Prompt Engineering for Marketing: 20 Copy Templates
Master prompt engineering for marketing copy with 20 tested templates. Ad headlines, email sequences, landing pages, social posts, and brand voice calibration.
Conversion Lift
CTR Improvement
Templates Included
Voice Accuracy
Key Takeaways
Every marketing team is using AI to generate copy. Most of them are getting mediocre results. The difference between teams that produce AI-generated copy indistinguishable from expert human writing and teams that produce obvious AI slop comes down to one skill: prompt engineering. Specifically, the ability to construct prompts that encode audience psychology, brand constraints, format requirements, and conversion principles into a repeatable system.
This guide provides 20 tested prompt templates that cover the core marketing copy needs: ad headlines for Google, Meta, and LinkedIn; email sequences for welcome, nurture, and win-back flows; landing page copy from hero sections to CTAs; social media posts optimized for each platform; and brand voice calibration prompts that ensure consistency across everything. Each template includes the complete prompt text, the variables you need to customize, and annotated examples showing the output quality you should expect.
These templates are not theoretical. They have been refined across hundreds of production campaigns and measured against real conversion data. The prompt structures work with Claude, ChatGPT, Gemini, and any other general-purpose LLM, because the principles of effective prompt engineering are model-agnostic.
Why Prompt Engineering Is the New Copywriting Skill
Traditional copywriting follows a process: research the audience, understand the offer, study the competition, draft multiple versions, test, and refine. Prompt engineering follows the exact same process, but instead of writing the copy yourself, you write the instructions that produce the copy. The skill shifts from sentence construction to strategic instruction design.
- Speed: A well-crafted prompt generates 10 headline variations in 15 seconds versus 2-3 hours of manual brainstorming for the same volume
- Consistency: Templates encode your brand voice and conversion principles, ensuring every output meets baseline quality regardless of who runs the prompt
- Scale: The same prompt template produces copy for 50 product variants, 10 audience segments, or 7 campaign themes without starting from scratch each time
- Testability: Prompt templates can be A/B tested against each other, creating a measurable improvement cycle for your copy quality over time
- Knowledge transfer: A prompt template captures expert copywriting knowledge in a format that junior team members can use to produce senior-level output
The critical insight is that prompt quality determines output quality far more than model selection. A well-structured prompt on Claude Sonnet consistently outperforms a generic prompt on Claude Opus. The model provides the capability. The prompt provides the direction. Without direction, even the most capable model produces average copy, because it defaults to the statistical average of all the marketing content in its training data. Your prompts are what pull the output above that average toward excellence.
This is why professional content marketing teams are investing in prompt libraries the same way they once invested in copywriting style guides. The prompt library becomes the codified expertise of the organization, capturing what works, why it works, and how to reproduce it at scale.
Ad Headline Templates (Google, Meta, LinkedIn)
Ad headlines are the highest-leverage copy in your marketing stack. A 1% improvement in headline CTR compounds across every impression, potentially worth thousands of dollars in improved campaign performance. The following five templates cover the major ad platforms and headline types you will use most frequently.
## Context
Product/Service: [name and one-sentence description]
Target Keyword: [primary search term]
Audience Pain Point: [what problem are they searching to solve]
Unique Value Prop: [why choose this over competitors]
## Task
Generate 15 Google Ads headlines. Each headline must be under 30 characters. Include:
- 5 headlines featuring the target keyword naturally
- 3 headlines emphasizing the unique value proposition
- 3 headlines with specific numbers or statistics
- 2 headlines with urgency or scarcity
- 2 headlines as direct questions
## Constraints
- Strictly under 30 characters per headline (count every character)
- No exclamation marks (Google policy)
- No ALL CAPS words
- Each headline must work independently (RSA combinations are unpredictable)
- Include character count next to each headline
Example output (for a project management tool):
- "Project Management Made Easy" (28 chars)
- "Ship Projects 40% Faster" (24 chars)
- "Where Teams Get Things Done" (27 chars)
- "Free Trial - Start Today" (24 chars)
- "Why Do Teams Keep Missing?" (26 chars)
## Context
Product/Service: [name and description]
Target Audience: [demographics, interests, behaviors]
Campaign Objective: [awareness/consideration/conversion]
Offer: [what the user gets if they click]
## Task
Write 5 variations of Meta ad primary text (the text above the image). Each variation should use a different hook strategy:
Variation 1: Problem agitation (describe the pain, twist it, present the solution)
Variation 2: Social proof lead (start with a result or testimonial)
Variation 3: Curiosity gap (reveal just enough to compel a click)
Variation 4: Direct benefit (lead with the most compelling outcome)
Variation 5: Story-based (micro-narrative in 3 sentences)
## Format
- First line: Hook (this shows above the "See more" fold, max 125 characters)
- Body: 2-3 short sentences expanding on the hook
- CTA: Clear next step (max 1 sentence)
- Total length: 70-150 words per variation
## Context
Company: [name and industry]
Content Asset: [whitepaper/webinar/case study/demo]
Target Titles: [job titles of ideal audience]
Industry Pain: [the business problem they face]
## Task
Generate 10 LinkedIn ad headlines (max 70 characters each) and 5 introductory text variants (max 150 words each).
