45 ChatGPT Prompts for Marketing That Actually Convert

Marketing is the discipline where prompt quality shows up fastest, because the output ships. A vague ask gets you the same beige copy your competitors are publishing. A prompt loaded with real context — your audience, your offer, your objections, your proof — gets you a draft that sounds like it was written by someone who actually knows the business.

That single difference is what this pack is built around. The 45 prompts below are grouped by job: positioning, ad copy, email, SEO briefs, social, landing pages, and reporting. Each one has placeholders in [brackets]. The prompts work as-is, but they get dramatically better the moment you replace those brackets with specifics instead of generics.

Why context beats clever phrasing

Most "best ChatGPT prompts" lists hand you a clever one-liner and promise magic. The magic isn't in the phrasing — it's in what the model knows when it starts writing.

Weak: Write me a Facebook ad for my software product.

Strong: Write 3 Facebook ad variations for [product], a [one-line description]. Audience: [specific role/segment] who currently [painful status quo]. Our edge over [main competitor] is [differentiator]. Tone: [confident but not hypey]. Each ad: a scroll-stopping first line, 2-3 sentences of body, and a CTA. Avoid the words "revolutionary," "seamless," and "game-changer."

The second prompt didn't get longer for the sake of it. Every added clause removes a guess the model would otherwise make badly. If you want the deeper logic here, the RCTCO prompt structure breaks down exactly which context slots earn their keep — and our guide on why most AI results are mediocre explains why thin prompts produce thin copy.

The fix is mechanical: before you ask for any asset, paste a short brief block — who you sell to, what you sell, the one objection that kills deals, and the proof that overcomes it. Reuse that block across every prompt in this pack.

Positioning & messaging

Get the foundation right and everything downstream — ads, emails, pages — writes itself faster.

Act as a positioning strategist. Here's my product: [product + what it does]. My three closest alternatives are [A], [B], and [C]. Write a positioning statement using the format: "For [audience] who [need], [product] is the [category] that [key benefit], unlike [alternative] which [shortcoming]." Then give me 2 alternative angles I might be underusing.

I sell [product] to [audience]. List the top 5 objections this audience has before buying, ranked by how often they kill the deal. For each, write a one-sentence reframe I can use in copy that defuses it without sounding defensive.

Turn this feature list into benefit-led messaging. Features: [paste 5-8 features]. For each, write the underlying benefit in the customer's own words, then the deeper emotional payoff. Output as a 3-column table: Feature / Benefit / Why it matters.

Write 7 distinct value propositions for [product] aimed at [audience]. Vary the angle: time saved, money saved, risk avoided, status gained, effort removed, outcome guaranteed, and identity ("the kind of company that…"). Keep each under 12 words.

Act as my toughest skeptical buyer, a [job title] at a [company type]. Read my pitch: [paste pitch]. Push back the way you actually would — name what's unclear, what sounds like every other vendor, and what you'd need to see to believe it.

Define the voice for [brand]. We're [3 adjectives] but never [3 adjectives to avoid]. Give me a voice chart with 5 rows: trait, what it means, do, don't, and a one-line example sentence in that voice.

Why these work: positioning prompts force the model to choose an angle rather than hedge. Naming real competitors and a real audience is what separates a sharp statement from a generic one. Role prompting — telling the model to be the skeptical buyer — surfaces objections you're too close to see; the role prompting guide goes deeper on this technique.

Ad copy

Paid copy lives or dies on the first line and the match between promise and audience.

Write 5 Google Search ad headlines (max 30 characters each) and 3 descriptions (max 90 characters) for [product]. Target keyword: [keyword]. Audience intent: [what they're trying to do]. Lead with the outcome, include one number or specific, and end at least one headline with a CTA.

Write 3 Facebook/Instagram ad variations for [product]. Audience: [segment] who [pain]. Angle 1: problem-agitate-solve. Angle 2: a surprising stat or claim. Angle 3: a short customer-style story. Each under 100 words with a clear CTA. No emojis, no "unlock," no "imagine if."

