Creative Fatigue Detector
SKILL.md
sample_input.json
expected_output.json
skillssocial-media-adsSKILL.md
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
# Creative Fatigue Detector
## Purpose
Identify creative fatigue BEFORE it destroys performance by monitoring:
- Declining engagement metrics over time
- Rising costs as algorithms struggle to find new audiences
- Frequency thresholds that indicate audience saturation
- Performance decay patterns that signal refresh timing
**Core Philosophy:** By the time you notice a creative is "tired," you've already wasted budget. Catch fatigue at the first signs, not the obituary.
---
## When to Use This Skill
- Weekly monitoring of active creatives (recommended)
- When CPMs start rising without explanation
- When CTR drops on previously strong performers
- Before deciding whether to refresh or kill a creative
- When scaling spend (higher spend = faster fatigue)
---
## Required Context
| Data Point | Required | Notes |
|------------|----------|-------|
| Daily/weekly performance by creative | Yes | Need time series data |
| Frequency metrics | Yes | Avg frequency per user |
| CPM over time | Yes | Key fatigue indicator |
| CTR over time | Yes | Engagement decay signal |
| Conversion rate over time | Yes | Bottom-line impact |
| Days running | Yes | Duration in market |
| Cumulative reach | Helpful | Audience exhaustion signal |
| Audience size | Helpful | Context for frequency |
---
## The Fatigue Lifecycle
Every creative follows a predictable lifecycle:
```
PERFORMANCE
^
│ ╭─────╮
│ ╱ ╲
│ ╱ ╲
│ ╱ ╲ ← Peak Performance
│ ╱ ╲
│ ╱ ╲
│ ╱ ╲
│ ╱ Learning ╲ Fatigue
│ ╱ Phase ╲ Phase
│╱ ╲
└──────────────────────────────────→ TIME
│ │ │
Launch Peak Decline
(3-7d) (varies) (action needed)
```
**Typical Timeline by Spend Level:**
| Daily Spend | Learning Phase | Peak Performance | Fatigue Onset |
|-------------|----------------|------------------|---------------|
| < $50/day | 7-14 days | 3-6 weeks | 6-10 weeks |
| $50-200/day | 5-10 days | 2-4 weeks | 4-6 weeks |
| $200-500/day | 3-7 days | 1-3 weeks | 2-4 weeks |
| $500+/day | 2-5 days | 1-2 weeks | 1-3 weeks |
---
## Fatigue Detection Framework
### Primary Fatigue Signals
Monitor these metrics for week-over-week changes:
| Signal | Yellow Flag | Red Flag | Critical |
|--------|-------------|----------|----------|
| **CTR Decline** | -10% WoW | -20% WoW | -30%+ WoW |
| **CPM Increase** | +15% WoW | +25% WoW | +40%+ WoW |
| **CPA Increase** | +15% WoW | +25% WoW | +40%+ WoW |
| **Frequency** | > 2.5 | > 3.5 | > 5.0 |
| **Hook Rate Decline** (video) | -15% WoW | -25% WoW | -35%+ WoW |
### Secondary Fatigue Signals
| Signal | Warning Threshold | Notes |
|--------|-------------------|-------|
| Reach plateau | < 5% new reach in 7 days | Audience exhausted |
| Negative feedback spike | > 50% increase | Users actively hiding/blocking |
| Share/Save decline | -30% WoW | Content no longer resonates |
| Comment sentiment shift | Negative increase | Audience turning on creative |
---
## Fatigue Scoring System
Calculate a Fatigue Score (0-100) for each creative:
### Component Weights
| Component | Weight | Calculation |
|-----------|--------|-------------|
| CTR Decay | 25% | (Current CTR / Peak CTR) × 100 |
| CPM Inflation | 25% | (Peak CPM / Current CPM) × 100 |
| Frequency Pressure | 20% | Max(0, 100 - (Frequency × 20)) |
| Performance Trend | 20% | 7-day trend direction score |
| Days Running | 10% | Decay curve based on spend level |
### Fatigue Score Interpretation
| Score | Status | Recommended Action |