Headlines should follow these proven LinkedIn patterns:
- "How [Company Type] [Achieved Result]"
- "[Number] [Professionals] Are [Doing X]"
- "The [Adjective] Guide to [Outcome]"
- "Stop [Common Mistake]. Start [Better Approach]."
- "What [Title] Need to Know About [Topic]"
## Tone
Professional but not corporate. Data-driven but accessible. Avoid buzzwords: "synergy," "leverage," "revolutionize," "game-changing." Use specific language that B2B decision-makers use in their own conversations.
## Context
Product: [name]
Retargeting Segment: [visited pricing page / added to cart / read blog post / watched video]
Days Since Visit: [1-3 / 4-7 / 8-14 / 15-30]
Likely Objection: [price / timing / trust / complexity]
## Task
Write retargeting ad copy that addresses the specific objection for this audience segment. The copy must acknowledge they have already seen the product without being creepy.
Structure:
- Hook: Reference the benefit they were exploring (not "we noticed you visited")
- Objection handler: Address the likely reason they did not convert
- Social proof: One specific result or customer reference
- CTA: Low-friction next step appropriate to their stage
Write 3 variations: short (under 50 words), medium (50-100 words), long (100-150 words).
## Context
Current headline: [paste your existing best performer]
Current CTR: [percentage]
Platform: [Google/Meta/LinkedIn]
Target improvement: [what specific metric to improve]
## Task
Analyze the current headline and generate 10 test variants. For each variant, explain:
1. What psychological lever it uses (curiosity, fear, authority, social proof, urgency)
2. What element changed from the original
3. Why this change might improve the target metric
Group the variants:
- 3 minor tweaks (same structure, different words)
- 4 moderate changes (different angle, same message)
- 3 radical departures (completely different approach)
The pattern across all five ad templates is the same: specific context, structured output requirements, and explicit constraints. Generic prompts like "write me 10 headlines for my SaaS product" give the AI no information about your audience, platform limitations, or conversion goals. The templates above force you to think through these elements before generating, which is why they produce better results.
Email Sequence Templates (Welcome, Nurture, Win-Back)
Email sequences are where prompt engineering delivers its highest ROI. A single well-designed prompt can generate an entire 5-7 email sequence that would take a copywriter a full day to write manually. The key is encoding the narrative arc, progressive disclosure of information, and escalating CTAs into the prompt structure.
## Context
Business: [name, industry, core offer]
Sign-up Source: [what the subscriber signed up for: lead magnet / newsletter / free trial / purchase]
Primary CTA Goal: [what action the sequence should drive]
Brand Voice: [3-5 adjectives + 1 example email if available]
## Sequence Structure
Email 1 (Day 0): Welcome + deliver promised value. Set expectations for what emails they will receive. Include one quick win. Subject line: personal, not promotional.
Email 2 (Day 2): Your origin story or mission in 3 paragraphs. Build emotional connection. End with a question that invites replies.
Email 3 (Day 4): Your best piece of free content. Frame it as solving their biggest pain point. Social proof: how many others found this valuable.
Email 4 (Day 7): Case study or testimonial. Specific numbers and timeline. Subtle introduction of your paid offer.
Email 5 (Day 10): Direct pitch for primary CTA. Recap the value they have received. Clear offer with deadline or scarcity if applicable.
## For Each Email
- 3 subject line options (under 50 characters each)
- Preview text (under 90 characters)
- Full body copy (150-300 words)
- One clear CTA per email
- P.S. line where appropriate (highest-read section after the subject line)
## Context
Business: [name and what you sell]
Audience Segment: [who these leads are and where they are in the funnel]
Common Objections: [top 3 reasons they do not buy]
Competitive Alternatives: [what else they might choose]
## Sequence Logic
Email 1: Educational content addressing their #1 pain point (no pitch)
Email 2: Framework or methodology that positions your approach as unique
Email 3: Address Objection #1 with data and a case study
Email 4: Comparison with alternatives (fair but favorable)
Email 5: Address Objection #2 with testimonial
Email 6: Address Objection #3 with guarantee or risk reversal
Email 7: Direct CTA with recap of all value delivered across the sequence
## Key Constraint
Each email must deliver standalone value even if the reader skipped previous emails. Do not assume sequential reading. Each email should be self-contained while building on the overall narrative for those who read all.
