SKILL.md
- name:
- creative-cadence-operating-system
- description:
- Use this skill when planning creative production for Meta — iteration hierarchy, concept sourcing, format playbook, testing cadence, and quality scoring.
Creative Cadence Operating System - Meta B2B
The complete system for what creative to build, when to build it, how to iterate, and when to retire it. Companion to the Meta Ads Operating System (which governs when to swap/graduate/scale ads). This system governs the CREATIVE PRODUCTION pipeline that feeds the OS.
Core Principle
Creative accounts for 56-70% of ad performance on Meta (post-Andromeda). The algorithm uses creative - copy, images, video transcripts, carousels - as the primary signal for who to serve your ads to. Targeting inputs are suggestions; creative is the real filter.
Your creative is your targeting. Your hook is your filter. Your iteration velocity determines your scaling ceiling.
1. The Iteration Priority Hierarchy
When iterating on a winning ad, change elements in this order. Each element has decreasing impact on performance.
| Priority | Element | Impact | Reasoning |
|---|---|---|---|
| 1 | Hook (first 3 seconds / first line) | Highest | 90% of viewers decide in the first 3 seconds whether to engage. The hook determines who stays. Changing the hook changes who your ad reaches - it's the single highest-leverage creative variable. |
| 2 | Visual Treatment | High | Human faces increase engagement 38%. 4:5 aspect ratio delivers 18% higher CTR and 22% lower CPA than other ratios. The visual is what stops the scroll. |
| 3 | Format (static/video/carousel/UGC) | Medium | Format determines consumption pattern. Video allows storytelling (21-35 day lifespan), carousel allows multiple hooks (21-35 days), UGC builds trust (28-42 days). Format change = "new ad" to the algorithm. |
| 4 | Body Copy / CTA | Lower | Once someone is hooked, body copy matters less than the initial capture. CTA changes produce marginal differences unless the CTA is fundamentally mismatched. |
Why Hook-First?
- 90% of people bounce if not captured in 3 seconds (video) or 1-2 seconds (static)
- Target hook rate: 20-25% (people who watch past 3 seconds / total impressions)
- Hook rate below 15% = the ad will never perform, regardless of body quality
- Changing the hook on a winning concept can extend its lifespan by 2-4 weeks
Hook Types That Work for B2B
- Provocative question: "Why are 73% of CFOs still using spreadsheets for cash flow?"
- Specific pain: "Your finance team spends 15 hours/week on manual reconciliation."
- Counterintuitive claim: "The best ERPs are making your cash flow worse."
- Social proof lead: "A mid-market SaaS company cut their close time from 14 days to 3."
- Number-first: Numbers in headlines improve CTR 20-30%. Lead with specifics.
2. Concept Sourcing System
Before creating any ads, source 15-25 buyer situations. Each situation is a unique creative angle.
Where to Source Concepts (Priority Order)
| Source | Method | Quality Signal |
|---|---|---|
| Organic winners | Audit LinkedIn/blog/webinar content from last 12 months. High engagement = validated message. | Market already told you it works |
| Customer interviews | "What triggered your search?" "What almost stopped you?" "What surprised you?" | Direct from ICP |
| Sales call recordings | Objections, "aha" moments, specific pain language | Real buyer language |
| Support tickets | Problems driving upgrades or churn | Active pain points |
| Meta Ad Library | Search competitors. Ads running 3+ months = likely profitable. Note angles and formats, not exact creative. | Competitor-validated |
| LinkedIn/forum discussions | What your ICP complains about, asks about, celebrates | Community language |
Concept Types
| Type | Description | Best For | Iteration Potential |
|---|---|---|---|
| Problem/Solution | Name specific pain, show the fix | Cold traffic, problem-aware | High - many pain points |
| Before/After | Show the transformation | Visual products, clear outcomes | Medium |
| UGC/Founder Video | Customer or founder talking naturally | Trust building, cold audiences | High - different people |
| Meme/Humor | Industry-specific humor | Pattern interrupt, cold | Low - hard to iterate |
| Data/Statistics | Lead with a compelling stat | Authority building | Medium - different stats |
| Testimonial | Real customer quote with photo/video | Social proof, warm audiences | High - different customers |
| Case Study | "[Company] achieved [specific result]" | Proof of concept, warm | High - different companies |
| Comparison | Your solution vs status quo or competitors | Solution-aware, competitive | Medium |
Image-First Concept Validation
Test new concepts as static images first before investing in video or other formats. Images are faster and cheaper to produce, making them ideal for proving whether a concept resonates before committing to higher-effort formats.
