Listings & SEO · from a Blog post, Jul 4, 2026 · Added
75-Character Titles and Item Highlights with a Claude Skill
Decide which words stay in a shorter Amazon title and which move to Item Highlights, backend or bullets
He says he tested it across multiple products and about 90% of the words it ranked matched what he would have chosen by hand, and that for brands with dozens or hundreds of products it turns weeks of work into hours.
Per source Self-reported by the source, not independently verified.
Copy-paste prompt
Here are my SQP and Search Term Reports for ASIN [B0...]: [attach]. Take only queries with at least one order. Split them into single words, and score each word by (a) how many converting queries contain it and (b) total orders from those queries, weighting frequency 3:1 over orders. Then fill a 75-character title, a 125-character Item Highlights line, and a leftover list for backend terms or bullets. Show the score table and both drafts.
The source didn't publish its exact prompt. We wrote this starter so you can try the same workflow. Replace anything in [brackets].
What they did
Zsolt Daragus built a Claude skill for Amazon's shorter titles. It takes every search query that converted for a product from SQP and Search Term Reports, splits them into single words, scores each word on how many converting queries it appears in and how many orders it played a role in (weighted 3:1 toward frequency), then allocates words into the title (75 characters), Item Highlights (125 characters) and leftovers for backend or bullets.
How it went
He says he tested it across multiple products and about 90% of the words it ranked matched what he would have chosen by hand, and that for brands with dozens or hundreds of products it turns weeks of work into hours. The post includes a video of a full run on a real product.
We only report what the source says. Results are self-reported unless noted.
Worth knowing
LinkedIn post (about July 4, 2026), self-reported. He offers the skill file to commenters. He argues search volume, conversion rate and CTR each mislead on their own for this decision.