13 Claude recipes for Facebook + Google Ads marketers

You are a senior direct-response copywriter for a B2B SaaS brand.
Brand voice samples (do not mimic exactly; match cadence and word choice):
<paste 5 of your existing high-performing organic posts>

Product: <one-line value prop>
ICP: <one-line audience>
Top 3 pain points: <list>
Top 3 outcomes the product delivers: <list>

Write 10 Facebook primary-text variants for a cold-traffic campaign.
Mix the formats evenly:
- 3 short hooks (max 80 chars, designed to stop scroll)
- 4 problem-agitate-solve (max 200 chars)
- 3 social-proof openers (max 150 chars)

Constraints:
- No emoji
- No clickbait phrasing ("you wont believe...")
- No unverifiable health/financial claims
- One clear CTA per variant

Output each variant numbered, with character count in brackets.

1. [76] Stop paying $1.72 average CPC on Meta to copy that does not convert your ICP.
2. [73] Your top competitor just shipped 12 ad variants this week. You shipped 1.
3. [194] Most B2B founders rewrite ad copy weekly. Top performers test 10 variants per week, kill 7, scale 3. Heres the system. CTA: Start free.
...
[10 numbered variants follow]

Offer: <product or service, one line>
Price: <exact price or pricing model>
ICP: <audience>
Top objection from current customers: <one line>

Write 5 distinct Facebook ad angles for this offer:
1. Pain-based (lead with the most expensive cost of the problem)
2. Identity-based (lead with who the buyer becomes after)
3. Social-proof (lead with the strongest customer line)
4. Contrarian (lead by attacking the dominant industry wisdom)
5. FOMO (lead with the time-bound reason to act now)

For each angle, write:
- 1 primary text (max 200 chars)
- 1 headline (max 40 chars)
- 1 description (max 30 chars)
- 1 image brief (one sentence for the visual concept)

Angle 1: Pain-based
Primary: "Every week without a content system costs you 12 buying-window leads..."
Headline: "Stop losing leads to silence"
Description: "Get the system"
Image brief: "Founder at laptop, calendar empty, late-evening glow."
...
[Angles 2-5 follow same shape]

Build a Google Responsive Search Ad asset set for a paid campaign.

Product / service: <one-liner>
Target keyword cluster: <main keywords>
Final URL: <landing page url>
Top customer outcome: <one line>
Top objection: <one line>

Generate:
- 20 headlines (max 30 chars each), mix of:
  - 4 with the exact keyword
  - 4 with a benefit-led claim
  - 4 with social proof or numbers
  - 4 with a CTA
  - 4 with a pain-point opener
- 4 descriptions (max 90 chars each), each pairing one outcome with one CTA

Constraints:
- No superlatives ("best", "#1") that breach Google trademark policy
- No "free" claims unless the offer is genuinely free
- No dynamic-keyword-insertion syntax

Return as a numbered table.

Headlines
1. [28] Brand Brain for Solo Founders
2. [29] 7-Day Trial, $1, Cancel Anytime
3. [27] 4x Faster LinkedIn Output
...
Descriptions
1. [86] Cancel anytime. Brand voice in 5 minutes. Trusted by 500+ founders. Start free trial.
...

Paste the full product page copy (hero, features, pricing, FAQ) between the markers.

<<<PRODUCT PAGE START>>>
<paste your full product page text here>
<<<PRODUCT PAGE END>>>

From this page, extract:
1. The single strongest pain point (one line)
2. The single strongest outcome (one line)
3. The 3 most credible proof elements (testimonials, numbers, brand mentions)
4. The price and trial terms

Then write a Meta ad creative brief that includes:
- Audience hypothesis (interests, behaviors, lookalikes)
- 3 visual concepts (one sentence each)
- 3 primary-text variants (max 200 chars)
- 3 headline variants (max 40 chars)
- A 15-second UGC video script (3 shots)
- A 6-second hook-only video script (1 shot)

Reference the verified Meta CPC benchmarks (WordStream): all-industry average CPC $1.72, Finance & Insurance $3.77, Apparel $0.45.

Pain: "Founders spend 4 hours per week posting and still get under 200 views..."
Outcome: "Ship 5 LinkedIn posts a week in 30 minutes, in your voice."
Audience hypothesis: Lookalike of last 90 days trial activations, layered with...
[3 visual concepts, 3 primary text, 3 headlines, 2 video scripts follow]

Write a 30-second 5-shot UGC ad script for a product I will describe.

Product: <one line>
Spokesperson: <persona, e.g. "27-year-old founder, gym clothes, kitchen background">
Hook angle: <pain, identity, or proof>
Single CTA: <one verb + benefit>

Shot structure (each shot has on-camera dialog + caption overlay + 1 visual cue):
1. [0-3s] Hook (must stop scroll in 1 second)
2. [3-9s] Pain or problem framing
3. [9-18s] Product demonstration
4. [18-25s] Proof or transformation
5. [25-30s] CTA

Constraints:
- No script lines longer than 12 words (UGC pacing)
- Caption overlays must be safe for Meta policy (no health/financial claims)
- Final CTA must work whether the viewer has sound on or off

Reference Meta CPC range: Apparel $0.45, all-industry average $1.72, Finance $3.77. Match the energy of the cheapest CPC vertical (Apparel) for hook intensity.