The flow:
- Proof the concept with an image ad - design a static that captures the hook, angle, and message
- Validate using the Decision Tree - if the image passes Stage 1 (delivery check) and Stage 2 Steps 2-4 (generates pixel leads, qualified leads, cost within range), the concept is proven
- Iterate the winning concept into other formats - take the proven concept and produce video (founder talking, AI-generated video, screen recording), carousel, or UGC versions
- Scale the best-performing format - the format that performs best on the proven concept gets the production investment
Why image-first:
- Static images are the fastest to produce (hours, not days)
- Lower production cost means more concepts tested per cycle
- A concept that works as an image has a validated hook and angle - video adds engagement and extends lifespan on top of that
- A concept that fails as an image would likely fail as video too - saving significant production time and cost
- Video iterations on proven image concepts have a higher win rate than untested video concepts
When to skip image-first and lead with video:
- UGC/testimonial concepts (inherently video-native - the person IS the format)
- Concepts that require demonstration or motion to communicate the value (e.g., product walkthrough)
- Iterations on already-proven concepts where the format change IS the variable being tested
The 50/30/20 Production Split
- 50% of creative production: Iterations on top performers - same concept, different hook/format/visual
- 30% of creative production: Iterations on other performers - same angle, different execution
- 20% of creative production: Completely new concepts - different angle, format, message
For each proven ad, produce 10 variations over its lifespan.
3. Creative Production Formula
All formulas reference TCPL (Target Cost Per Lead) from the Meta Ads Operating System.
Tests Per Month
Tests/month = Testing Campaign monthly budget / (3 x TCPL)
Logic: The Testing Campaign gets 20% of total budget. Each test needs 3x TCPL spend for Stage 2 quality evaluation (95% statistical confidence, Poisson distribution: e^-3 = 0.0498). Tests that fail Stage 1 (delivery check at Day 7) free up budget faster, increasing actual throughput.
| Monthly Budget | Testing Campaign (20%) | Tests/Month (TCPL = EUR 500) | Tests/Week |
|---|---|---|---|
| EUR 20K | EUR 4K | ~3 | ~1 |
| EUR 30K | EUR 6K | ~4 | ~1 |
| EUR 50K | EUR 10K | ~7 | ~2 |
| EUR 100K | EUR 20K | ~13 | ~3-4 |
Win Rates
| Type | Win Rate | Tests for 1 Winner | Source |
|---|---|---|---|
| Iterations on winners | ~25% (1 in 4) | 4 | Foxwell, Pilothouse ($50K-500K/month) |
| New concepts | ~10% (1 in 10) | 10 | Foxwell, Pilothouse ($50K-500K/month) |
| Blended (50/50 mix) | ~17% (1 in 6) | 6 | Derived from iteration/concept mix |
Proven Ads Needed
Minimum proven ads = Monthly budget / EUR 5,000
Each proven ad absorbs ~EUR 5K/month before frequency causes fatigue. Above EUR 6-8K on one ad, performance degrades.
Scaling Readiness
Creative deficit = Proven ads needed (Scaling Campaign) - Current proven ads
Tests to close gap = Creative deficit / 0.17
Months to close = Tests to close / Tests per month
You cannot scale budget ahead of creative.
4. Format-Specific Playbook
Static Images
- Aspect ratio: 4:5 (1080x1350) for Feed, 9:16 for Stories. 4:5 delivers 18% higher CTR and 22% lower CPA.
- Text on image: Under 20% of image area. Old rule removed but algorithm still penalizes text-heavy images.
- Lifespan: 14-28 days. People register static images quickly, then scroll past.
- Key stats: Numbers in headlines improve CTR 20-30%. Human faces increase engagement 38%.
- Refresh tactic: Change hook text/headline while keeping visual = buys 7-14 extra days. Different color scheme = "new" to algorithm.
- Duplication warning: Meta's visual recognition flags similar images as duplicates. Changes must be visually distinct.
Video
- Hook window: First 3 seconds are everything. 90% bounce if not captured.
- Hook rate target: 20-25% (ThruPlay-to-impression). Below 15% = never performs.
- Captions: Always on. 85% of users watch with sound off.
- Aspect ratio: 9:16 vertical for Stories/Reels, 4:5 for Feed. 9:16 with audio = 12% higher conversions.
- Length: Under 30 seconds for prospecting, up to 60 seconds for retargeting.
- Lifespan: 21-35 days. More engaging, fatigues slower than static.
- Refresh tactic: Swap opening hook while keeping body = extends lifespan significantly.
- Production styles: Founder talking, customer testimonial, screen recording, animation. Different styles = different audience signals.
Carousel
- Card count: 3-5 cards (10 max). First 3 cards do most work.
- First card: Must stand alone as a hook. Some placements show only the first card.
- Lifespan: 21-35 days. Multiple cards give built-in variety.
- Refresh tactic: Rearrange card order = "new" ad to algorithm. Add 1-2 new cards while keeping best cards.
