We recommend signing up with your work email to keep all your marketing content in one place
At Anyword, we draw on millions of real A/B tests to deliver sophisticated performance predictions. Here are some recent observations and actionable insights from that data.
Too often, ad copy gets written on instinct or vibes. To see which instincts hold up, we analyzed 1,000 real Meta A/B tests from fashion advertisers (500 headline pairs and 500 primary-text pairs). Fashion is a fast-moving, copy-heavy category, so it's a good place to look for patterns that carry over to other verticals.
Every pair is a head-to-head test with a statistically significant winner (p<0.05). The tests span five years and more than 23 million impressions, and the winning version's click-through rate was a median of about 60% higher than the loser's.
These aren't just vibes or things that sound nice at first blush. They're battle-tested outputs from real money spent in real feeds, and you can apply them to your next campaign.
Headlines in first person (I, my, we) won 81% of the time. Conversational phrasing like "it's" won 88%.
Loser: Sneaker Waterproofing Spray review
Winner: I don't have to worry about ruining my sneakers, thanks to this $15 waterproofing spray
Headlines with a number won 69% of the time. And coincidentally, experience verbs like "tried," "replaced," "found" and "swear by" also won 69% of the time.
Loser: Activewear leggings review and summer sale deal
Winner: These activewear leggings are comfortable, flattering — and only $19 in the summer sale
Headlines with generic filler ("stylish," "trendy," "perfect," "chic," "essentials," "won't break the bank") won only 31% of the time, so use terms like this sparingly, if at all. A brief line can often sound like a TikTok opening hook.
Loser: This foldable bucket hat is the perfect summer vacation accessory — and it's only $20
Winner: This $20 bucket hat is going viral
Want to know which version of your copy will win before you spend on it? Check out Anyword's predictive scoring. Book a demo today!
Headlines using the word "review" won 7% of the time. "Shop" as a command ("shop these pants!") won only 28% of the time, so use it sparingly and don't lead with it.
Loser: Women's hiking sandal review
Winner: I replaced my sneakers with these $23 sandals — here's how they held up
The winning headline was longer 62% of the time, with a median of 76 characters versus 62 (about 13 words versus 11).
But long doesn't mean padded: the long winners added specific detail, not adjectives. We didn't see this effect in primary text, where length didn't separate winners from losers (the medians were 102 vs 103 characters).
Loser: Sweatpants review
Winner: I wore these $28 sweatpants every day for a month, and they still haven’t pilled or lost their shape
None of these separated winners from losers in a meaningful way:
Primary text copy showed almost no pattern. Length didn't matter, and first-person voice leaned positive (60%) but not strongly. Only one trait stood out: experience verbs won 66% of the time. So test your variations before you launch.
Takeaway: Headline craft has repeatable patterns, but primary text needs to be tested before you launch.
1. Write in first person where it's natural.
2. Put a real number or price in.
3. Replace adjectives with a specific detail.
4. Cut "review," "shop" and filler words.
5. Use more words, not fewer.
6. Don't count on exclamation points or urgency to rescue weak copy.
Anyword: Marketing AI that knows what works — before you launch. Book a demo today to see how you can drive better results, faster.
Transform the way you do business with our AI writing tool. Sign up for free and explore our powerful tools.
.svg)