## Context
Business: [name]
Inactive Period: [how long since their last engagement]
Previous Relationship: [what they purchased / used / engaged with]
What Changed: [new features / improvements / offers since they left]
## Sequence
Email 1 (Re-engagement): "We miss you" without being desperate. Highlight 1-2 major improvements since they left. Ask one simple question to re-engage. Subject line should create curiosity about what is new.
Email 2 (Value reminder, 3 days later): Share a specific success story from a customer who returned. Include a limited-time incentive (discount, extended trial, bonus). Create FOMO without being manipulative.
Email 3 (Last call, 5 days later): Direct and honest. "Should we keep sending you emails?" Give them a clear opt-out alongside a final offer. Respect their attention.
## Tone
Human, not corporate. Acknowledge the gap without guilt. Focus on value, not obligation.
## Context
Email Purpose: [what this email contains/promotes]
Audience: [who is receiving this]
Previous Best Performers: [paste your top 3 subject lines with their open rates]
## Task
Generate 15 subject line options in these categories:
- 3 curiosity-based (open loops that compel clicking)
- 3 benefit-led (immediate value statement)
- 3 question-based (relevant questions they want answered)
- 3 personalized (uses name or segment data)
- 3 urgency/scarcity (time or quantity limited)
## Constraints
- Under 50 characters (optimal for mobile preview)
- No spam trigger words (free, guarantee, act now, limited time)
- No ALL CAPS
- Include preview text suggestion for each (under 90 chars)
- Rate each 1-5 on curiosity, clarity, and urgency
The email templates above produce complete, production-ready sequences. The key technical insight is encoding the narrative arc into the prompt. Without explicit sequence structure (Email 1 does X, Email 2 does Y), the AI generates 5-7 emails that all sound the same and make the same points. The structured approach ensures each email has a distinct job within the conversion journey.
Landing Page Copy Templates
Landing page copy follows a predictable structure: hero section, social proof, feature-benefit blocks, objection handling, testimonials, and CTA. The prompt templates below generate each section with the conversion principles already encoded.
## Context
Product/Service: [description]
Primary Audience: [who is landing on this page]
Traffic Source: [Google Ads / social / email / organic]
Primary Conversion Goal: [sign up / buy / book demo / download]
#1 Benefit: [the single most compelling outcome]
## Generate
5 hero section variants, each containing:
1. Headline (max 10 words): Communicate the primary benefit or transformation. Not the product name. Not a feature. The outcome the reader wants.
2. Subheadline (max 25 words): Expand on the headline with specifics. Include who it is for and what makes it different.
3. CTA button text (max 5 words): Action-oriented, specific to the conversion goal. Not "Learn More" or "Submit."
4. Supporting text under CTA (max 15 words): Reduce friction. "No credit card required" or "Free for teams under 10"
## Think Step by Step
Before writing each variant, briefly state: (a) what psychological angle this variant uses, (b) what objection it preemptively addresses, (c) why this approach might work for this audience.
## Context
Product Features: [list your top 6 features with brief descriptions]
Target Audience: [who and what they care about]
## Task
For each feature, generate a feature-benefit block:
1. Feature headline (5-8 words): Lead with the benefit, not the feature name. "Send invoices in 30 seconds" not "Invoice automation."
2. Supporting paragraph (40-60 words): Explain how the feature works and why the user should care. Use second person ("you"). Include one specific detail that builds credibility.
3. Proof point (1 sentence): A statistic, customer quote, or specific result tied to this feature.
## Rule
Every feature block must answer the reader's implicit question: "So what? Why should I care?" If the block does not answer this, rewrite it.
## Context
Product: [name]
Top 5 Objections (ranked by frequency):
1. [objection]
2. [objection]
3. [objection]
4. [objection]
5. [objection]
## Task
For each objection, write:
- Objection statement (phrased as the customer would say it)
- Response (3-4 sentences): Acknowledge the concern, reframe it, provide evidence that addresses it, end with confidence
- Supporting proof: One specific data point, testimonial, or guarantee that neutralizes this objection
Format as an FAQ-style section or an "Addressing your concerns" block. Tone: empathetic but confident. Never dismissive.