Shot 1 (0-3s): ON CAMERA: "I post 5 LinkedIn posts a week in 30 minutes."
CAPTION: "5 posts. 30 min. Same voice."
VISUAL: Kitchen counter, phone propped, founder mid-bite of toast.
...
[Shots 2-5 follow]

Calculate the break-even CPC for a paid campaign.

Inputs:
- Landing page conversion rate (visit to lead): <percent>
- Lead to customer conversion rate: <percent>
- Average revenue per customer (LTV): <dollars>
- Target customer acquisition cost as percent of LTV: <percent, default 30>
- Industry CPC benchmark (Databox / WordStream): <dollars; default $1.27 median across industries>

Walk me through:
1. Expected CAC at break-even
2. Maximum allowable CPC at my conversion rates
3. Comparison to industry CPC benchmark (am I above or below median)
4. Three concrete moves to lower CAC if I am above benchmark (rank by speed)

Cite the WordStream / Databox numbers I gave you. Do not invent additional benchmarks.

Break-even CAC: $150 (30% of $500 LTV)
Conversion chain: 5% landing-page conv * 20% lead-to-customer = 1% visit-to-customer
Maximum allowable CPC: $1.50
Industry CPC median: $1.27 (Databox 2023)
Verdict: You can pay up to $1.50 CPC profitably; you are at par with industry median.
3 moves to lower CAC: 1) Tighten landing-page CTA; 2) ...; 3) ...

Original ad (paste verbatim):
<paste original primary text + headline + description>

Why it is losing (one line): <e.g. CTR dropped 40% in 14 days>
Audience: <one line>
Offer: <unchanged from original>

Produce 6 rewrites that keep the offer identical but change:
- Variant A: Same hook structure, swap proof element
- Variant B: Swap from question to declarative
- Variant C: Swap from feature lead to outcome lead
- Variant D: Swap to contrarian frame
- Variant E: Swap to identity-based frame
- Variant F: Swap to time-bound urgency

For each variant: primary text (max 200 chars), headline (max 40 chars), description (max 30 chars).
End with a one-line recommendation on which variant to test first and why.

Variant A: Same hook, swapped proof:
Primary: "Most founders post 1x/week. Our users ship 5..."
Headline: "5 posts per week in 30 min"
Description: "Start free"
...
[Variants B-F follow]
Recommendation: Test Variant D first; the contrarian frame outperforms when CTR drops 40%+ on fatigue.

Paste your product page copy and your target keyword cluster.

<<<PRODUCT PAGE START>>>
<paste full product page>
<<<PRODUCT PAGE END>>>

Target keywords: <comma-separated list>

Build a Google Ads negative-keyword list (50+ terms) covering:
- Free-intent searchers (e.g. "free", "torrent", "cracked")
- Wrong-job-function searchers (if B2B SaaS, exclude students, hobbyists)
- Wrong-stage searchers (e.g. "review", "complaint", "lawsuit")
- Adjacent products you do NOT sell (e.g. if you sell Brand Brain not Buffer, exclude Buffer-specific queries)
- Off-platform terms (e.g. job titles, education, internships)

Output as a single comma-separated list, alphabetized.
Then add a 5-line rationale block explaining the categories of exclusions.

alternative free, buffer pricing, buffer reviews, buffer vs, careers, complaint, cracked, ...
[50+ comma-separated terms]

Rationale:
- Free-intent exclusions: 12 terms
- Adjacent-product exclusions: 14 terms (Buffer, Hootsuite, Loomly)
- Off-platform exclusions: 8 terms
- ...

Paste 30-day ad performance data in this exact format (one ad per line):

ad_name,spend,impressions,clicks,ctr,cpc,leads,cost_per_lead
<paste data>

Industry CPC benchmark (Databox May 2023):
- eCommerce $0.38, Apparel $0.54, Real Estate $0.70, Technology $1.30, SaaS $1.40, Consulting $1.76, Education $2.91, Health Care $3.08, all-industry median $1.27.

Industry CTR benchmark (WordStream Facebook 2017 baseline):
- all-industry average 0.90%, Legal 1.61%, Retail 1.59%, Apparel 1.24%.

Tell me:
1. Which 3 ads to KILL (with the deciding metric and why)
2. Which 3 ads to HOLD (give them another 7 days, deciding metric)
3. Which 3 ads to SCALE (the deciding metric and recommended new daily budget)
4. The 3 new angles I should test based on what is winning

Cite the benchmark numbers above when justifying decisions. Do not invent additional benchmarks.

KILL:
1. ad_summer_v3: CPC $3.21 vs your category benchmark $1.40 (SaaS). 2.3x over. Spend wasted.
2. ...

HOLD:
1. ad_pain_point_v2: CTR 0.74% (below 0.90% all-industry avg) but cost-per-lead within target. Give 7 days.
...

