User Feedback Analyzer

The User Feedback Analyzer is a Claude skill that takes a file of product reviews and hands back a themed report a founder can act on. Drop in a CSV of App Store reviews, a Play Store export, a G2 dump, or a markdown file of YouTube comments. Get patterns, priorities, and verbatim quotes — not a vague vibe.

The Problem

You scraped the reviews. Now you have a spreadsheet with 487 rows and no idea what to do with it.

You scroll. You read the first ten. You highlight a few in yellow. You open a fresh doc and start typing notes.... "lots of crashes," "people want dark mode," "someone mentioned Notion three times." By row 80 you've lost the thread. By row 200 you're skimming. By row 400 you've stopped reading entirely and you're just looking for the worst one-stars because at least those are fast.

An hour later you close the file. You have a vague impression that "users are mostly happy but the latest update broke something." You can't remember which version. You can't remember which quote made you write that down. You don't have a list. You can't show your team anything. So next week, you do it again.

This skill clears the fog. Hand Claude the file... get back a themed breakdown of every review, topped with a clear, ranked plan of what to do next.

What this Claude skill does

Classifies every review into one of 12 categories: Praise & Love, Bug Reports, Feature Requests, Pricing & Value, Competitor Comparisons, UX & Onboarding, Performance, Customer Support, and five more.
Tags sentiment independently: positive, negative, mixed, neutral — so a 5-star review that hides a complaint doesn't slip through.
Rolls up the top patterns: top 10 pros, top 10 cons, top feature requests, top bugs (with affected app versions), and every competitor named.
Writes a Priorities & Focus section: Fix First, Ship Next, Double Down, Investigate — ranked, with evidence and the cost of ignoring each item.
Quotes verbatim: every theme is backed by the actual reviews underneath it, sorted by relevance.
Works on CSV, markdown, TSV, or Excel: auto-detects the format and the column that holds the review body.

Why it's amazing!

Read it in 60 seconds and understand exactly where the customers stand.
Stop manually tagging reviews in a spreadsheet at 11pm.
Catch the version-specific regression hiding in the one-star pile.
See which competitor your churned users keep naming and why.

Who it's for

Anyone trying to make a product decision from messy customer feedback. SaaS founders staring at their first 100 App Store reviews, PMs running a quarterly user-voice readout, founders doing pre-launch competitive research on a rival's review pile, support leads spotting refund-driver bugs, and anyone who's ever paid for a "voice of customer" report and gotten 50 pages of nothing.

How it works

Ask Claude something like:

"Analyze the reviews in habit-app-reviews.csv"
"Group these App Store reviews by theme and tell me what to fix first"
"I just scraped 500 reviews from a competitor, what's the story?"
"Pros and cons of LinearApp from these G2 reviews"
"I'm deciding whether to ship dark mode or fix sync next quarter. Here's the review file... what should I do?"
"Analyze these reviews but only the 1-star and 2-star ones, I want to see the pain"

You'll get a markdown report back with sections like:

Top-line insight: "Bug Reports are 18% of reviews. 80% name the 2.4.1 release — version-specific regression."
🔴 Fix First: Sync crash on shared notes (v2.4.1) — 9 reviews, all iOS 17.x, all in the last 30 days.
🟡 Ship Next: Dark mode — 12 requests across iOS and Android over 6 months.
🟢 Double Down: Fast cross-device sync — 23 mentions, the #1 praise theme.
🔵 Investigate: Why are power users churning to Obsidian? 3 reviews mention the switch, none say why.
Top Bugs Reported: Each with affected app version and quote count.
Competitors Named: Every rival product mentioned, with sample quotes.
Per-category sections: Every theme followed by the actual reviews, verbatim, with author and rating.

Every claim in the report traces back to a quoted review underneath it. The categorization is the analysis. The words are the proof.

Pairs well with these other skills

This skill analyzes a file you hand it, so it works best at the end of a scrape — pull the feedback first, then point this at the output:

Apple App Store Reviews Scraper Skill: Pull up to 500 iOS reviews per country into a file, then feed that file straight in for a themed, ranked punch list. The natural first step before this one.
Google Play Reviews Scraper Skill: Scrape the full Android backlog — thousands of reviews — and run them through here to see what's breaking on the other platform.
YouTube Comments Scraper Skill: Grab the comments off a launch video or competitor's channel and analyze them as feedback, same as any review pile.
ICP Generator Skill: Once the report surfaces the real pains and desires in customers' own words, hand those verbatim quotes over to sharpen your ideal customer profile.

FAQ

Does it cost anything to run?

Only a Claude subscription, it reads a local file of reviews.

How accurate is the categorization?

Reliable on English-language reviews with clear signal. Long mixed reviews get classified by the strongest actionable signal (a crash beats a vague complaint) and the secondary themes are noted in the per-review gist. Single-emoji and gibberish reviews go to "General" and don't pollute the real buckets.

Can it scrape reviews for me?

No. It analyzes a file you hand it. Use the companion apple-app-store-reviews-scraper skill for iOS reviews or youtube-comments-scraper for YouTube comments, then point this skill at the output.

Does it work on Reddit threads?

No. Use the reddit-comment-theme-analyzer skill for Reddit — it uses a different category set tuned for demand signals like "Money Talk" and "Solution Requests."

How many reviews can it handle?

Hundreds in one run, no problem. Larger files get batched internally so nothing gets dropped. Every non-skipped review ends up classified.

Does it remember what I analyzed before?

No. Each run is independent. Re-running with the same file overwrites the previous report. If you want a longitudinal view, keep the dated output files yourself.

Does it handle non-English reviews?

Best on English. It will still attempt classification on other languages but the pattern statements and rollups assume English phrasing, so accuracy drops.

Can I focus on just the bad reviews?

Yes. For example just ask Claude to focus on the 2 star reviews and it'll skip everything above two stars. Useful for "what's burning right now" reads on a noisy review pile.

Safe To Install

Please read before installing. Any skill you copy online is someone else's code running inside your Claude, and a lot of what's out there should give you pause. Most authors who give skills away don't give a second thought to quality or security. Most are vibe-coded slop pushed out untested, and some are the occasional bad actor hiding an exploit in plain sight. (Read about OpenClaws ClawHub security nightmare). Every skill we publish is vetted by a programmer with 20+ years of professional experience and checked line by line for anything that touches your files without reason. With Stim-Pack Studios Claude Skills, you get peace of mind knowing you're installing safe, vetted code, not a stranger's guess.

Add it to your Claude workflow

If you spend any time reading customer feedback to decide what to build next, this one earns its slot. It collapses the spreadsheet-tagging-and-vibing process into a single conversation.

Get the User Feedback Analyzer Claude skill and turn your next review dump into a decision in 60 seconds.