## Task
Assemble a complete landing page using the following sections in order:
1. Hero (headline, subheadline, CTA, friction reducer)
2. Social proof bar (logos, numbers, or one-line testimonial)
3. Problem statement (3-4 sentences describing the pain)
4. Solution overview (3-4 sentences on your approach)
5. Feature-benefit blocks (top 4 features)
6. Testimonial block (3 customer quotes with names and titles)
7. Objection handler (top 3 concerns addressed)
8. Final CTA section (different headline from hero, same button)
## Important
The page must tell a coherent story from top to bottom. Each section should flow naturally into the next. The reader should feel progressively more convinced as they scroll. Do not repeat the same points across sections.
The landing page templates work best when used together. Start with Template 10 to generate hero variants, Template 11 for features, Template 12 for objections, and then Template 13 to assemble everything into a cohesive page. This modular approach lets you optimize individual sections without regenerating the entire page.
Brand Voice Calibration Prompts
Brand voice consistency is the biggest challenge in AI-generated marketing copy. Without calibration, AI produces content that sounds generically professional but lacks the specific personality, vocabulary, and rhythm that makes a brand recognizable. These three templates solve this problem.
## Input
Here are 5 examples of our brand's best writing:
[Paste 5 pieces of copy: website, emails, ads, social]
## Task
Analyze these examples and create a Brand Voice Guide that documents:
1. Voice Attributes (5 adjectives with definitions and examples)
2. Vocabulary Preferences (words we use vs. words we avoid)
3. Sentence Structure Patterns (average length, complexity, active vs. passive)
4. Tone Spectrum (formal to casual scale, where we sit and when we shift)
5. Signature Phrases (recurring patterns or structures that define our voice)
6. Anti-Patterns (things our brand would never say or do)
Format this as a reference document I can paste into future prompts as brand context.
## System Context
[Paste brand voice guide from Template 18]
## Input
Here is a draft piece of marketing copy:
[Paste AI-generated or human-written draft]
## Task
Rewrite this copy to match our brand voice exactly. Make the following adjustments:
- Replace any vocabulary that conflicts with our preferences
- Adjust sentence structure to match our patterns
- Ensure the tone sits at the right point on our spectrum
- Add any signature phrases or patterns where natural
- Remove any anti-patterns
After rewriting, highlight the 3 most significant changes you made and explain why each change better aligns with the brand voice guide.
## System Context
[Paste brand voice guide from Template 18]
## Input
Here are 10 pieces of copy we plan to publish this week:
[Paste all 10 pieces]
## Task
Evaluate each piece against our brand voice guide. For each:
- Score 1-10 on voice consistency
- Flag any off-brand phrases or patterns
- Suggest specific rewrites for flagged sections
- Note any pieces that are particularly strong examples of our voice (for future reference material)
Summarize: overall consistency score, most common deviations, and 3 recommendations for improving voice consistency in future generation prompts.
Advanced Techniques: Chain-of-Thought and Few-Shot
The 20 templates above use two advanced prompt engineering techniques that significantly improve output quality: chain-of-thought (CoT) reasoning and few-shot learning. These are not abstract concepts. They are specific structural patterns you can add to any prompt to immediately improve results.
CoT prompting asks the AI to reason through a problem before producing output. For marketing copy, this means reasoning through audience psychology, competitive positioning, and emotional triggers before writing.
## Before writing:
1. Identify the primary emotion this audience feels about [topic]
2. List the top 3 objections they would have to [offer]
3. Determine the strongest proof point available
4. Choose the hook type that best addresses emotion + objection
## Then write:
[The actual copy request]
Result: 28% higher CTR on ad headlines vs. direct generation
Few-shot learning provides examples of desired output within the prompt. For marketing copy, include 3-5 examples of your best-performing content to anchor the AI's output style and quality.
## Examples of our best copy:
Example 1: [paste full text]
Why it worked: [1 sentence]
Example 2: [paste full text]
Why it worked: [1 sentence]
Example 3: [paste full text]
Why it worked: [1 sentence]
## Now generate:
[Your copy request, matching the quality and style of these examples]
Result: 90%+ brand voice consistency with 5 examples
You can combine both techniques in a single prompt. Start with few-shot examples to establish the quality bar and style, then add chain-of-thought instructions to ensure the AI reasons through the strategic context before writing. This combined approach produces the highest-quality output across all 20 templates.
## Brand Voice (from Template 18 output)
[Paste your brand voice guide]
## Examples of Our Best [Content Type]
[Paste 3-5 examples with performance context]
## Reasoning (complete before writing)
1. What is the core message we need to communicate?
2. What emotional state is our audience in when they encounter this?
3. What objection or hesitation needs to be addressed?
4. Which example above is closest to what we need? Why?
5. What specific angle will make this piece stand out?
## Generate
[Your specific copy request with format constraints]
## Self-Review
After writing, check: Does this match our brand voice? Does it address the right emotion? Is it as specific and compelling as our examples? If not, revise before presenting.