I'm running a retargeting campaign for people who viewed [page] but didn't buy. Write 4 ad variations that address why they hesitated — likely [objection 1] and [objection 2] — and give a reason to come back now. Keep urgency honest, not fake-scarcity.

Write 5 LinkedIn ad hooks for [product] targeting [job title]. Each hook should name a specific frustration of that role and feel like it was written by a peer, not a vendor. First line only — punchy, no buzzwords.

Generate 10 ad angles for [product] I can A/B test. For each, give the angle name, the core promise in one line, and the audience emotion it targets. Cover at least: fear of missing out, frustration with status quo, aspiration, skepticism, and convenience.

Rewrite this underperforming ad to lift click-through. Current ad: [paste ad]. It's getting impressions but few clicks. Diagnose the likely reason in one line, then give me 3 rewrites that fix it, each testing a different first line.

Why these work: ad prompts that specify character limits, a single target keyword, and a named emotion produce usable output instead of mush. Asking for variations by angle turns one prompt into a test plan — you ship the winner, not a guess.

Email & nurture

Email rewards specificity more than any channel, because you know exactly who's on the list and what they did to get there.

Write a 4-email welcome sequence for new subscribers to [product/newsletter]. Email 1: deliver the promised value + set expectations. Email 2: the origin story / why we exist. Email 3: address the #1 objection ([objection]). Email 4: a soft offer for [offer]. Give subject line + preview text + body for each, under 180 words per body.

Write a cart-abandonment email for [product]. The reader added it but didn't check out. Lead with the likely hesitation ([price / timing / trust]), reframe it, add one piece of proof ([testimonial type / guarantee]), and close with a single CTA. Friendly, not pushy.

I need 10 subject lines for an email announcing [offer/news]. Mix styles: curiosity, benefit, number, question, and plain-direct. Keep each under 50 characters. After the list, tell me which 2 you'd test first and why.

Write a re-engagement email for subscribers who haven't opened in 90 days. Acknowledge the silence honestly, give them one genuinely useful thing, and ask a simple yes/no question to re-confirm interest. No guilt-tripping.

Turn this blog post into a 3-email mini-course for my list. Post: [paste or summarize]. Each email teaches one idea, ends with a takeaway they can use today, and the last one points to [product] as the natural next step. Keep them skimmable.

Write a sales email for [product] to a warm list that already knows us. Audience: [segment]. The offer: [offer + deadline]. Structure: hook on the result they want, the offer, what's included, the deadline, one objection handled, CTA. Confident, specific, under 200 words.

Why these work: the model can write a competent email blind, but it can't know your subscriber actually abandoned a cart or went quiet for 90 days unless you tell it. Each prompt above hands the model the trigger and the state of the relationship, which is exactly what changes the right thing to say.

SEO & content briefs

Use the model to plan and brief — not to mass-produce thin articles. The leverage is in the structure you hand a writer (human or AI).

Build a content brief for an article targeting "[keyword]". Search intent: [informational/commercial]. Give me: the likely searcher question, 6-8 H2 sections in logical order, 3 questions to answer for featured-snippet/AI-overview capture, related entities to mention, and 2 internal-link ideas. Don't write the article — just the brief.

I want to rank for "[keyword]". Generate 15 related long-tail keywords and group them into 3-4 content clusters. For each cluster, suggest one pillar page and the supporting articles around it.

Analyze the search intent behind "[keyword]". Tell me what the searcher actually wants, what format wins (list / how-to / comparison / definition), what they'll do next, and the one thing most existing content gets wrong about this query.

Write 12 article-title options for the topic "[topic]" targeting [audience]. Mix formats: how-to, listicle, "X vs Y," definition, and contrarian. Each title should imply a clear benefit and stay under 60 characters where possible.

Draft a meta title (≤60 chars) and meta description (≤155 chars) for a page about [topic], targeting "[keyword]". Make the description promise a specific outcome and include a soft reason to click. Give me 3 variations.