- Best for: Step-by-step processes, multiple benefits, before/after comparisons.
UGC / Testimonial
- Lifespan: 28-42 days (longest of all formats). Authenticity fatigues slower.
- Refresh tactic: Different customer, same concept. Multiple UGC videos from different people = built-in rotation system.
- B2B caveat: UGC outperforms polished content, but B2B needs credibility markers (job title, company, specific result).
- Video vs quote card: Video > quote card for cold audiences. Quote cards work for retargeting.
- Production tip: Record 3-5 customers at once. Edit into 10+ variations with different hooks and cuts.
5. The Testing Cadence
| Action | Frequency | What to Do |
|---|---|---|
| Check Stage 1 (Delivery) | Wednesday (Day 7 for new ads) | Check CBO spend distribution in Testing Campaign. Swap ads below underdelivery threshold (1/N/2). |
| Launch new iterations | Every 2 weeks | 2-3 iterations on current top performers. Change hook first (Priority 1). Launch in Testing Campaign. |
| Launch new concepts | Monthly (minimum) | 1-2 completely new angles as static images. Source from buyer situations list. Launch in Testing Campaign. |
| Retire depleted ads | When CTR drops 30%+ from peak OR frequency > 5.0 | Don't iterate - concept is exhausted. Remove from Scaling Campaign. |
| Iterate winning concepts into new formats | When image concept passes Stage 1 + Stage 2 | Produce video/carousel/UGC version of the proven concept. The concept is validated - now test which format maximizes it. |
| Full creative refresh | Quarterly | Re-audit buyer situations, source new angles, refresh visual identity. |
| Check fatigue signals | Weekly (Monday) | Pull frequency, CTR trend, CPM trend, ad relevance diagnostics for Scaling Campaign ads. |
Testing Protocol
- New concepts: Launch as a static image in the Testing Campaign first (image-first validation). Skip image-first only for video-native concepts (UGC, testimonials, product demos).
- Launch in the Testing Campaign (two-campaign structure from Meta Ads OS)
- Stage 1 - Delivery Check (Day 5-7): Check CBO spend distribution. If ad gets less than half its expected share (1/N/2 where N = active test ads), swap immediately - don't wait for 3x TCPL.
- Stage 2 - Quality Evaluation: Minimum spend 3x TCPL before judging (95% confidence threshold)
- Minimum runtime: 14 days (learning phase + stabilization)
- Judge by qualified lead rate and cost per QL (not CTR or pixel CPL)
- If winning (all 5 graduation criteria met): graduate to Scaling Campaign
- If the concept is proven AND was image-first: produce video/carousel/UGC versions of the same concept
- If losing: classify per Decision Tree - swap reason determines next creative
What to Build When an Ad Fails
| Failure Type | What It Means | What to Build Next |
|---|---|---|
| Delivery failure (Stage 1) | Creative doesn't stop the scroll | Different hook/visual/format. Don't iterate on copy - audience never got that far. |
| Zero pixel leads (Step 2 fail) | Creative doesn't resonate | Abandon concept. Completely different angle/format/message. |
| Zero QLs despite leads (Step 3 fail) | Attracts wrong audience | Keep format, change the pain point to filter for ICP. |
| Low quality < 40% (Step 3 fail) | Not enough ICP filtering | Add revenue figures, industry terms, ICP-specific language. |
| Cost > 1.5x TCPL (Step 4 fail) | Right audience, wrong execution | Same angle, different hook/format/visual/CTA. |
| Fatigue (Step 6) | Audience exhausted | New hook + different format + different customer/color. |
6. Fatigue Detection System
Signals (in order of urgency)
| Signal | Threshold | Urgency |
|---|---|---|
| Frequency > 4.0 (cold prospecting) | Immediate action needed | URGENT |
| Frequency > 6.0 (retargeting) | Immediate action needed | URGENT |
| CTR drops 20%+ from baseline over 7 days | Creative dying | WARNING |
| CPM rising 30%+ over 2 weeks | Algorithm struggling to deliver | WARNING |
| Ad Relevance: "Below Average" in any metric | Quality issue | WARNING |
| CPA increasing with stable targeting | Likely fatigue (rule out other causes first) | MONITOR |
Frequency Thresholds by Campaign Type
| Campaign Type | Safe | Warning | Critical |
|---|---|---|---|
| Cold prospecting | 1.0-2.5 | 2.5-3.5 | > 4.0 |
| Retargeting | 2.0-4.0 | 4.0-5.5 | > 6.0 |
| ABM | 2.0-5.0 | 5.0-7.0 | > 8.0 |
ABM tolerates higher frequency because audiences are small and message reinforcement is intentional. But even ABM hits diminishing returns past 5-6 impressions per person per week.
