20 ChatGPT Prompts for Writing Google Ads That Actually Convert

A generic prompt produces generic ads. ChatGPT doesn't know your offer, persona, competitive landscape, or Google Ads character limits (30 chars for headlines, 90 chars for descriptions). Without these elements, you get marketing filler that nobody clicks.

The 3 classic mistakes:

1. **No context**: "Write a Google Ads headline for a shoe store." Output: generic, zero differentiation. 2. **No constraints**: ChatGPT generates 45-character headlines. You waste time cutting. 3. **No angle**: no unique benefit, no urgency, no social proof.

The solution: a structured prompt built around 5 blocks.

CPROP = Context, Persona, Result, Output format, Proof.

**Base template:**

``` Context: [Your business, product/service, target market, competitive advantage] Persona: [Ideal customer: age, problem, purchase motivation] Result: Generate 10 Google Ads RSA headlines (max 30 chars each) and 4 descriptions (max 90 chars) Format: Table with columns "Headline" | "Char count" | "Angle" Proof: [Numbers, testimonials, certifications to integrate if possible] ```

This format forces ChatGPT to count characters (imperfectly, but much better than a free prompt) and vary angles.

**Category 1: Headline Variations**

Prompt 1: Price-focused headlines ``` You are a Google Ads expert. Generate 10 RSA headlines (max 30 chars) for [product/service at [price]]. Angles: price, savings, value for money. Format as table with character count. ```

Prompt 2: Urgency headlines ``` Generate 10 RSA headlines (max 30 chars) with urgency for [product]. Use: "Today Only", "Same-Day Delivery", "Limited Stock", "Offer Ends [date]". Table with char count. ```

Prompt 3: Functional benefit headlines ``` For [product], generate 10 RSA headlines (max 30 chars) focused on the main benefit: [benefit]. Avoid vague words like "quality" or "great". Table. ```

Prompt 4: Social proof headlines ``` Generate 10 RSA headlines (max 30 chars) integrating social proof: [X customers], [rating] stars, [award]. Table. ```

Prompt 5: Exact keyword in position 1 ``` Generate 10 RSA headlines starting EXACTLY with "[keyword]" (max 30 chars total). Keyword must be the first word. Table. ```

**Category 2: Descriptions**

Prompt 6: AIDA descriptions ``` Write 4 Google Ads descriptions (max 90 chars) following the AIDA model for [product]. Each description = 1 AIDA step. Table. ```

Prompt 7: CTA-focused descriptions ``` Write 4 descriptions (max 90 chars) for [product] with these CTAs: "Shop Now", "Get a Quote", "Start Free Trial", "Book Today". 1 description per CTA. Table. ```

Prompt 8: Objection-handling descriptions ``` My customers' 3 main objections before purchase: [objection 1], [objection 2], [objection 3]. Generate 3 Google Ads descriptions (max 90 chars) addressing each objection. Table. ```

**Category 3: Improving Existing Ads**

Prompt 9: Fix a low CTR ``` This Google Ad has a CTR of [X]%. Headline: "[current headline]". Description: "[current description]". Generate 5 improved variations (max 30 chars headlines, 90 chars descriptions). Table with the angle for each. ```

Prompt 10: Seasonal adaptation ``` Adapt these ads for [season/period]: [current headlines and descriptions]. Keep best-performing elements, add seasonal components. Table. ```

**Category 4: A/B Testing**

Prompt 11: A/B pairs ``` Generate 5 A/B headline pairs (max 30 chars) for [product]. Each pair tests 1 variable: rational vs emotional angle. Table: Headline A | Headline B | Variable tested. ```

Prompt 12: Value angle testing ``` For [product], generate 6 RSA headlines (max 30 chars) testing 3 angles: savings, speed, reliability. 2 headlines per angle. Table with angle identified. ```

**Category 5: Special Cases**

Prompt 13: RLSA (remarketing) ads ``` Generate 5 headlines (max 30 chars) and 2 descriptions (max 90 chars) for remarketing ads targeting visitors who viewed [product page] without purchasing. Tone: gentle urgency, reassuring proof. ```

Prompt 14: Competitor targeting ``` I'm targeting queries containing competitor name [name]. Generate 5 headlines (max 30 chars) and 2 descriptions (max 90 chars) that favorably compare [my product] without explicitly naming the competitor. ```

Prompt 15: Long-tail ads ``` For the long-tail query "[query]", generate 5 RSA headlines (max 30 chars) that hyper-specifically address this search intent. Table. ```

**Category 6: Advanced Optimization**

Prompt 16: Strategic pinning ``` I have these RSA headlines: [list]. Indicate which headlines to pin in positions 1, 2, 3 for maximum message consistency. Justify each choice. ```

Prompt 17: Ad-to-landing page consistency ``` My ad: [headline + description]. My landing page starts with: "[first paragraph]". Analyze message consistency and propose 3 ad adjustments to improve Quality Score. ```

Prompt 18: Multi-location ads ``` Adapt these ads for 3 cities: [city 1], [city 2], [city 3]. Include city name in each headline. Max 30 chars. Table: City | Adapted headline | Char count. ```

Prompt 19: Specific product ads ``` For product [exact name, SKU, price], generate 10 ultra-specific RSA headlines (max 30 chars). Include: [technical spec], [differentiating advantage], [price if relevant]. ```

Prompt 20: Full ad brief ``` Here's my complete brief: Product: []. Target: []. Advantage vs competitors: []. Proof: []. Monthly budget: []. Goal: [target CPA or ROAS]. Generate a complete RSA ad: 15 headlines (max 30 chars) + 4 descriptions (max 90 chars) + pinning recommendations. Table. ```

ChatGPT makes character counting errors. Always validate with an external tool (Google Ads Editor, or the character counter in the Google Ads interface).

**Validation checklist:**

1. **Length**: each headline ≤ 30 chars, each description ≤ 90 chars 2. **Relevance**: primary keyword appears in at least 1 headline 3. **Clarity**: a reader understands the offer in 3 seconds 4. **CTA**: at least 1 headline or description contains a call to action 5. **Differentiation**: the ad doesn't look like competitor #1's ad

Tip: paste your final ads into Claude or ChatGPT and ask "Play the role of a Google user searching for [keyword]. Would these ads make you want to click? Why?"

**Step 1: Brief (5 min)** Fill a template with product/persona/proof information. Save it in Notion or a Google Doc.

**Step 2: Generation (10 min)** Run prompts from categories 1 and 2. Generate 50-60 raw options.

**Step 3: Selection (10 min)** Keep the 15 best headlines and 4 best descriptions. Apply the validation checklist.

**Step 4: Integration and testing (5 min)** Import into Google Ads Editor. Check Ad Strength (aim for "Excellent"). Launch and monitor after 2 weeks.

Real time savings: this process takes 30 minutes vs 2-3 hours manually. Quality is comparable or better thanks to the volume of variations tested.