Verified against ChatGPT · 2026-08-14
Map backend and title keywords for a marketplace listing without stuffing the title into gibberish
Builds a keyword placement map for a marketplace product listing — title, backend search terms, and bullet points — that respects the platform's actual character limits and search-relevance rules.
The prompt
Ready to copy — highlighted parts are example details you can swap.
Build a keyword placement map for one marketplace product listing. The output has to respect the actual character limits and indexing rules of the named platform — a keyword list on its own is not the deliverable, placement is. PRODUCT A stainless steel 32oz insulated water bottle with a straw lid. MARKETPLACE Amazon US, Sports & Outdoors category. CANDIDATE KEYWORDS (with rough search-volume signal if known) 'insulated water bottle' (high), 'water bottle with straw' (high), '32 oz tumbler' (medium), 'gym water bottle' (medium), 'leak proof' (low). CURRENT TITLE "Water Bottle 32oz Stainless Steel Insulated with Straw Lid Leak Proof Gym Sports" CHARACTER LIMITS 200-character title limit, 249-byte backend search term field. STEP 1 — SORT KEYWORDS BY PLACEMENT Sort the candidate keywords into three buckets: title-worthy (high relevance, should be human-readable in the title), bullet-worthy (relevant but would make the title unreadable if forced in), and backend-only (relevant for indexing but not natural language a shopper would read). Say explicitly why each keyword landed where it did. STEP 2 — WRITE THE TITLE Rewrite the title using only the title-worthy keywords, staying within the stated character limit, and keep it readable as an actual sentence a shopper would parse at a glance — do not chain keywords with no grammar between them just to fit more in. If the current title already does this reasonably well, say so rather than rewriting for the sake of rewriting. STEP 3 — BULLET POINTS Draft bullet points that each work in one or two bullet-worthy keywords naturally, phrased as a benefit or spec a buyer would actually read, not a keyword dump with commas. STEP 4 — BACKEND SEARCH TERMS List the backend-only keywords as a comma-or-space-separated string within the platform's actual backend character limit, removing any duplicate word stems already covered by the title or bullets (repeating a word already indexed elsewhere wastes limited backend character space). OUTPUT FORMAT The four steps as labeled sections in order, with the final title and bullets shown as ready-to-paste text.
Customize
Optional — swap in your own details for the highlighted parts above.
Why this works
Sorting keywords into title/bullet/backend buckets before writing anything counters GPT-5.1's default behavior when asked to 'optimize a listing title with these keywords,' which is to chain every given keyword into the title regardless of whether the result reads as a sentence — that produces the exact keyword-stuffed, comma-free title strings that dominate marketplace listings and that marketplace ranking algorithms increasingly discount relative to natural-language titles with strong click-through, since a title nobody can parse at a glance suppresses the click-through signal the algorithm also weighs. Explicitly permitting the model to say the current title is already fine, rather than always producing a rewrite, matters because a model asked to 'improve' something will almost always generate a changed version even when the original was adequate, since producing no change can read as an incomplete response — stating this permission upfront removes that pressure and lets a genuinely fine title stay as is. Requiring duplicate word stems to be stripped from the backend terms once they're covered by the title or bullets addresses a specific and well-documented Amazon indexing rule: backend search term fields have a hard byte limit, and most marketplace search algorithms already index visible on-page text (title, bullets) separately from backend terms, so repeating an already-covered word there is a pure waste of a scarce, non-renewable character budget that could otherwise hold a genuinely new search term the listing isn't yet indexed for.
What you get back
Step 2 title: "32oz Insulated Water Bottle with Straw Lid – Leak-Proof Stainless Steel, Gym & Sports" (98 chars, within limit, reads as a sentence). Step 4 backend terms: "tumbler 32 oz sports outdoor hydration flask" — duplicate stems like 'water bottle' and 'insulated' dropped since they're already indexed via the title.
Verified against
ChatGPT GPT-5.1 · 2026-08-14
Changelog
- 2026-08-14 — Initial publish, verified against ChatGPT GPT-5.1.
Need this built into your business?
If a prompt isn't enough — Google Ads management, built and maintained for you — that's Scult's day job.
EXPLORE GOOGLE ADS MANAGEMENT
