Vibe marketing: what it is, where it came from, and how to do it

Vibe marketing is a spin-off of vibe coding, and the dates are well documented.

• February 2, 2025: Andrej Karpathy posted on X about "a new kind of coding I call 'vibe coding'", where you "fully give in to the vibes" and "forget that the code even exists." He described talking to an AI coding tool, accepting its changes without reading them, and pasting error messages back in until things worked, adding that it was "not too bad for throwaway weekend projects."
• February 2025: According to The Vibe Marketer, a site run by James Dickerson (known online as @boringmarketer), Greg Isenberg and Jordan Mix, Dickerson was sharing marketing automation workflows in a private Slack channel when Isenberg suggested he tweet about it and call it "vibe marketing." Dickerson says he then tweeted it and Isenberg amplified it. That is their own account of the origin.
• March 22, 2025: Isenberg posted a widely viewed thread arguing that the "20x acceleration" of vibe coding "is hitting marketing teams RIGHT NOW. It's called VIBE MARKETING." In it he credited his co-founder @boringmarketer (Dickerson) with introducing him to the idea. Five days later he followed with "VIBE MARKETING IS HERE," a ten-point list of what it changes, including campaign planning, ad testing, targeting, content, analytics and budgets.
• June 2025: The Wall Street Journal reported that Meta aims to let advertisers fully automate ad creation with AI by 2026, supplying little more than a product image and a budget goal (as summarized by Marketing Dive on June 2, 2025). In an interview with Ben Thompson of Stratechery, Mark Zuckerberg described advertisers who "connect to your bank account, you don't need any creative."
• October 8, 2025: Digiday wrote that "vibe marketing used to be a punchline" until OpenAI's developer conference announced tools that pull ChatGPT deeper into marketing work.
• November 6, 2025: Collins Dictionary named "vibe coding" its Word of the Year, crediting Karpathy with coining it.

So the short answer to "who coined vibe marketing" is: Greg Isenberg popularized it in March 2025, and The Vibe Marketer's account says the name came out of a Slack conversation between Isenberg and James Dickerson a few weeks earlier. Many agencies and software vendors now publish their own definitions.

Search results for "vibe marketing" mix two different ideas, so it helps to know which one someone means.

1. The AI meaning (the one this guide covers). One person directs AI models and automations to do work that used to need a team. This is the Isenberg and Dickerson usage, and it is the meaning used by most of the explainers cited in this guide.
2. The brand-feel meaning. Salesforce's small business blog (October 22, 2025) defines vibe marketing as a strategy focused on "creating an emotion around a brand rather than explicitly promoting specific products or features." That is closer to classic brand marketing: selling a mood or identity rather than a feature list.

The two overlap in practice, because the AI version still depends on a human deciding what the brand should feel like. When someone describes themselves as a "vibe marketer," they usually mean the AI and automation skill set.

The proponents describe the shift in bold terms. Isenberg's March 2025 thread contrasted "marketing teams with 10+ specialists" taking "weeks (and thousands of dollars) to launch anything" with "a single smart marketer with AI agents and workflows" launching "in days not weeks." The Vibe Marketer's definition talks about enabling one marketer to "execute at the level of five while maintaining strategic oversight and creative control."

Those are claims from the people who named the trend, not measured results. What actually changes, stripped of the hype:

• Who does the production. Drafts of copy, images, short videos and simple landing pages come from AI tools. The human writes the brief and picks what ships.
• How many variations you test. Drafting 30 ad hooks takes little more effort than drafting three, so testing can shift from a couple of options to many.
• Where the bottleneck sits. As Isenberg put it, "what to say becomes more important than how to say it." Positioning, taste and knowing your customer become the scarce skills.
• Who builds the tools. Marketers use vibe coding tools and automation platforms to build small internal tools (a calculator page, a lead scraper, a reporting script) instead of filing a request with engineering.

What does not change: you still need a product people want, a clear customer, and distribution channels where that customer already is. AI makes execution cheaper; it does not supply demand.

The Vibe Marketer describes the shift as moving from "doing tasks manually" to "managing automated systems." Isenberg's March 27 list similarly said marketing careers will require "systems thinking and prompt engineering." The Vibe Marketer's guide lays out four levels:

1. Using AI tools one at a time (a chatbot for copy, an image tool for graphics).
2. Connecting them with workflow automation (n8n, Make or Zapier).
3. Marketers who code: using vibe coding tools to build their own pages, scripts and internal apps.
4. Full autonomy, which the same guide says is "not here yet."

The same guide says most marketers are stuck at level 1 or 2. For a solo founder, moving from level 1 to level 2 usually means the same prompts, but triggered automatically and routed into a review queue.

The loop below is our own summary, drawing on the descriptions from Hightouch, Go Fractional, Inc. and The Vibe Marketer:

1. Set the context. Write down the audience, the problem you solve, your proof, your tone, and words you never use. Hightouch calls this a "brand context layer" and pairs it with customer data as the two foundations AI needs.
2. Generate. Ask the model for many options: hooks, headlines, post drafts, email subject lines, image concepts.
3. Curate. A human picks, edits and approves. This is the part that is easy to skip and expensive to get wrong.
4. Ship. Publish through scheduling tools or automations.
5. Measure. Track which versions drove visits, signups or sales.
6. Feed back. Put the winners and losers back into the context so the next batch starts smarter.

The loop is only as good as step 1. Hightouch's warning is the one to remember: "AI executing in the wrong direction produces more of the wrong thing faster."

