What follows is what I've learned from mining 10M+ customer comments across 6,151 subreddits and testing Meta ads for my own product. These are patterns, not rules. Steal what works, test it, kill what doesn't.
1. A winning ad is already written
Copy can't create desire. It can only channel desire that already exists. Your customers typed the exact words into a review, a Reddit thread, a comment section somewhere. Your job is finding them, not inventing them. That's the whole premise of voice of customer research.
2. A winning ad wins in the first second
Nobody decides to watch your ad. They decide not to scroll past it. The hook is 80% of the outcome. Spend 80% of your time on it.
3. A winning ad creates an itch
Withhold the why. Withhold the what happened. If the viewer can predict your next line, they scroll. Curiosity is just information with a gap in it.
4. A winning ad enters at problem aware
Cold traffic doesn't know your product exists and doesn't care. Meet them at the problem they already complain about, then walk them toward the solution. Product-first ads are for retargeting.
5. A winning ad describes the problem better than the customer can
Before anyone trusts your solution, they need to believe you understand their problem. Describe it in their words, sharper than they would, and half the sale is done before the product shows up.
6. A winning ad sells one angle
One reason to buy per ad. Blend three angles and you get something that kind of speaks to everyone and sells to no one. Pick one, save the rest for the next ad.
7. A winning ad sounds like the customer wrote it
"Revolutionary formula" is copywriter language. "My scalp stopped itching in a week" is customer language. One reads as selling. The other reads as a person. Throw out the thesaurus.
8. A winning ad shows the receipt
"Feel energized" tells. "Ran 5k before my kids woke up" shows. Write the concrete thing that happened and let the viewer draw the conclusion. Conclusions people reach on their own don't feel like marketing.
9. A winning ad survives three questions
Can you picture it? Can you prove it wrong? Could a competitor run it word for word? If a competitor could run your ad verbatim, it was never your ad.
10. A winning ad calls its person out
People scroll past ads and stop for mirrors. A label they already own, night shift nurses, dads who cook, buys more attention than any visual trick.
11. A winning ad raises the stakes
Make the problem bigger. Add a witness. Make it urgent, permanent, or expensive to ignore. Low stakes are forgettable by design.
12. A winning ad answers the objection before it forms
You already know the top objection. It's sitting in the comments of your last ad and in your competitor's one-star reviews. Handle it mid-ad with a fact or a because. Don't leave it for the landing page.
13. A winning ad never defends its price
Defending price accepts the wrong frame. Reframe the cost instead: per day, against the alternative, against what staying stuck costs.
14. A winning ad doesn't look like an ad
The feed trains people to skip anything polished. A shaky phone video with a real voice outruns the studio shoot most weeks. Native wins.
15. A winning ad is different, not better
"Better" starts an argument you can't win in 15 seconds. "Different" ends it. Find the hole your competitors leave open and stand in it.
16. A winning ad has an enemy
The old way. The overpriced incumbent. The advice that never worked. An enemy gives the viewer someone to blame that isn't themselves.
17. A winning ad is one of twenty
Nobody writes the winner on the first try. Write twenty hooks, rank them, test five, keep one. Creative volume is the only edge the algorithm hasn't automated away. If you need a starting bench, the ad hook generator is free.
18. A winning ad gets killed by a rule, not a feeling
Set the kill rule before launch, while you're still objective. I kill anything that doubles target CPL with nothing to show for it. Feelings are not a metric.
19. A winning ad dies. The angle doesn't
Creative fatigues in weeks. The desire underneath it doesn't. When the winner fades, remake the same angle with a new hook, new face, new format. Don't start the hunt from zero.
20. A winning ad pays off its hook
An open loop you never close doesn't create a customer. It creates someone who feels tricked. Curious keeps watching. Confused scrolls.
21. A winning ad works with the sound off
Most of the feed plays muted. Captions, text on screen, visible proof. If the story needs audio to make sense, there's no story yet.
22. A winning ad does its own targeting
Broad won. The creative is the targeting now. The hook decides who stops, the label decides who feels seen, the angle decides who buys.
23. A winning ad was never brainstormed
Brainstorms produce what the room already believes. The customer said something better in a review last Tuesday. Two hours reading Reddit comments beats two weeks of workshops.
24. A winning ad borrows its proof
A claim from you is marketing. The same claim from a customer is evidence. Screenshots, reviews, ugly testimonial clips. Let someone else say your best line.
25. A winning ad runs on one emotion
Fear, shame, hope, relief, anger. Pick one driver and commit. An ad that tries to make people feel everything makes them feel nothing.
26. A winning ad reads out loud without embarrassment
Read the script to a friend. Any sentence you'd be embarrassed to say to their face is copywriter language. Cut it.
27. A winning ad's comment section is the next ad
Your ad comments are free research from the exact people the algorithm already found for you. The objections, the jokes, the questions. Your next hook is in there, and scraping the comment section takes one click.
The best operator I've talked to sells $300k/month and still reads comments by hand. These are patterns, not rules.
And if you've found a 28th, I want to read it.
About the author
Daniel is the founder of Adlicio. He builds the scrapers behind it and uses them daily to turn customer comments and reviews into ad angles.