Outline a comparison article: "[Product A] vs [Product B]." Cover the dimensions a buyer actually weighs: [price, ease, support, X]. Include a verdict-by-use-case section ("choose A if…, choose B if…") and an honest note on who shouldn't pick either.

Why these work: SEO prompts should produce structure and intent, not finished prose to publish unchecked. A brief tells you what to cover and why; you keep editorial control over voice and accuracy. For more on getting reliable, well-reasoned output, chain-of-thought prompting helps the model think through search intent before it answers.

Social media

Social copy needs a hook in the first line and a reason to stop scrolling — the model is good at volume here once you set the constraints.

Write 5 LinkedIn posts for [founder/role] about [topic]. Each opens with a one-line hook that earns the click-to-expand, delivers one concrete idea, and ends with a question or takeaway. No hashtag spam — max 3 relevant tags. Vary the format: story, list, hot take, lesson learned, and how-to.

Turn this blog post into 8 social posts across formats. Post: [paste/summarize]. Give me: 2 LinkedIn posts, 2 X/Twitter threads (first tweet + 4 follow-ups each), 2 Instagram captions, and 2 short hooks for Reels/TikTok. Keep each platform's native voice.

Write a 6-tweet thread teaching [topic] for [audience]. First tweet must stop the scroll with a specific claim or number. Each following tweet = one idea, one line of proof. Last tweet recaps and invites a follow. No "🧵" cliché opener.

Give me 10 short-form video hooks (first 3 seconds) for content about [topic] aimed at [audience]. Each should create an open loop or challenge an assumption. Then pick the 3 most likely to retain viewers and say why.

Write 5 Instagram captions for [brand] promoting [post/product]. Each: a scroll-stopping first line, 2-3 lines of value or story, one CTA, and a tight set of tags. Match a [voice descriptor] tone. No "link in bio" as the whole CTA — give a reason to click.

I post on [platform] about [topic]. Generate a 2-week content calendar: 10 post ideas with the format, the hook, and the goal (awareness / engagement / conversion) for each. Balance educational, personal, and promotional so it's not all selling.

Why these work: social prompts that name the platform, the format, and the job of the first line beat "write me a post" every time. Asking the model to pick its top few and justify them turns raw volume into a shortlist you can actually ship.

Landing pages & CRO

Conversion copy is where structure matters most — every section has a job, and the prompt should name it.

Write the above-the-fold section for a landing page selling [product] to [audience]. Give me: a headline that states the core outcome, a subhead that adds the how or the proof, and a primary CTA. Then 2 alternative headlines testing different angles ([outcome vs pain vs identity]).

Draft the full copy for a landing page for [product]. Sections: hero, problem, solution, how-it-works (3 steps), social proof slot, objection-handling FAQ (4 Qs), and final CTA. Audience: [segment]. Keep it skimmable with short paragraphs and benefit-led subheads.

Write 5 CTA button variations for [offer]. Move beyond "Get started" — make each one specific to the value or the next step ([e.g. "Start my free brief"]). Then note which fits a high-intent vs low-intent visitor.

My landing page converts at [X]%. Here's the current copy: [paste]. Act as a CRO specialist and give me a prioritized list of 5 changes most likely to lift conversion, with the reasoning for each and roughly how much effort it takes.

Write 6 FAQ entries for a [product] landing page that quietly handle objections: price, switching cost, trust/security, time-to-value, "will this work for my case," and cancellation. Answer honestly and in a reassuring voice — each answer 2-3 sentences.

Write 3 versions of a hero headline for [product], each anchored to a different stage of awareness: unaware (lead with the problem), problem-aware (lead with the solution category), and solution-aware (lead with our specific edge over [competitor]).

Why these work: CRO prompts win when you assign each section a single job and give the model the conversion math. "Act as a CRO specialist" plus your real conversion rate produces prioritized, reasoned suggestions instead of a generic checklist — that's role prompting doing real work.