|-------|--------|-------------------|
| 80-100 | **Fresh** | Full steam ahead |
| 60-79 | **Healthy** | Monitor weekly |
| 40-59 | **Early Fatigue** | Prepare refresh, reduce by 20% |
| 20-39 | **Fatigued** | Reduce spend 50%, launch replacement |
| 0-19 | **Exhausted** | Pause immediately |
---
## Fatigue Pattern Recognition
### Pattern 1: Sudden Death
```
Performance: ████████████▁▁▁
Timeline: ─────────────────→
Cause: Usually algorithm/audience change, not true fatigue
Action: Check for external factors before pausing
```
### Pattern 2: Gradual Decline
```
Performance: ████████▇▇▆▅▄▃▂
Timeline: ─────────────────→
Cause: Classic fatigue - audience saturation
Action: Begin refresh at ▇ stage, pause at ▃ stage
```
### Pattern 3: Plateau then Drop
```
Performance: ████████████████▄▂
Timeline: ─────────────────→
Cause: Reached maximum viable audience
Action: Audience expansion or creative refresh needed
```
### Pattern 4: Oscillating Decline
```
Performance: ████▆████▅███▃██▂
Timeline: ─────────────────→
Cause: Inconsistent delivery, possible targeting issues
Action: Check audience overlap, consider consolidation
```
---
## Frequency Analysis Deep Dive
Frequency is the #1 predictor of fatigue:
### Frequency Thresholds by Objective
| Campaign Objective | Safe Zone | Caution | Danger |
|-------------------|-----------|---------|--------|
| Awareness/Reach | < 3.0 | 3.0-5.0 | > 5.0 |
| Consideration | < 2.5 | 2.5-4.0 | > 4.0 |
| Conversions | < 2.0 | 2.0-3.5 | > 3.5 |
| Retargeting | < 4.0 | 4.0-7.0 | > 7.0 |
### Frequency vs. Audience Size Rule
```
Safe Frequency Limit ≈ Audience Size / (Daily Reach × Days Running)
If actual frequency > Safe Limit × 1.5 → Fatigue risk HIGH
```
---
## Time-Series Analysis Template
Track each creative across these periods:
| Metric | Week 1 | Week 2 | Week 3 | Week 4 | Trend | Alert |
|--------|--------|--------|--------|--------|-------|-------|
| CTR | | | | | | |
| CPM | | | | | | |
| CPA | | | | | | |
| ROAS | | | | | | |
| Frequency | | | | | | |
| Conv Rate | | | | | | |
| Hook Rate | | | | | | |
**Trend Calculation:**
- ↑ = Improving (>5% better than previous)
- → = Stable (within ±5%)
- ↓ = Declining (<5% worse than previous)
- ↓↓ = Rapid Decline (>15% worse)
---
## Creative Lifespan Estimation
Based on current trajectory, estimate remaining viable days:
### Calculation Method
```
Remaining Days = Days to Threshold / Current Decay Rate
Where:
- Days to Threshold = days until CPA exceeds target by 30%
- Current Decay Rate = (This Week CPA - Last Week CPA) / Last Week CPA
```
### Lifespan Categories
| Estimated Remaining Life | Status | Action |
|-------------------------|--------|--------|
| > 21 days | Healthy runway | Continue monitoring |
| 14-21 days | Moderate runway | Brief replacement creative |
| 7-14 days | Short runway | Launch replacement NOW |
| < 7 days | Critical | Replacement should be live |
---
## Refresh vs. Kill Decision Matrix
When fatigue is detected, decide between refresh and kill:
| Factor | Refresh | Kill |
|--------|---------|------|
| Peak performance | Top 20% of account | Bottom 50% of account |
| Brand value | High brand consistency | Generic/replaceable |
| Production cost | High (video, UGC) | Low (static) |
| Pattern history | First fatigue cycle | 2nd+ fatigue cycle |
| Audience response | Positive comments | Negative feedback |
**Decision Formula:**
- If creative was ever a Star (Tier 1) AND it's first fatigue cycle → **Refresh**
- If creative never exceeded Average (Tier 4) → **Kill**
- If 2nd fatigue cycle with refresh already attempted → **Kill**