SCALE:
1. ad_proof_v5: CTR 1.92% (2.1x Apparel benchmark of 1.24%), CPC $0.51 (below your category $1.40). Double daily budget.
...

New angles:
1. Contrarian "you do not need 50 ad variants" (proof_v5 won on contrarian framing)
2. ...

Build a 5-email re-engagement sequence for paid leads who downloaded our <lead magnet> and have not booked a call or upgraded in 14+ days.

Product: <one-line value prop>
Original lead magnet topic: <one line>
Cost per lead (last 30 days): <dollars from your account>

Industry-standard cost-per-lead reference: SaaS Google Ads CPC $1.40, with a 3-5% landing page conversion gives roughly $28-$47 per lead. Anything materially above that is worth re-engaging.

Sequence shape (one email per day for 5 days):
Day 1: Pattern-interrupt subject line + the contrarian frame from <lead magnet>
Day 2: Customer story (specific name, outcome, time period)
Day 3: Common objection + reframe
Day 4: Show vs tell (link to a free tool or demo)
Day 5: Hard close ("if not now, when, and what will change")

For each email:
- Subject (max 60 chars)
- Preview text (max 90 chars)
- Body (under 150 words)
- Single CTA
- Send time (use IST timezone if I do not specify)

Day 1
Subject: "Most ad copy advice is wrong"
Preview: "The 3 reframes that work in 2026"
Body: ...
CTA: "Open the framework"
...
[Days 2-5 follow]

Stress-test this offer.

Offer: <full offer, including price, deliverables, guarantee>
ICP: <audience>
Price point: <exact dollar amount>

Industry benchmark context:
- Facebook all-industry CPC average $1.72 (WordStream)
- Google all-industry CPC median $1.27 (Databox)
- Average CAC at break-even on $500 LTV with 1% visit-to-customer: $5

Predict the 10 most likely objections this ICP will raise, ranked by frequency. For each objection:
1. Verbatim phrasing the prospect uses
2. Underlying fear or belief driving it
3. The shortest ad-copy line that pre-empts it
4. The landing-page section that should address it
5. Whether this is a "kill" objection (lose the sale) or "delay" objection (push to next quarter)

End with the 3 objections you would address in the ad itself vs the 7 you would push to the landing page.

Objection 1: "I tried Buffer/Hootsuite, this is the same thing."
Fear: Wasted previous SaaS spend.
Pre-empt line: "Built for solo founders who fired Buffer."
LP section: Hero
Type: Kill
...
[Objections 2-10 follow]
Ad-copy objections: 1, 4, 7. Landing-page objections: 2, 3, 5, 6, 8, 9, 10.

Seed keyword: <one phrase>
Product / service: <one-liner>
Locations: <countries / states>
Budget per month: <dollars>

Build 30 Google Ads ad-group themes covering:
- 6 informational ("what is X", "how does Y work")
- 6 comparison ("X vs Y", "alternatives to Z")
- 6 commercial-intent ("best X for Y", "X pricing")
- 6 transactional ("buy X", "X free trial", "X demo")
- 6 brand-defense ("[your brand] reviews", "[your brand] vs [competitor]")

For each theme:
- Ad group name
- 3 keywords (broad match modifier, phrase match, exact match) using current Google Ads match-type conventions
- 1 headline matching that intent
- 1 negative-keyword to exclude

Constraints:
- No keywords overlap across groups (avoid self-cannibalization)
- All keywords meet Google Ads minimum 1 word
- No exact-match keywords longer than 6 words (Google quality-score guidance)

Group 1 (Informational): "what is brand voice ai"
Keywords: brand voice ai, "ai brand voice", [what is brand voice ai]
Headline: "Brand Voice AI Explained"
Negative: "career"
...
[Groups 2-30 follow]

Paste a structured summary of competitor ads pulled from the Meta Ads Library (facebook.com/ads/library).

For each competitor ad, give me:
- Headline / primary text (verbatim)
- Visual type (UGC, product shot, motion graphic, talking head)
- Days running
- Apparent CTA
- Apparent funnel destination

<<<COMPETITOR ADS START>>>
<paste 5-10 competitor ads in the format above>
<<<COMPETITOR ADS END>>>

Then:
1. Identify the 3 angles the competitor is testing
2. Identify which 1 angle is the winner (deciding clue: longest days-running)
3. Identify the 3 angles they are NOT testing (gaps you can exploit)
4. For each gap angle, draft 1 primary text (max 200 chars) and 1 headline (max 40 chars) you could run
5. Flag any policy or claim risk in their copy you should avoid

Cite Meta Ads Library URL: facebook.com/ads/library.

Angles competitor is testing:
1. Pain-based (3 ads, all running 7-14 days)
2. Social-proof (2 ads, longest one 32 days; winner)
3. Identity-based (1 ad, 4 days; too new to judge)

Gap angles to exploit:
1. Contrarian (no competitor running this)
   Primary text: ...
   Headline: ...
...
Policy risk flag: Competitor ad 4 includes "guaranteed results"; do NOT copy this phrasing; Meta policy violation risk.