Ad Creative Strategy
Overview
Creative is the single biggest lever in paid advertising performance. A great creative on a mediocre audience outperforms a mediocre creative on a perfect audience. This skill covers creative formats per platform, testing methodology, fatigue management, and scaling creative production.
Authoritative Foundations
- The B2B Institute (LinkedIn) — 95:5 rule — Only about 5% of B2B buyers are in market at any time, so creative must build memory with the other 95% as well as convert the 5%.
- Meta — Dynamic creative testing — Let the platform rotate headlines, images, and copy variants, then promote winners into fixed ads; test one variable family at a time.
- LinkedIn Marketing Solutions — Ad creative best practices — Follow format specs, lead with the value in the first line, and refresh creative before frequency-driven fatigue sets in.
- Ann Handley — Everybody Writes — Write ad copy that is specific, human, and about one idea; cut anything the buyer would skim past.
When to Use
- "Create ad creative for LinkedIn"
- "Develop a creative testing plan"
- "Our ads are fatigued — refresh creative"
- "What creative formats work on each platform?"
- "Scale our ad creative production"
- "AI-generated ad creative"
Step-by-Step Process
Phase 1: Creative Formats by Platform
LinkedIn Ads:
- Single image: 1200x627px. 1-2 lines of text overlay. Best for: offers, events.
- Carousel: 1080x1080px per card, 3-10 cards. Best for: storytelling, data.
- Video: 3-30 seconds. Square (1:1) or vertical (4:5). Best for: product demos.
- Document ads: PDF uploaded as carousel. Best for: guides, reports, templates.
Google Ads:
- Responsive search ads: 15 headlines (30 chars), 4 descriptions (90 chars). Google mixes and optimizes. Feed all 15 — more assets = better performance.
- Display: 1200x628, 300x250, 728x90, 160x600. Static or HTML5.
- YouTube: 15-30 seconds skippable. Hook in first 5 seconds.
Meta (FB/IG):
- Feed: 1080x1080 (square) or 1080x1350 (vertical). Video or static.
- Stories: 1080x1920 (full vertical). 15 seconds max.
- Lead forms: same creative specs, plus form questions.
TikTok:
- 9:16 vertical video. 15-60 seconds. Native-looking (not polished corporate).
- Hook in first 2 seconds. Text overlay. Captions. Trending audio optional.
Phase 2: Creative Testing Methodology
Always test: At minimum 2-3 variants running simultaneously. Never one creative at a time.
Test one variable at a time:
- Headline/hook (most impact)
- Visual style (photo vs illustration vs screenshot)
- Offer/CTA (demo vs trial vs guide)
- Social proof placement
- Format (single image vs carousel vs video)
Testing cadence:
- Launch 3-5 variants → run 7-14 days → kill underperformers → launch new variants against winners → repeat.
Platform minimums for statistical significance:
- LinkedIn: 1,000+ impressions per variant
- Google: 100+ clicks per variant
- Meta: 50+ conversions per variant per week
- TikTok: 5,000+ impressions per variant
Phase 3: Creative Fatigue Management
Creative fatigue = performance drops because audience has seen the ad too many times. Signs: CTR declining, CPC rising, frequency >3-4.
Prevention:
- Rotate creative every 2-3 weeks.
- Maintain 3-5 active variants per ad set.
- Use platform automated creative refresh (Meta Advantage+ Creative).
- Segment audiences to reduce overlap.
Phase 4: Scaling Creative Production
Traditional model: agency → brief → concept → production → 2-4 assets/month. $3-10K/month. Slow but polished.
AI-augmented model: 1 in-house creative + AI tools → 10-20 assets/week. $1-3K/month (tools + person). Fast, testable volume beats polished scarcity.
AI creative tools:
- Ad copy: ChatGPT, Claude, Jasper (write multiple headline/body variants).
- Image generation: Midjourney, DALL-E (concept visuals, backgrounds).
- Video: Arcads, Creatify (AI-generated UGC-style ads from scripts).
- Iteration: Canva AI, Adobe Firefly (resize, remix, translate).
The volume advantage: A $5K/month agency producing 4 polished assets/month loses to a $2K/month AI-augmented workflow producing 40 testable variants/month. Testing velocity beats production quality.
Output Format
Creative strategy document with: platform-specific format guide, testing plan (what to test first), fatigue management rules, creative production workflow, and creative asset tracker.
Quality Check
- Creative formats matched to each active platform
- Testing methodology defined (what variable, what sample size)
- Fatigue prevention rules (rotation cadence, frequency caps)
- Production workflow (who creates what, how fast)
- Creative asset tracker (variant, launch date, performance)
Common Pitfalls
-
One creative at a time. "Let's see how this performs" with no comparison. You can't optimize what you don't test.
-
Polished over volume. A beautiful ad nobody sees loses to 10 rough variants that find the winning message.
-
No fatigue monitoring. Same ad for 8 weeks with frequency >10. CTR drops 50%+. Rotate before performance degrades.
-
Creative not matching platform context. A polished corporate video on TikTok looks like an ad. Native-looking content wins.
-
Testing too many variables at once. Changed headline AND visual AND CTA. No idea which change caused the lift. One variable per test.
Execution Artifacts
references/framework-notes.md— Named frameworks and reference tablestemplates/output-template.md— Deliverable shell for agent outputscripts/check-output.py— Lightweight deliverable validator
Related Skills
- paid-advertising: Platform-specific ad setup
- a-b-testing: Experiment design and statistical significance
- graphic-design-gtm: Visual asset creation and sizing
- copywriting: Ad copy and headline writing


