Ad copy A/B matrix — platform-specific constraints, testing framework, and conversion refinement

'You are a performance-marketing copywriter who optimizes for conversions, not cleverness. For my product: [PRODUCT + TARGET AUDIENCE + key benefit + price point if relevant], write ad copy for [META / GOOGLE / BOTH].

PLATFORM CONSTRAINTS (follow these exactly):

FOR META (Facebook/Instagram):
• Primary text: 125 characters before the "See More" truncation — the hook MUST land in those 125 characters. Write the full primary text (up to 250 chars) but front-load the punch.
• Headline: Under 40 characters (truncates on mobile)
• Description: Under 30 characters
• If this is an image ad: no more than 20% text in the image area (suggest text placement in the "safe zone" — center 60% of the image)

FOR GOOGLE RSA (Responsive Search Ads):
• Headlines: 30 characters max, write 10-15 (Google will mix and match)
• Descriptions: 90 characters max, write 4
• Pin the most important headline to Position 1
• Include the keyword in at least 3 headlines
• At least 2 headlines should work as standalone statements

Now generate using this TESTING MATRIX — don't just test copy, test the COMBINATION of:

HOOK TYPE (pick 4 to test):
□ Pain-point ("Tired of...")
□ Benefit-led ("Get X without Y")
□ Social proof ("Join 10,000+ who...")
□ Curiosity gap ("The X nobody talks about")
□ Urgency ("Last chance / Limited")
□ Objection-handling ("Think X is too expensive? Here's why...")
□ Contrarian ("Stop doing X")
□ Founder story ("I built this because...")

× CTA TYPE (pick 2 to test):
□ Direct ("Buy now" / "Sign up")
□ Low-commitment ("See how it works" / "Try free")
□ Curiosity ("See the difference" / "Find out")

× SOCIAL PROOF FORMAT (pick 2 to test):
□ Number-based ("50,000+ customers")
□ Testimonial snippet ("Changed my mornings — Sarah K.")
□ Authority ("Featured in Forbes")
□ Result ("Average 3x improvement")

Generate 8 complete ad variations — each with a DIFFERENT combination from the matrix above. Label each with its matrix combination: [Hook: X | CTA: Y | Proof: Z].

After the 8 variations, provide:

TESTING PLAN:
1. Which 2 ads to test FIRST (highest expected impact, most different from each other)
2. What to test NEXT based on which one wins
3. The minimum budget/impressions needed for statistical significance (rule of thumb: 100 conversions per variant, or 1000 clicks for CTR testing)

CONVERSION REFINEMENT CYCLE:
After I tell you which ad won, be ready to:
1. Analyze WHY it won (which matrix element drove it)
2. Generate 3 variations of the winner, changing only ONE element at a time
3. Repeat until we hit diminishing returns

Match this voice: [VOICE]. Keep claims honest — no guarantees we can't back up.'

Tips: the biggest ad testing mistake is testing whole ads against each other — you learn nothing about WHY one won; the matrix approach isolates variables; start with hook type testing (biggest impact), then refine CTA, then proof format; for Meta, always write mobile-first — 85%+ of impressions are mobile; add 'write the image text overlay too (under 8 words, in the safe zone)' for complete creative briefs.