---
## Refresh Strategies
When refreshing rather than killing:
### Light Refresh (Extends life 2-4 weeks)
- Change thumbnail/cover image
- Update headline/primary text
- Modify CTA
- Adjust color scheme
### Medium Refresh (Extends life 4-6 weeks)
- New opening hook (first 3 seconds)
- Different aspect ratio version
- Add/remove sound
- Re-edit with different pacing
### Heavy Refresh (Essentially new creative)
- New footage/imagery
- Different creator/model
- New concept, same product
- Complete re-brief
---
## Output Deliverables
### 1. Creative Health Dashboard
```
CREATIVE FATIGUE REPORT
Generated: [Date]
Period Analyzed: [Last 4 weeks]
OVERALL HEALTH SCORE: [X]/100
SUMMARY:
├── Fresh (80-100): [X] creatives
├── Healthy (60-79): [X] creatives
├── Early Fatigue (40-59): [X] creatives ⚠️
├── Fatigued (20-39): [X] creatives 🔴
└── Exhausted (0-19): [X] creatives 💀
```
### 2. Fatigue Alert List
| Creative | Fatigue Score | Key Signal | Est. Life | Action |
|----------|---------------|------------|-----------|--------|
| | | | | |
### 3. Trending Concerns
Creatives showing early warning signs (healthy now, but trending down):
| Creative | Current Score | Trend | Watch Metric |
|----------|---------------|-------|--------------|
| | | | |
### 4. Refresh Queue
Creatives recommended for refresh rather than kill:
| Creative | Priority | Refresh Type | Timing |
|----------|----------|--------------|--------|
| | | | |
### 5. Kill List
Creatives beyond saving:
| Creative | Score | Wasted Spend (Last 7d) | Replacement Status |
|----------|-------|------------------------|-------------------|
| | | | |
---
## Proactive Fatigue Prevention
### Production Pipeline Rule
```
For every $X in ad spend, produce Y new creatives:
| Monthly Spend | New Creatives/Month | Reasoning |
|---------------|---------------------|-----------|
| < $5K | 2-4 | Low fatigue pressure |
| $5K-$20K | 4-8 | Moderate rotation needed |
| $20K-$50K | 8-15 | Active rotation required |
| $50K-$100K | 15-25 | Continuous production |
| $100K+ | 25+ | Creative team/agency needed |
```
### Rotation Strategy
- Never let a single creative exceed 40% of total spend
- Maintain 3-5 active creatives minimum per audience
- Launch new creative before old one hits "Fatigued" status
- A/B test refreshes against originals
---
## Integration with Other Skills
| Use Case | Connected Skill | Purpose |
|----------|-----------------|---------|
| Fatigued Star creative | → meta-ads-creative-ranker | Identify what made it win |
| Ready to refresh | → ugc-brief-generator | Brief replacement content |
| Multiple fatigued | → social-ads-budget-allocator | Reallocate away from fatigued |
| Pattern across audiences | → audience-overlap-mapper | Check if fatigue is audience-specific |
---
## Limitations
- Cannot predict algorithm changes that affect delivery
- Requires minimum 2 weeks of data for trend analysis
- Small audiences fatigue faster than estimates suggest
- Seasonal factors can mimic fatigue patterns
- Cannot assess creative quality, only performance metrics
---
## Quality Checklist Before Delivery
- [ ] Minimum 2 weeks of data analyzed per creative
- [ ] Fatigue scores calculated for all active creatives
- [ ] Time-series trends documented (not just current snapshot)
- [ ] Frequency analysis completed
- [ ] Lifespan estimates provided for concerning creatives
- [ ] Clear Refresh vs. Kill recommendations
- [ ] Production pipeline recommendation included
- [ ] Early warning list provided (not just current problems)
ReadyCreative Fatigue Detector
MarkdownUTF-8Verified