CPG prompt of the day

Review mining: turn a thousand reviews into a product and copy fix list

Ratings are slipping and the team is guessing at why from the last ten reviews they happened to read.

You are a consumer insights analyst mining product reviews for actionable product and content fixes. I will paste review text. Produce:

A) A THEME TABLE — each recurring theme with its approximate share of the reviews it appears in, sentiment, star-rating association, and whether it is a PRODUCT issue, a PACK/DELIVERY issue, an EXPECTATION-SETTING issue (content promised something the product does not do), or a competitor comparison.
B) A FIX LIST split by owner — what R&D/ops must change versus what the PDP copy or imagery can fix this week, ranked by frequency × rating impact.
C) FIVE VERBATIMS worth quoting internally, chosen because they are representative, not because they are dramatic — each with the theme it evidences.

Input: [PASTE REVIEW TEXT WITH STAR RATINGS AND DATES; NOTE THE RETAILER AND ITEM]

Rules: Do not invent, estimate, or extrapolate any figure — if a number is not in the data I give you, write "not provided" and flag it. Mark every claim I should verify against my syndicated data or internal reporting before using it externally. Never include retailer-confidential terms or personally identifiable shopper data. Do not fabricate or paraphrase a verbatim into something the reviewer did not write, and remove any name or personal detail before quoting.

Why it works — The most valuable review signal is usually expectation-setting, not product failure — and it is fixable in an afternoon by rewriting the PDP. Splitting the fix list by owner keeps the fast fixes from being queued behind an R&D roadmap.