B2B Creative Lifespan
| Format | Typical Lifespan | Why |
|---|---|---|
| Static image | 14-28 days | Seen, registered, scroll past |
| Video (< 30s) | 21-35 days | More engaging, lasts longer |
| Carousel | 21-35 days | Multiple cards = more novelty |
| UGC / testimonial | 28-42 days | Authenticity fatigues slower |
B2B audiences are smaller so frequency builds faster. But B2B users pay less attention to ads while scrolling past work content. Net result: 14-21 day refresh cycle for most B2B campaigns.
7. Competitive Analysis Framework
Meta Ad Library Process
- Go to facebook.com/ads/library
- Search competitor company pages
- For each competitor record:
- Total active ads (volume signal)
- Ads running 3+ months (likely profitable)
- Dominant format (static/video/carousel)
- Angles used (pain point, testimonial, comparison, etc.)
- Offer types (demo, guide, webinar, calculator)
- CTA language
- Creative style (polished/UGC/meme)
What to Learn vs What NOT to Copy
Learn: Which angles they keep running (validated by spend), format preferences, offer types that persist, how they handle ICP specificity.
Don't copy: Exact creative (won't work for your brand), larger competitors' scale strategy (they absorb losses you can't), assume volume = quality.
Cadence
- Full competitive audit: quarterly
- Quick check (top 3 competitors): monthly
- Reactive check: when launching new concept types
8. Quality Scoring System
Standard Meta metrics (CTR, CPL) don't tell you which ads drive revenue. Validate ads against downstream quality.
The Scoring Framework
After ~20 demos, score each prospect:
| Factor | 0 Points | 1 Point | 2 Points | 3 Points |
|---|---|---|---|---|
| Urgency | "Just browsing" | Some pain, no timeline | Active problem, 3-6 month timeline | Burning problem, need it NOW |
| Budget | No budget, no authority | Budget exists, unclear | Budget allocated for category | Budget approved, ready to spend |
| Fit | Not ICP | Partially matches | Good ICP match | Perfect ICP match |
Max score: 9 per prospect.
How to Use
- After 20 calls, calculate average quality score per ad
- Sort ads by quality score, NOT by CPL or CTR
- Top 3 ads by quality score = real winners
- Scale those 3 with 10 variations each (see Production Formula)
- Kill any variation where average quality drops below 5
Critical insight: The ad with the lowest CPL is often NOT the best ad. A $15 CPL with 2% CTR might generate garbage leads. A $40 CPL with 1.2% CTR might generate $100K deals.
9. The Ad Copy Formula
7-Step Structure
- DIRECT OFFER - state the value proposition clearly
- PAIN - name the specific problem your ICP has
- SOLUTION - show how you solve it
- PAIN EXPLAINED - deeper on consequences of not solving
- SOLUTION EXPLAINED - deeper on benefits of solving
- SOCIAL PROOF - customer quote, stat, or logo
- CTA - clear next step with motivation
Copy Rules for B2B Meta
- Call out who the ad is for explicitly - "mosquito repellent" for non-ICP
- Name the specific problem, not the category
- Include time-bound element when possible ("in 60 days", "this quarter")
- First 2-3 lines must hook - most users won't expand "See More"
- "Sell the click, not the product" - not selling a $30K contract in an ad
Quick Reference - All Creative Formulas
| Formula | Calculation | Source |
|---|---|---|
| Delivery check threshold | (1 / N) / 2, where N = active test ads | Two-stage evaluation (Meta Ads OS) |
| Tests per month | Testing Campaign budget / (3 x TCPL) | Foxwell, Pilothouse, Poisson |
| Min proven ads | Total budget / EUR 5,000 | Creative fatigue data ($25M+ spend) |
| Creative deficit | Proven needed - Current proven | Derived |
| Tests to close gap | Deficit / 0.17 | Blended win rate |
| Iteration win rate | ~25% (1 in 4) | Foxwell, Pilothouse |
| New concept win rate | ~10% (1 in 10) | Foxwell, Pilothouse |
| Hook rate target | 20-25% | Industry standard |
| Static lifespan | 14-28 days | $25M+ spend data |
| Video lifespan | 21-35 days | $25M+ spend data |
| UGC lifespan | 28-42 days | $25M+ spend data |
| Fatigue warning | Frequency 2.5-3.5 (cold) | $25M+ spend data |
| Fatigue critical | Frequency > 4.0 (cold) | $25M+ spend data |
Related Files
- meta-ads-operating-system.md - THE decision framework for swap/graduate/scale (read first)
- creative-strategy.md - deeper detail on concept types, copy formulas, placement optimization
- creative-fatigue-detection.md - full fatigue detection workflow, format-specific notes
- message-validation.md - quality scoring process, winner scaling, buyer situations
- optimization-playbook.md - decision trees for CPA/CTR/quality issues