← ClaudeAtlas

onsite-searchlisted

Turns onsite search query logs, including zero-result queries, into specific merchandising fixes: synonym rules, naming corrections, and collection-page gaps. Use when visitors search on the site but don't buy, when collection pages convert poorly, or when the team wants demand signal from real search behavior instead of a keyword tool. Boundary: differs from `conversion-funnel`, which diagnoses funnel drop-off against conversion benchmarks; this works specifically from search query exports, not funnel stage data.
sidchaudhary/gtm-skills · ★ 9 · Data & Documents · score 71
Install: claude install-skill sidchaudhary/gtm-skills
# The Search Merchandiser Turn onsite search query logs into specific fixes: synonym rules, naming corrections, and collection pages for demand the store isn't answering. > **Findings discipline.** Read `references/audit-findings-discipline.md` before writing the > output. It covers what happens to a finding after it is written: the audit's date and exact > scope, a re-audit trigger stated as an event, severity paired with effort so the list > resolves into a sequence, and a baseline captured before anything changes so the fixes are > attributable. Query findings need effort alongside volume, since a synonym rule and a new collection page are not the same cost. Where zero-result logging is off, instrumenting it is the prerequisite finding and belongs first. ## Before you write **Run the input list below before you write anything. If one of those inputs is missing, ask for it and stop. Do not return a draft with a warning on it.** The user copies the draft and leaves the warning behind, so a caveat protects you and not them. **Ask at most THREE questions. Hard cap.** Before anything becomes a question, get it yourself: read `.agents/product-context.md`, fetch the site or page they named, compute it from numbers they already gave, or look up the platform default. Whatever is left after that, and everything past the third question, becomes a stated assumption the user corrects in one word rather than a question that stops the work. Number them, and say what you will assume i