went from $18 to $28. Frequency is at 4.2. Help diagnose."

Output: Classification of each creative into fatigue categories; identification that 3 creatives are exhausted (FV >1.2%/day), 2 are confirmed fatigued, and 1 remains healthy. Root cause: audience over-saturation (frequency 4.2) compounding creative fatigue. Action plan: pause 3 exhausted creatives immediately, expand lookalike audience from 1% to 3%, introduce 4 new variants with UGC angle, implement weekly fatigue monitoring.

Input: "All our ads across Google and Meta are showing rising CPAs over the last 10 days. Is it creative fatigue or something else?"

Output: Systemic decline diagnosis — uniform CPA increase across platforms and creatives suggests external factor. Analysis points to Q4 CPM inflation (industry-wide 35% CPM increase in this period). Recommendation: hold creative constant, adjust CPA targets for seasonal benchmarks, increase bids, and revisit in January when auction pressure normalizes.

Guidelines

• Never diagnose fatigue with fewer than 14 days of data — early fluctuations are noise
• Always check frequency before blaming the creative; high frequency is often the root cause
• Distinguish between absolute metric declines and relative (rate-based) declines
• Account for spend changes — doubling budget will naturally raise CPA 10–20%
• For small-audience retargeting, fatigue thresholds should be more aggressive (lower frequency caps)
• Recommend creative diversification, not just replacement — maintain learnings from winners
• Platform-specific thresholds matter: TikTok creatives fatigue faster (7–14 days) than Meta (14–28 days)
• For CPG, seasonal purchase cycles affect conversion rates independent of creative quality

Validation Checklist

• Minimum 14 days of data analyzed per creative
• Baseline period is clearly defined and reasonable
• All primary and secondary fatigue indicators are evaluated
• Each creative receives a fatigue classification with supporting evidence
• Root cause analysis differentiates creative fatigue from confounding factors
• Fatigue velocity and DTE are calculated for actionable prioritization
• Refresh recommendations are specific and actionable (not generic)
• Monitoring cadence is established for ongoing fatigue prevention