Paste your ASIN and up to 10 competitors. ReviewIQ collects every review, image and customer photo, reads them with an AI analyst, maps who wins which search keyword, and hands you a strategy: where you win, where you lose, and exactly how to win more.
Why · Features · How it works · Quick start · Costs · Sample report · About
Reading hundreds of competitor reviews by hand takes days, and it's easy to miss what matters. Spreadsheet exports don't tell you why customers love or hate a product, and keyword tools don't tell you what to say to win the click.
ReviewIQ puts both together:
| Without ReviewIQ | With ReviewIQ |
|---|---|
| Skim a few top reviews per product | Every review from 5★ to 1★, for you and 10 competitors, in one database |
| "People seem to complain about quality" | e.g. "12.4% of all customers complain about battery life, vs 3.1% for the best competitor" |
| Guess which keywords to target | See every search term you win, lose, or could win, with volume and suggested bid |
| Generic listing advice | A rewritten title & bullets, a PPC plan, and a 30/60/90-day plan built from real evidence |
Built for Amazon brand owners, agencies and product teams who make listing, PPC and product decisions with real money.
- Amazon competitor analysis - see exactly why shoppers pick a competitor over you, aspect by aspect.
- Amazon review scraper + sentiment analysis - every 5★ to 1★ review with emotion, pain points and praise tagged by AI.
- Listing optimization - rewrite your title, bullets, images and A+ content around real customer language and objections.
- Amazon keyword research & SEO - find striking-distance keywords, competitor keyword gaps and open white space.
- Amazon PPC strategy - defend, attack and conquest campaigns with suggested bids.
- Product development - turn recurring 1★–3★ complaints into your next product improvement.
- Agency client audits - produce a full review-intelligence report for a brand in under an hour.
 📥 Collect - Enter ASINs or product links. Choose how deep to go (up to ~400+ reviews per star level). See the cost before you start. |
 🛍️ Products - Side-by-side listings: price, rating, star breakdown, full-size gallery, bullets, specs and Amazon's own AI review summary. |
 💬 Reviews - Every review as a readable card or a table: stars, date, verified, helpful votes, customer photos, plus AI-tagged emotion, pain point and praise. |
 📸 Customer photos - A wall of what buyers actually photograph. Low-star photos reveal defects and packaging problems fast. |
 📊 Insights - Complaints & praise by product aspect, emotional mix (EQ), and a win / lose table vs competitors, all weighted to the real star mix. |
 🔎 Search keywords - Share of search voice, the keywords you win, the ones competitors win, striking-distance terms and open white space (via DataDive). |
 🧠 AI Analyst - One click produces the full strategy report and a prioritized action plan with impact, effort, timeframe and KPI for every action. |
|
flowchart LR
A[Your ASIN +<br/>up to 10 competitors] --> B[Apify scraper<br/>5★→1★, recent + helpful<br/>+ keyword views]
B --> C[(SQLite database<br/>reviews · images · photos<br/>product data · coverage)]
D[DataDive API<br/>search volume + every<br/>product's organic rank] --> C
C --> E[AI Tagger - Claude<br/>emotion · sentiment · aspects<br/>pain point · praise · persona]
E --> F[Metrics engine<br/>star-weighted % · win/lose<br/>share of voice · keyword gaps]
F --> G[AI Strategist - Claude<br/>full report + action plan]
G --> H[Dashboard · Markdown · Excel]
- Collect - Amazon only shows ~100 reviews per filter view, so ReviewIQ requests several views (each star level × most recent and most helpful, plus optional keyword searches) and merges them, de-duplicating by review ID. Listing images and customer photos are saved at full resolution.
- Store - Everything goes into a local SQLite database. Re-running never duplicates rows; it updates them.
- Tag - Claude reads every review and tags the dominant emotion, sentiment, product aspects, the specific pain point or praise, use case, buyer persona and purchase driver.
- Measure - Because reviews are sampled evenly per star, raw counts over-represent 1★ reviews. Every "% of customers" figure is re-weighted by Amazon's real star breakdown, so the numbers reflect the whole customer base.
- Keywords - DataDive supplies the niche's search terms with volume, suggested bid and each product's organic rank. ReviewIQ classifies every keyword (win · contested · striking distance · gap · open) and estimates share of search voice.
- Strategize - A second Claude pass reads only computed evidence (so numbers are quoted, not invented) and writes the report.
