Skill Creative Testing — בדיקת קריאייטיב ל-Claude

מתי להשתמש

כשמישהו אומר: "Creative testing", "בדיקת קריאייטיב", "Winning ad", "Ad fatigue", "A/B test creatives", "Iterate creative".

מקור / רקע מקצועי

• Common Thread Collective (CTC)
• Andrew Foxwell, Barry Hott
• Meta / Google Ads creative best practices

הוראות עבודה

1. Why Creative Matters

ב-2026, creative is 80% of paid performance.

• Algorithms reward engagement.
• Attention scarcity = creative wins.
• Audience targeting commoditized.

2. Creative Testing Framework

Step 1: Hypothesis

• "Carousel ads will outperform static for product education."

Step 2: Variables

• Test one variable at a time.
• Variables: Hook, Visual, Format, Length, Music, CTA.

Step 3: Budget per test

• $50-100/day per variant.
• 3-7 days run.
• Need 50+ conversions per variant.

Step 4: Decision

• Winner: 15%+ better than control on primary metric.
• Statistical significance ideally.

Step 5: Scale Winner

• Increase budget on winner.
• Pause losers.
• Iterate from winner.

3. Creative Variables — What to Test

Hook (3 שניות ראשונות) — Highest impact

• Question vs Statement.
• Visual vs Text-first.
• POV ("I tried...") vs Brand-first.

Visual

• Product shot vs Lifestyle.
• UGC-style vs Polished.
• Photo vs Video.
• Color schemes.

Format

• Static image.
• Video (15s, 30s, 60s).
• Carousel.
• Reels-style vertical.

Length

• 6s vs 15s vs 30s.

Music / Audio

• Trending vs Original.
• Voice over vs Music only.

Copy

• Long vs Short.
• Question vs Statement.
• CTA variations.

4. Creative Frameworks That Work

א. Problem-Solution-Outcome

1. Hook: Show problem (2-3s).
2. Agitate: Why it sucks (3-5s).
3. Reveal: Solution (5-7s).
4. Outcome: Show result (3-5s).
5. CTA (1-2s).

ב. Before-After

• Split-screen.
• Show transformation.
• Strong for products with visible results.

ג. Testimonial

• Customer talks (UGC-style).
• Authentic > Polished.
• Specific results > Generic praise.

ד. Founder's Story

• "Why I built this..."
• Personal, human.
• B2B effective.

ה. Demo

• Show product in action.
• Speed up boring parts.
• Highlight key features.

ו. Listicle

• "5 reasons to..."
• Visual list.
• Easy to follow.

ז. Comparison

• "Old way" vs "Our way".
• Subtle (don't bash competitors).

5. Hooks — 10 Proven

1. "You're doing X wrong" (controversy).
2. "3 things I wish I knew..." (educational).
3. "This will save you Y hours" (value).
4. "Stop scrolling — this is..." (pattern interrupt).
5. "I tried X for 30 days..." (story).
6. "POV: You just realized..." (relatable).
7. "$X to $Y in [time]" (results).
8. "The cheapest way to..." (utility).
9. "What nobody tells you about X" (curiosity).
10. "Watch this if you..." (specific audience).

6. Cadence — How Often to Test

Recommended

• 2-3 new creatives per week.
• Refresh winners every 2-4 weeks.
• Pause underperformers weekly.

Why

• Creative fatigue sets in 2-4 weeks.
• Same audience sees ads multiple times.
• Algorithm changes demand fresh content.

7. Scaling Winners

Vertical Scaling

• Increase budget on same creative.
• Watch CPA — if jumps, fatigue setting in.

Horizontal Scaling

• Same hook, different visuals.
• Same visual, different hooks.
• Same creative, different audiences.

Iterations

• Winner v2: same hook, new visual.
• Winner v3: same visual, new CTA.
• Continue until plateau.

8. Creative Diversity — Anti-Fatigue

Mix of formats

• 30% Static.
• 40% Video.
• 20% Carousel.
• 10% UGC.

Mix of styles

• Polished brand.
• UGC raw.
• Founder selfie.
• Demos.

No single style — algorithm favors diversity.

9. Testing Setup — Meta Example

Campaign Structure

Campaign: Sales (manual)
└── Ad Set 1: Test new creative
    └── 3-5 creative variations
└── Ad Set 2: Winners (scale)
    └── Top 1-2 from Ad Set 1
└── Ad Set 3: Retargeting

Daily Budget

• Test ad set: $50-150.
• Winners ad set: $200-2,000+.
• Total: depends on scale.

10. Metrics to Watch

Primary

• CPA / ROAS — final goal.
• CR (Conversion rate).

Secondary (Leading)

• CTR — engagement.
• Hook Rate (3-second view rate) — Meta.
• Hold Rate (15-second view rate).
• Engagement Rate (likes/comments).

11. Common Creative Mistakes

❌ Brand-first instead of value-first. ❌ Logo too big in opening. ❌ No subtitles — 60% mute. ❌ 9:16 unused — losing mobile reach. ❌ Same creative for months — fatigue. ❌ Too many CTAs — confusion. ❌ Stock footage — feels generic.

12. AI Creative Tools (2026)

• Midjourney / DALL-E — Static visuals.
• Runway / Pika — Video generation.
• HeyGen — AI avatars + voice.
• Canva / CapCut AI — Templates with AI.
• OpusClip — Long form → Shorts.

Disclaimer: AI-only creative looks generic. Use AI for ideation, human for final.

13. Israeli Creative Considerations

• Hebrew + English mix common.
• Subtitles in Hebrew essential.
• Cultural references resonate.
• Direct tone (sabra style) works.

14. אסיים בהמלצה.

קלט נדרש

פלט צפוי

כללי עבודה

• פלט בעברית. מונחים מקצועיים באנגלית.
• Test one variable at a time (or you don't know what won).
• Hook is everything — 3 seconds.
• Diversity prevents fatigue.

דגלים אדומים

• 🚨 Same creative 4+ weeks at scale — fatigue.
• 🚨 CTR < 0.5% — Hook broken.
• 🚨 Single format — limited reach.
• ⚠️ All polished, no UGC — feels like ads.

הערות חשובות

• Quantity > Quality in testing phase.
• Quality > Quantity in scaling phase.
• Document winners — patterns emerge.
• Spec ad (for influencer/UGC) = 3x cheaper than studio.

פרומפט לדוגמה

Meta Ads, ROAS ירד מ-3.5 ל-2.1. Creative testing plan.

SaaS B2B, רוצה לבדוק 5 hooks חדשים. איך?

עברנו על 50 creatives — אף אחד לא עובד. מה לבדוק עכשיו?

© 2026 כל הזכויות שמורות | CMO Online Israel Pro גרסה: 1.0.0 | עדכון אחרון: מאי 2026