Final project: a campaign in 7 days
Lesson 16 of 16 in the course “Marketing 2027”. Free, with animations and a quiz.
Putting it all together
You’ve gone through the whole course: from how the auction and other brands’ ad libraries work to agents in your ad account and honest measurement. Now let’s pull it all into one plan you can complete in a week. Day 1: competitor research. Day 2: the offer. Day 3: five creatives. Day 4: the signal, so the platform sees purchases. Day 5: launch an automated campaign with guardrails and a test. Day 6: a helper agent in read-only mode. Day 7: review and the next test. Each day builds on a lesson you’ve already done, so there’s nothing new to learn. The key is to do it, not just read it. A week of practice beats a month of marketing videos.
Days 1–2: research and offer
Day 1. Open Meta Ad Library, Google Ads Transparency Center and TikTok Top Ads and find 10 ads in your niche. For each, write down the hook, the main promise and the proof. Look for repeating techniques: if five competitors open with the same type of hook, it probably works. Important: we break down techniques, we don’t copy videos, that’s been the rule since the first lessons. Day 2. Build your offer using the value equation from Alex Hormozi’s book “$100M Offers.” Value goes up when you strengthen the dream outcome and the buyer’s confidence it will happen, and reduce the waiting time and effort. For Nika’s hoodie that could be: “lasts for years, keeps its shape after 30 washes, 2-day delivery, no-questions returns.”
Day 3: five creatives
Make five creatives on five different ideas: customer review, product test, humor, story, comparison. Remember the automated campaigns lesson? Different ideas find different people, copies with a new button color don’t. Build each video on Russell Brunson’s “hook, story, offer” formula: grab attention, tell, propose. The first three seconds decide whether someone keeps watching. Back in a 2017 Facebook study, nearly half of a video ad’s value came in the first three seconds. If you use AI to create a person, voice or event that didn’t exist, add a label visible without a click and declare AI when uploading. Better still, film at least one video with a real person: those earn more trust.
Days 4–5: signal and launch
Day 4 is about the boring but most important part. Connect the pixel and Conversions API, or the equivalent on your platform. Make a test purchase and confirm the platform saw it and didn’t count it twice. Without this, everything else is built on sand: automation will learn from clicks, and reports will lie. Day 5 is launch. Create an automated campaign: Advantage+, Performance Max or Smart+, depending on the platform. Upload all five creatives. Build the fence: negative keywords, exclusions, text rules. And build in a test with one difference right away, for example two versions of the hook in your best video. Write down the length, at least a week, and the metric you’ll judge by. Promise yourself not to peek early.
Days 6–7: agent and review
Day 6. Connect an AI agent via the official MCP server in read-only mode. Give it the playbook from the media-buying lesson: the morning report format, four alerts with thresholds, and stop rules that, for now, you carry out yourself on its prompt. Day 7 is review. Add up the “sales” from all platforms and compare with real orders. Look at what the agent proposed over these days and where it was right. Write everything into the test log: what you launched, what you saw, what you decided. And plan the next test right away. Key point: don’t expect profit in week one. Its result is knowledge: which ideas hook your audience and how honest your numbers are. That knowledge is what profitable advertising is built on.
Five mistakes and five habits
Let’s sum up what most often costs money. One: copying someone else’s creative one to one. Habit: break down techniques and build your own. Two: hiding AI without a label. Habit: the three decision questions before every launch. Three: stopping a test on day one. Habit: length and metric set in advance, and no peeking. Four: trusting platform numbers without checking. Habit: compare with real orders weekly and run a lift test quarterly. Five: giving an agent full access at once. Habit: the trust ladder, from read-only upward. Notice that none of these habits needs a big budget or rare skills. All it takes is discipline, and that’s exactly what separates people whose ads pay off.
Your final project
Pick a real product or service: yours, a friend’s, or an invented one if you have nothing to sell yet. Go through the seven days on this lesson’s map. Only use a budget you can afford to lose: week one is about learning, not getting rich. If you have no ad money at all, still do days 1–3 and 7: research, offer, creatives and the log. That’s already real experience. At the end of the week, answer three questions in your log. Which creative idea worked best and why? How closely did the platforms’ numbers match reality? What will you test next? If you have answers, you’ve truly finished the course. Good luck, and may your ads be honest and profitable!
Check yourself
- How should day one start?
Correct answer: Break down 10 competitor ads and note the techniques. Research in ad libraries shows which techniques repeat in your niche. We take techniques, not other people’s videos. - What will strengthen the offer most?
Correct answer: Prove the result and cut waiting and effort. Per Hormozi’s value equation, value grows with confidence in the result and falls with waiting and effort. A 30-wash test, fast delivery and easy returns hit all four parts. - What’s the right move?
Correct answer: Set up Conversions API and make a test purchase before launch. The signal is the foundation. Without purchase data automation learns from clicks, and every report and test that week is skewed. - How should she judge week one?
Correct answer: It’s knowledge: log the takeaways and plan a test. Week one’s goal is learning which ideas hook the audience and how honest the numbers are. Zero profit with new knowledge is a normal start.