Analytics & reporting

The model can't see your dashboard, but it's excellent at turning numbers you paste into a narrative your boss or client will read.

Here's my campaign data: [paste metrics — spend, impressions, clicks, conversions, CPA]. Write a 5-bullet executive summary: what happened, what worked, what didn't, the one number that matters most, and the recommended next move. Plain English, no jargon.

Turn these monthly marketing numbers into a client-ready report: [paste data]. Structure: headline result, performance by channel, one win, one concern, and next month's priorities. Confident and honest — flag underperformance instead of hiding it.

I have two A/B test results: Variant A [metrics], Variant B [metrics]. Tell me which won, whether the difference looks meaningful or could be noise given the sample, and what to test next based on what the result implies about the audience.

Write 5 hypotheses for why [metric] dropped this month. Context: [what changed — campaigns, seasonality, site changes]. Rank them by likelihood and tell me what data I'd check to confirm or rule each one out.

Summarize this customer-feedback dump into themes: [paste reviews/survey responses]. Give me the top 5 recurring themes ranked by frequency, a representative quote for each, and one marketing angle each theme suggests.

Draft 3 data-backed talking points for a stakeholder update, based on: [paste key metrics]. Each point = the claim, the number that supports it, and the "so what" for the business. Keep it to one tight paragraph total.

Why these work: reporting prompts turn the model into a translator. You bring the numbers; it brings the structure and the plain-English framing. Asking it to flag noise vs. signal in A/B tests keeps you from over-reading a result that's really just sample variance.

How to make any of these convert better

Three habits separate marketers who get usable output from those who paste, sigh, and rewrite from scratch.

Front-load real context. Keep a saved brief block — audience, offer, top objection, proof — and paste it above every prompt. It's the single highest-leverage habit in this entire pack.

Ask for variations, then choose. "Give me 5 and tell me which to test first" beats "give me one." You stay the editor; the model does the volume.

Iterate, don't restart. When a draft is 80% right, tell the model what to fix — "punchier first line, cut the second paragraph, more specific CTA" — instead of rewriting the prompt. The conversation is the tool.

If you want to stop pasting prompts into a blank box and instead build and version your marketing prompts visually, that's exactly what Studio is for — Goal, Role, Context, and Instruction as draggable blocks you can reuse across every campaign. And before you write your next ad, drop a rough request into the free brain to see what an expertly structured prompt returns.

Frequently asked questions

What are the best ChatGPT prompts for marketing?

The best marketing prompts aren't clever one-liners — they're prompts loaded with your real context: who you sell to, what you sell, the main objection, and your proof. A prompt like "write 3 ad variations for [product] targeting [specific segment] whose biggest frustration is [pain], with our edge being [differentiator]" outperforms any generic template, because specifics remove the guesses the model would otherwise make.

How do I write a ChatGPT prompt that actually converts?

Name the asset, the audience, the constraint, and the angle. Specify the format (subject line, ad headline, FAQ), the character or word limit, the emotion you're targeting, and one thing to avoid. Then ask for several variations and pick the strongest to test. The RCTCO structure is a reliable template for assembling those slots.

Can ChatGPT replace a marketing copywriter?

No — it replaces the blank page, not the judgment. The model is fastest at volume and structure: ten ad angles, a content brief, a first draft. A marketer still decides which angle fits the brand, fact-checks every claim, and edits for voice. Treat it as a tireless junior writer you direct closely, not an autopilot.

Should I use my own brand voice in the prompt?

Yes. Define your voice once — three adjectives you are, three you're never — and paste a one-line example sentence in that voice. Including it in the prompt is the difference between copy that sounds like a template and copy that sounds like you. Save it in your reusable brief block so it rides along with every request.

The fastest way to level up is to start from prompts that already work. Browse the Library — 2,750+ expert prompts you can copy, tweak, and run today, including dozens built for marketing — or open any of them in Studio to rebuild it around your exact product. Or just type your next marketing task into the free brain and watch a rough ask come back as expert-level copy.