The self-review step at the end is critical. It acts as a built-in quality gate that catches generic output before it reaches you. The AI compares its own output against the quality bar set by your examples and revises if it falls short. This single addition reduces the amount of human editing required by approximately 40%.
Measuring Prompt Performance
Prompt templates are not static. They should evolve based on performance data, the same way ad creative and landing pages are continuously tested and refined. The measurement system below turns your prompt library from a collection of templates into a self-improving engine that produces better copy every month.
Tag every prompt template with a version number (headline-v1.0, welcome-email-v2.3). Track which version produced each piece of published copy. Store prompt versions in a shared document or prompt management tool.
Connect each prompt version to its downstream performance metrics. Ad headlines map to CTR. Email templates map to open rates and conversion rates. Landing page templates map to bounce rate and conversion rate. Social templates map to engagement rate.
Each month, compare prompt versions against each other. Did headline-v2.1 produce higher CTR than headline-v2.0? If yes, promote v2.1 to the default. If no, analyze what changed and revert or iterate. Maintain a changelog of what was modified in each version.
Feed performance data back into the prompts themselves. Update your few-shot examples to use your current top-performing copy. Adjust constraints based on what format characteristics correlate with higher performance. This creates a compounding improvement cycle.
- Click-through ratePrimary
- Quality scoreSecondary
- Conversion ratePrimary
- Cost per clickSecondary
- Open ratePrimary
- Click ratePrimary
- Reply rateSecondary
- Unsubscribe rateSecondary
- Engagement ratePrimary
- Save/bookmark ratePrimary
- Share rateSecondary
- Follower growthSecondary
The teams that get the best results from AI-generated marketing copy are not the ones with the best initial prompts. They are the ones with the best measurement and iteration systems. A mediocre prompt that gets refined monthly based on performance data will outperform a brilliant prompt that never changes within 3-4 months. This is the same principle that drives all effective content marketing strategy: measure, learn, iterate.
Elevate Your Marketing Copy
Our team builds prompt engineering systems that produce high-converting marketing copy at scale. From ad headlines to full landing pages, we help you turn AI into your competitive advantage.
Related Guides
Continue exploring these insights and strategies.
Social Media Post Templates by Platform
Social media copy has the strictest format constraints and the highest volume demands. These templates are designed for batch generation: run each one to produce a week's worth of platform-specific content in a single session.
Write 5 LinkedIn posts for [week theme].
Each post: 1,300-1,800 characters. Start with a hook that challenges conventional wisdom. Body: share a specific experience, lesson, or data point. End with a question that invites discussion.
Format: Short paragraphs (1-3 sentences). Use line breaks generously. No hashtags in the body; add 3-5 at the end.
Tone: [brand voice]. Write as a practitioner sharing real experience, not as a brand promoting itself.
Write 5 Instagram carousel captions for educational content about [topics].
Each caption: 150-300 words. First line is the hook (shows in feed). Include a micro-story or specific example. End with CTA: "Save this for later" or "Share with someone who needs this."
Also provide: slide-by-slide text outline for each (10 slides: cover, 8 content, CTA slide). Max 25 words per slide.
Hashtags: 3-5 relevant tags per post. No generic tags (#business #success).
Write 3 X threads on [topics]. Each thread: 8-12 tweets.
Tweet 1: Bold, counterintuitive hook under 200 chars.
Tweets 2-10: One insight per tweet. Use specific examples, numbers, comparisons. Each tweet delivers standalone value.
Final tweet: Summary + CTA (follow, retweet, reply).
Each tweet strictly under 280 characters. Use numbered format (1/, 2/). No hashtags within thread. Write like a knowledgeable human sharing expertise, not a brand account.
Write 5 Facebook posts designed to maximize comments and shares for [topics].
Each post: 50-150 words. Conversational tone. Open with something relatable (shared experience, common frustration, universal truth).
Post types to include:
- 1 poll/question post
- 1 behind-the-scenes story
- 1 hot take / opinion
- 1 tip/hack with image concept
- 1 celebration/milestone
No external links (algorithm penalty). Use native features (polls, images) over links.
The batch generation approach across all four social templates produces a full week of multi-platform content in 30-45 minutes. Run Template 14 for LinkedIn, Template 15 for Instagram, Template 16 for X, and Template 17 for Facebook. Each template produces 3-5 posts, giving you 12-20 pieces of platform-optimized content per batch session. This is the same methodology used in professional social media management operations.