My work and the thinking behind it

Every brand in the category sells the same benefit in the same words. Nothing in the feed sounded like the customer's own experience.

Two ads from the Black Friday sale, same warm audience:

• The conventional ad: polished product photography with a clear discount.
• The MS Paint ad: a single sentence lifted word for word from a real customer review, with nothing else on screen but my drawing.
• The best line was found in a review. I didn't write it.
• It showed her pain in vivid imagery before any offer was shown.
• With no product on screen, nothing competed with the line for attention.

The obvious thing to sell was not the thing the buyer needed to hear. Product and offer got 1.02%. The customer's own words got 11.22%. We later used this same format in our email creatives.

Steady ad spend and empty workshops. On top of that, Facebook kept flagging the ads for non-compliance.

The client's target audience had already paid for courses and had nothing to show for it, while they watched peers collect awards. The buyer needed someone to point out the frustration and then introduce the solution.

So I mapped the buyer before writing a word, then built a five-stage funnel around what I found. I launched 4 winning ad ideas:

• Pain: mirror the audience's frustration.
• Origin story: build trust with the audience.
• Direct offer: state the offer plainly, with no warm-up.
• Cultural moment: ride what people are already talking about.
• Later added testimonial video ads when winning ads started showing signs of fatigue.

Scaled ad spend from $2k a week to $10k a week over six months. It hit 36x ROAS at $10k a week.

Earlier this year I joined a 10-week creative strategy program to sharpen my creative strategy skills. There was no written exam at the end. To graduate, I had to make real ads.

I made mine for the founder's own 7-figure business, and he judged them on one question: would he put his own money behind them? Copy that sounded like AI was an automatic fail.

So I made three ads for one buyer, a copywriter stuck on small retainers:

• Comic: me against the guy I learned copywriting with, and the very different months we are having.
• WhatsApp chat: a friend asking how on earth he is closing 5k retainer clients.
• Instagram story: a sceptic reacting to the idea that two words in a bio could change what a copywriter gets paid.

All three look like something already in the feed, so they get read before they get judged as ads.

The product is a shampoo and conditioner bar for mature hair. Each brief has two readers, the creator and the editor, and everything they need to shoot and cut the ad is on the page.

So both briefs follow the same ten parts. The ones that do the most work:

• The concept in one sentence, before any script.
• Five hooks on one body, shot back to back in the same setup, so each one becomes its own ad to test.
• A shot list in two columns: what she says on the left, what the viewer sees on the right.
• Format notes with the reason attached. Captions are burned in on every line because most viewers watch on mute.
• Hard constraints. The claims and shots that would stop the ad from running, written down before anyone films.

The ads started with research into the US women's fragrance buyer, across three brands and their reviews. One thing kept coming up: she is buying someone else's reaction.

So each ad sells one small reaction, sung as a story:

• Waitress Asked: she doesn't believe the videos where ten men walk up to a girl. Then a waitress stops to ask what she's wearing. (That idea came from a customer review.)
• Longer Hugs: her boyfriend hugs her like a coworker, one arm, phone in the other. By the chorus it's both arms.
• The song came first. No scene was made until the song was approved.
• Each ad has its own primary text and headline, written to run with the video.

The song, the characters and the animation were all made with AI tools.