📄 Read a full sample report → (Apple AirTag vs 4 competitors)
| # | Section | What it answers |
|---|---|---|
| 1 | Executive summary | Where you win, where you lose, the biggest opportunity, the #1 move |
| 2 | Market snapshot | Price, rating, sales, listing assets for every product |
| 3 | Where we win vs lose with customers | Per aspect: your complaint & praise rate vs competitors, with quotes |
| 4 | Customer emotional intelligence (EQ) | What triggers delight vs frustration, and the emotional job the product is hired for |
| 5 | Pain-point deep dive | Root causes; product defects vs expectation gaps |
| 6 | Search keyword battlefield | Share of voice, keywords you own, keywords each competitor pulls, gaps, white space |
| 7 | Where we can win | Rewritten title & bullets, PPC plan with bids, messaging that exploits competitor weaknesses, product fixes |
| 8 | 30/60/90-day plan | Actions with KPIs and targets |
| 9 | Risks & data limits | What to watch and how confident to be |
Requirements: macOS or Linux, uv, and API keys for Apify and Anthropic. DataDive is optional (needed for search keywords; Standard or Enterprise plan).
git clone https://github.com/Only-Pro-Marketer/Review-IQ.git
cd Review-IQ
cp .env.example .env # then paste your keys into .env
./run.sh # installs everything on first run, then opens the dashboardOpen http://localhost:8501.
- Collect - pick the marketplace, enter your ASIN and competitors, choose a depth, click Collect now.
- Keywords - Find my niches in DataDive → Pull keywords (or create a new dive seeded with your ASIN).
- AI Analyst - Tag reviews, then Write full analysis report.
- Export - Excel workbook (15+ sheets), reviews CSV, or the report as Markdown.
| Step | Typical cost |
|---|---|
| Reviews (Apify, ~$0.005 per review) | Standard: ≤ $27.50 for 11 products · Extended: ≤ $55 · usually 30–60% less in practice |
| AI tagging (Claude Opus 5) | ≈ $1.50–2.00 per 1,000 reviews |
| Strategy report | ≈ $1–3 |
| DataDive keywords | Uses your DataDive plan (a new dive costs ~1 token per competitor) |
The app shows an estimate before every paid step. You can switch the analyst to Claude Sonnet 5 to cut AI costs.
- Review coverage - Amazon exposes ~100 reviews per filter view. Extended depth (recent + helpful) gets ~170 per star; Maximum adds keyword searches (measured 400+ per star). The Coverage table shows collected vs available for every star, so partial data is never hidden.
- Weighted percentages - "% of customers" is re-weighted by Amazon's star mix. Figures marked "(sample)" are not.
- Share of voice - uses an estimated click-through curve by organic position. Treat it as directional.
- Currency - prices are shown in the marketplace currency. Never compare marketplaces without converting.
- API keys live only in your local
.env, which is git-ignored. Nothing is sent anywhere except the three APIs you configure. - Your database, downloaded images and reports stay on your machine (
data/is git-ignored). - Reviewer names and profile IDs are not collected (
includeGdprSensitive: false). - Test fixtures contain anonymized sample data only.
app.py Streamlit dashboard (9 pages)
review_intel/
apify.py Apify client: multi-view collection, retries, pagination
parse.py Raw scraper output → clean records (full-size photos, dates, votes)
db.py SQLite schema + idempotent upserts
metrics.py Coverage, star re-weighting, trends, aspect & emotion prevalence
keywords.py Keyword battlefield, share of voice, customer win/lose
datadive.py DataDive API client
analyst.py AI tagger + strategist (Claude, structured JSON output)
images.py · export.py Image downloads · Excel export
ui.py Theme and visual components
tests/ 54 tests (parsing, storage, math, API clients, every UI page)
Run the tests:
.venv/bin/pytest -qPro Marketer is a Toronto-based ecommerce growth agency. We help Amazon, Shopify and DTC brands win with data-driven marketing: Amazon SEO & PPC, paid social, email, CRO and review intelligence like ReviewIQ.
Want us to run this analysis for your brand, or build custom tools for your team? Get in touch:
| 🌐 Website | promarketer.ca |
| ✉️ Email | info@promarketer.ca |
| @onlypromarketer | |
| OnlyProMarketer | |
| ✖️ X / Twitter | @onlypromarketer |
| @onlypromarketer | |
| 🎵 TikTok | @promarketer |
⭐ If ReviewIQ helps you, star the repo and share it!
MIT © 2026 Pro Marketer. You're free to use, modify and share it; please keep the copyright notice.
Made with ☕ in Toronto by Pro Marketer
Not affiliated with Amazon, Apify, Anthropic or DataDive. Use responsibly and in line with each platform's terms.