Here is how the loop maps onto a typical launch for a small software product. It assumes one person, a small budget, and an app that already works.

1. Write a one-page context doc before any prompting. Who the app is for (be narrow), the job it does, three things users say about it, how it is different from the two closest alternatives, and a short style guide. Paste this into every AI session or save it as a project file so every output starts from the same brief.
2. Pick one social channel and one owned channel. The social channel is wherever your buyers already talk (X, LinkedIn, Reddit, TikTok). The owned channel is your site plus an email list. Spreading thin across six networks is an easy way for AI-generated volume to turn into noise.
3. Generate in batches, publish in small doses. Ask for 20 to 30 post hooks at a time from your context doc, keep the five you would actually say out loud, and edit them in your own voice. Schedule them with a tool like Buffer so posting does not depend on your mood that day.
4. Build the pages you need with vibe coding. A launch page, a pricing page, a changelog, a free mini tool related to your product. If you built the app with Lovable, Bolt, Replit or Cursor, the same tools can build these.
5. Automate one boring task at a time. Good first automations: turn each changelog entry into a draft post; send new signups a welcome email; save mentions of your product to a sheet. Keep a human approval step in anything that posts publicly.
6. Get listed where people look for new tools. Launch platforms, niche directories and communities can give you discovery and links that outlast launch day. See the next section for how to do this without wasting time.
7. Tag every link and check results weekly. Use UTM parameters per channel and read them in Google Analytics (free). Drop what is not working after two or three weeks; double down on what is.
8. Update the context doc. Add the hooks that worked, the objections you heard and the words customers used. This is what makes the next month better than the first.

Directories and launch sites are a different kind of launch work from content: a listing either exists on someone else's site or it does not, and a relevant one can keep being found after the first week of social posts.

Two cautions from Google's own spam policies before you submit anywhere:

• Google lists "low-quality directory or bookmark site links" as an example of link spam. A hundred listings on irrelevant, auto-generated directories are not an asset. Pick directories your buyers actually browse, in your category.
• Paid placements are fine for visitors, but Google's policies say links bought for advertising should be qualified with rel="nofollow" or rel="sponsored." Do not pay for a listing expecting a ranking boost from the link alone.

The work itself is repetitive: each directory wants different tagline lengths, descriptions, categories, logos and screenshots, and a submission can be rejected or never published without you hearing about it. VibeLaunch, the site publishing this guide, is a tool for this step. It matches a product to the directories that fit it, drafts each submission to that site's requirements, and then checks that the listing and its backlink actually went live. You can also do all of this by hand with a spreadsheet; the point is to track which listings exist, not which forms you filled in.

The honest case against vibe marketing is mostly about quality and trust, not technology.

• It takes setup. Inc.'s June 2025 piece, "Vibe Marketing Is Not as Easy as It Sounds," points out that it needs real setup time, several tools rather than one, good prompting, brand material to train on, and guardrails to keep output on brand.
• Speed without direction backfires. Hightouch warns that without customer data, AI produces "content that appears relevant but has no connection to the audience," and without brand knowledge, output "drifts from the brand messaging."
• Search engines judge value, not method. Google's guidance (February 2023) says "Using AI doesn't give content any special gains. It's just content." Its spam policies list using generative AI or similar tools to "generate many pages without adding value for users" as an example of scaled content abuse. Publishing hundreds of thin AI pages is a risk to your whole site.
• Fake social proof is banned by an FTC rule in the US. The FTC's final rule announced August 14, 2024 prohibits fake reviews and testimonials, including "AI-generated fake reviews," and prohibits buying fake indicators of social media influence, such as bot-generated followers or views, when the buyer knew or should have known they were fake. It lets the FTC seek civil penalties from knowing violators. Testimonials attributed to people who do not exist are exactly what it describes.
• Not every campaign suits it. Go Fractional notes that "long-term, highly regulated campaigns" may still need a traditional approach. Health, finance and legal claims need human review no matter how the draft was made.
• Mistakes scale too. An automation that posts a wrong price or a made-up statistic posts it everywhere at once. Keep a human approval step on anything public.

What we don't know. There is no independent, controlled data showing that vibe marketing outperforms a conventional small team. The growth figures in circulation (for example, the "686% search growth" and "2,600+ practitioners" cited by The Vibe Marketer) come from people who run communities and courses around the term. There is also no agreed definition: vendors define it in ways that fit what they sell. Meta's fully automated ads were reported as a goal for 2026 in the 2025 coverage; we have not confirmed whether that has shipped.

How this guide was put together. We did not run marketing experiments for this page. We compiled it from primary sources (the original posts on X, Google's and the FTC's own documentation, official pricing pages) and from the most-cited explainers, and we attribute every claim to where it came from. Prices were checked on each vendor's pricing page on September 29, 2026 and will change.

You do not need a special "vibe marketing platform." The usual stack is a general AI model, an automation tool, a design tool, a scheduler and free analytics. The Vibe Marketer's recommended starting stack is Claude or ChatGPT for content and Zapier or Make for automation, with Claude Code, n8n and Cursor added at the more advanced level.

A workable starting stack for a solo founder can cost about $20 a month: one paid AI model plan (ChatGPT Plus or Claude Pro) with the free tiers of Canva, Buffer, Make and Google Analytics. Upgrade a tool only when its free limit is what is actually holding you back.

Prices below are as of September 29, 2026, from each vendor's official pricing page.