Person climbing a spiral staircase of classic books into a golden sky surrounded by floating cameras, brushes and sound waves

Quick answer

To become a marketing engineer in 30 days, spend week 1 learning marketing fundamentals from classic books, week 2 turning your notes into a “second brain” your AI can use, week 3 learning one AI creative tool at a time, and week 4 doing a small free project for a local business to earn your first case study. A marketing engineer, as Greg Isenberg describes it, is someone who does marketing with AI agents and runs many small experiments. Thirty days will not make you an expert, but it will give you a working system and real proof you can show.

Key takeaways

  • Fundamentals first. AI makes more ads; it does not tell you which ideas are good.
  • Your second brain should include timeless marketing principles, not just this year’s posts.
  • One tool at a time beats trying every new model at once.
  • A real case study, even unpaid, is worth more than a certificate.
  • Start with a few experiments a week. The “100 a day” Greg mentions is for experienced teams.

This plan expands on one answer from our article Claude Whiteboards, ChatGPT Business Agents and Gemini 4 Carbon: Greg Isenberg’s Live Reaction. During his livestream, Greg Isenberg was asked what steps he would take to become a marketing engineer from zero, and he gave a four-part answer. Below we turn it into a practical week-by-week plan.

📘 Main article: See Greg’s full answer, and the AI launches he reacted to, in Claude Whiteboards, ChatGPT Business Agents and Gemini 4 Carbon: Greg Isenberg’s Live Reaction.

What is a marketing engineer?

A marketing engineer is a marketer who uses AI agents and automation to create, launch and measure lots of marketing experiments quickly. Greg traces the role through marketing history: from print, to radio and TV, to the internet, each step letting marketers run more tests. He says the best marketing engineers now run dozens, sometimes 100-plus, experiments a day.

He also argues that marketing is harder than coding, because it blends psychology, culture, writing and design, and that “learn marketing is the new learn to code”. That is his opinion, but it is a useful reminder that the tools are the easy part.

What does the 30-day plan look like?

WeekFocusOutput by the end of the week
1Marketing fundamentalsOne page of your own notes per book or course
2Second brainA folder or notebook your AI can search, plus a short “marketing rules” file
3AI creative toolsThree finished test assets made with one tool
4First real projectA small result for a friendly business, written up as a case study

💡 Tip: Block 45 to 60 minutes a day. Consistency matters more than long weekend sessions.

Week 1: How do you learn the fundamentals?

Read or listen to two or three classic marketing books and write one page of notes on each in your own words. Greg names David Ogilvy as his starting point. Good choices include Ogilvy on Advertising, Claude Hopkins’s Scientific Advertising and Eugene Schwartz’s Breakthrough Advertising.

For each book, capture:

  • The three to five principles that still apply today
  • One example ad or idea that shows each principle
  • One way you could test that principle this month

Week 2: How do you build a marketing second brain?

Put your notes, plus your own examples and results, into one place your AI can read, and write a short “marketing rules” file it follows. Greg’s point is that many people fill their second brain only with recent posts, when the most valuable material is often decades old.

⚠️ Warning: Use books you have bought and your own notes or summaries. Do not download pirated copies or upload whole copyrighted books to AI tools; your own notes are more useful anyway, because they are in your words.

A simple structure works:

  • /principles: your book notes
  • /swipe: ads and pages you admire, with a line on why (our ad swipe file guide shows how)
  • /results: what you tested and what happened
  • marketing-rules.md: ten rules your AI should follow, such as “lead with one clear benefit”

Week 3: Which AI creative tools should you learn first?

Pick one type of tool, learn it properly, and only then add the next. Greg lists image, video, sound and world-model tools as the areas to learn platform by platform.

OrderTool typePractice task
1Image generationThree ad images for one offer, each testing a different angle
2Short videoOne 15-second product or service video
3Voice and soundA short voiceover or jingle for that video
4Copy and scoringAsk an AI to score your assets against your marketing rules

Greg gave an example from just before his stream: he asked Claude to create a thumbnail, and says it took about six minutes and cost only a few cents, compared with the hundreds of dollars he has paid for thumbnails before. Results like that depend on your brief and judgement, which is why weeks 1 and 2 come first.

To score your work before it goes live, see how to set up an AI creative director to score ads.

Week 4: How do you get your first case study?

Offer a small, clearly defined project to a friendly local business, do it free or at a low price, and write up what you did. Greg’s example is helping a friend who works at an ice cream shop, getting a case study, and then starting to charge.

  1. Agree one goal and one measure, such as email sign-ups from a counter sign, in writing.
  2. Get permission to use their name and results in your case study.
  3. Produce two or three variations and run them for one to two weeks.
  4. Write a one-page case study: the problem, what you tried, what happened and what you would do next.

⚠️ Warning: Be honest about results, including small or disappointing ones. Never promise a business owner a sales figure, and never spend their money on ads without a written budget limit.

How do you run and log experiments?

Every experiment needs a hypothesis, a single change, a measure and a result, written down. A simple log is enough to start:

DateHypothesisWhat changedMeasureResultNext step
3 DecA price in the headline gets more clicksHeadline onlyClick rateTo fill inKeep or drop
5 DecA photo of the owner builds trustImage onlySign-upsTo fill inKeep or drop

Start with two or three experiments a week. As you get faster, your AI agents can help draft variations and fill in the log, which is how marketing engineers scale up. Our article on AI marketing agents shows what that looks like at a bigger scale.

What mistakes should you avoid?

  • Chasing every new AI model instead of mastering one tool
  • Skipping the fundamentals, so you make more average ads faster
  • Running tests without writing down what changed
  • Changing several things at once, so you cannot tell what worked
  • Calling yourself an expert before you have a single case study

Frequently asked questions

Do I need to code to be a marketing engineer?

No. Many people direct AI agents with clear written instructions. Basic comfort with spreadsheets and simple automation tools helps.

Is 30 days really enough?

It is enough to build a system and get a first case study. Becoming skilled takes much longer and comes from running many experiments.

Where does the term marketing engineer come from?

Greg Isenberg describes it as a job title that emerged in the last year for marketers who work with AI agents and run many experiments.

How many experiments should a beginner run?

Two or three a week is plenty at first. Increase the number only when you can log and learn from each one.

author avatar
Garry Knight
I'm Garry Knight, the person behind Prodify Digital. I write about email list building, email marketing, SEO, AI search and the tools that connect them. My aim is to make online marketing easier to understand, so creators and small business owners can make informed decisions about building an audience and keeping people engaged. Here you'll find straightforward guides and product reviews that explain what something does, where it fits and which limitations matter. The focus is on clear explanations and useful next steps—not hype, shortcuts or promises of easy earnings.

Leave a Reply