Customer reviews and star ratings show how products, sellers and businesses are actually received. Infovium collects publicly available reviews and ratings from ecommerce sites, marketplaces and business listing pages, cleans them, and delivers them as structured datasets for analysis. It can be a one-time collection or a recurring feed that tracks how reviews change over time.
What review and rating scraping covers
- Product review scraping: reviews and ratings for products listed on retail and brand websites.
- Ecommerce and marketplace reviews: product and seller feedback from online stores and marketplaces.
- Business and location reviews: ratings and review text for companies, branches, venues and service providers shown on public listing pages.
- Customer rating extraction: star ratings, rating counts and rating distributions where the source displays them.
- Competitor review intelligence: the same fields collected for your products and your competitors’ products, so they can be compared.
Review data fields
The fields depend on what each source publishes and are agreed in your project scope.
| Field | What it contains |
|---|---|
| Review title | The headline of the review, where the source has one |
| Review text | The body of the review |
| Star rating | The rating given with the review |
| Review date | The date shown for the review |
| Product or business identifier | The product name, SKU, listing ID or business name the review belongs to |
| Reviewer display name | The public display name, only where it is publicly available and appropriate to collect |
| Verified purchase indicator | Whether the source marks the review as a verified purchase, where available |
| Review URL | The page where the review appears |
| Source | The website or marketplace the review was collected from |
| Collection timestamp | When the review was collected |
Not every field is available on every source. We confirm what each source provides before the project starts.
What we do not collect
We collect only reviews and ratings that are publicly visible. We do not offer access to private reviews, logged-in or restricted data, or personal information beyond a public display name where that is appropriate. Reviewer details are minimized according to the project’s requirements.
Review monitoring, history and trends
| Option | Suited to | How it works |
|---|---|---|
| One-time collection | Product research, a market study, a competitor snapshot | Reviews for an agreed list of products, businesses or sources are collected once and delivered as a dataset |
| Recurring monitoring | Quality tracking, competitor watching, trend analysis | Collection is repeated on an agreed schedule. New reviews are added, and each run is timestamped so review history and rating trends can be analyzed |
The schedule depends on the source and on feasibility. Recurring collection is run as a managed data service, which includes monitoring the collection and maintaining it when a source changes.
Cleaning, normalization and deduplication
Raw review data is messy. The same review can appear more than once, dates and ratings can use different formats, and text can contain stray markup. Review data is subject to quality checks appropriate to the project requirements, which can include:
- removing duplicate reviews collected more than once;
- standardizing dates, rating scales and language or country codes;
- cleaning markup and encoding problems in the review text;
- matching the same product or business across sources, see entity matching and resolution.
See data cleaning, validation and normalization.
Sentiment analysis, where separately scoped
Collecting reviews and analyzing them are different jobs. If you want customer sentiment, topic or theme labels added to the review text, we can scope that as a separate task, using AI-assisted processing where it fits. It is not part of a standard collection project unless it is agreed in the scope. See AI-powered data extraction.
Business uses for review data
Product research
See what customers praise or complain about for products in a category, before you build, source or stock them.
Customer experience analysis
Group review text by topic and rating to find recurring problems, such as sizing, delivery or support.
Competitive intelligence
Compare ratings, review volume and review themes for your products against competing products over time.
Quality monitoring
Track rating changes for your own products or locations, and spot a drop soon after it begins.
AI-ready review datasets
Prepare review text and ratings as clean, labeled-ready datasets for machine learning and language model projects. See AI-ready datasets.
How a review data project works
- You send the products, businesses or sources you want covered and the fields you need.
- We review feasibility and confirm which fields each source provides.
- We agree the scope, schema, volume and schedule in writing.
- We collect the reviews, clean and deduplicate them, and check the output against the agreed schema.
- You receive the data in your chosen format, once or on a recurring schedule.
Delivery formats
Review data can be delivered as CSV, Excel/XLS, XML, HTML, JSON or JSON Lines files, through a REST API, over SFTP or FTP, to cloud storage or directly into a database, depending on project requirements. See data delivery.
Related services
- Scraping product prices and reviews
- Web scraping services
- Web data collection and extraction
- Data enrichment services
Review and rating scraping FAQ
What review data can be collected?
Typical fields are the review title, review text, star rating, review date, product or business identifier, review URL and source. A public reviewer display name and a verified purchase indicator can be included where the source shows them and collecting them is appropriate. The exact fields depend on the source and your scope.
Can product and business reviews be extracted?
Yes. We can collect product reviews from ecommerce sites and marketplaces, and public reviews and ratings for businesses and locations shown on listing pages. Only publicly visible reviews are collected.
Can review data be updated regularly?
Yes. Recurring collection adds new reviews on an agreed schedule and timestamps each run, so you can analyze review history and rating trends. The schedule depends on the source and on feasibility.
Can reviews be cleaned and deduplicated?
Yes. Duplicate reviews can be removed, and dates, rating scales and text can be standardized, as part of quality checks appropriate to the project requirements.
Is sentiment analysis included?
Not by default. Sentiment or topic labeling is scoped separately, so you can choose whether you want collected review data only or analyzed data as well.
What delivery formats are available?
CSV, Excel/XLS, XML, HTML, JSON and JSON Lines files, REST API, SFTP or FTP, cloud storage and database delivery, depending on project requirements.
Can you get private reviews or customer details?
No. We collect only publicly visible reviews and ratings. We do not provide private reviews, restricted data or personal information.
Tell us what reviews you need
Send the products, businesses or sources you want covered and the fields you need, and we will reply with questions or a written scope.