Give it a BestBuy product URL or a bare SKU and get that listing's customer reviews back as rows: the rating, the headline, the full text, the author and the date - plus whether the reviewer recommended the product, whether the purchase was verified, and how many people voted the review helpful or not. Eleven columns, one row per review.
The input is the listing, not the review - you name the products you want feedback on, pick the order, and the job collects what buyers wrote.
One row per review, each tagged with the query it came from - so a run across fifty products still reconciles back to your input list.
helpful and not_helpful are separate columns rather than one netted score, so a review that split the room is distinguishable from one nobody voted on. Alongside recommended and verified, that is four signals about a review beyond the rating and the text.
The form takes a product or reviews URL, or the bare SKU on its own line - so a column of SKUs in a spreadsheet is already a valid input list. Either way the sku arrives in its own column, pulled out of the address, which is what lets an export join back to a catalogue table.
In the JSON export the rating and the vote counts come back as numbers, recommended and verified as booleans, and the date as an ISO YYYY-MM-DD string. That is unusual enough in this family to be worth stating, and it means less cleaning before the first query.
Each row carries the review as BestBuy displays it, the product it belongs to, and the three flags shoppers attached to it.
Review text comes back verbatim, with no preprocessing applied - which is worth knowing before you write a parser, because the bodies contain line breaks, commas and double quotes. The column list below comes from three agreeing sources, and unlike most pages in this family the descriptions can tell you the types as well as the names, because the archived runs of this service actually contain reviews. Read the note under the table for what that sample does and does not settle.
The platform's own column list, in export order, matched against the JSON keys and the workbook header row of real runs. Types are as observed in the JSON export; the note underneath says what the sample does not settle.
YYYY-MM-DD form.This page can describe types because the archived runs of this service actually contain reviews - and the eleven names come from three agreeing sources: the platform's published column list, the JSON keys of those runs, and the header row of their workbooks. The types above are what the JSON holds: numbers for the rating and the vote counts, booleans for recommended and verified, an ISO date string, and free text that carries line breaks, commas and quotes. What the sample does not settle is anything about values. Every archived query pointed at a single product under the default relevance sort, which is plenty to establish the shape of a field and nowhere near enough to establish a rating scale, a distribution, or a range for the vote counts - so this page states none of those, and none of the archived queries used a bare SKU either. Pull one product on the free tier and read your own first rows before you fix a scale in code.
A per-query limit decides what a run costs and how much lands in the file. The sort order decides which reviews that limit spends itself on - both are chosen before the job starts.
A few examples of how teams use electronics review data to answer a question they actually have.
Sort by lowest rating and then rank what comes back by the helpful count. A one-star review nobody voted on is one person's bad day; a one-star review hundreds of shoppers found helpful is a product problem with an audience.
The sku column arrives filled whether you submitted a URL or a SKU, so a review export drops straight into a table keyed on SKU. That is the difference between a folder of review files and a dataset you can query alongside price and stock.
Feed the products you compete with and read the recommended and verified flags next to the ratings. What a category's buyers say, and how strongly other buyers back them, is a fuller picture than an average star count.
No subscription, no minimum, no recurring bill. Your first 500 rows are on us - after that, pay-as-you-go at the same flat rate as every other scraper here.
Every new account, one-time. No credit card required. Per-query limits, every sort order, file upload and every export format included.
Roughly $2 per 1,000 reviews. The pre-flight estimator shows the row count and credit cost before a run starts - no surprise bills, no compute units to translate.
Volume pricing, dedicated workers and an SLA for continuous monitoring or very large historical pulls. Tell us your numbers and we will quote.
LIVESCRAPER10 at checkout.The same idea at another big-box retailer, and it splits the helpful vote in two as well - helpful_positive and helpful_negative. Different column names, comparable signal.
The third place shoppers check before buying electronics. Running two of them across the same shortlist shows whether a complaint follows the product or the storefront.
Short answer: yes for the public review content - and this export carries nothing about the purchase behind it.
Customer reviews on BestBuy are published to be read. The rating, the headline, the review text, the date and the helpfulness votes are shown to anyone who opens the product page, signed in or not. Collecting publicly visible feedback for research is long-established practice, and nothing here touches a login, an account or a checkout.
The author field carries what the site itself publishes, which is a display name rather than a full identity - there is no email and no address in this output. The verified flag says only that the site marked a purchase as verified; nothing about the order, what was paid or when it shipped is collected. If you are processing the review text in the EU, the usual rules still apply to what you do with it downstream.
BestBuy's own terms restrict automated access, so this is a terms question as well as a legal one - if you have a contractual relationship with the retailer, check it. We run no third-party trackers on the data layer, and your exports auto-delete after 30 days.
The questions we hear most. Anything else? Talk to us - humans, not bots, write the answers.
500 one-time free rows on every new account - no expiry. After that it is $0.002 per row, pay-as-you-go - no card on file until you say so.
Livescraper's BestBuy Reviews Scraper turns a list of products into review data. You submit BestBuy product or reviews URLs, or bare SKUs - typed one per line, or uploaded as a CSV, XLSX, TXT or Parquet file - set a limit per query or clear it to take everything, choose the sort order, and download what buyers wrote as a clean CSV, Excel or JSON file. Requests go out through the proxy pool and your real IP is never used.
Each row carries the review as BestBuy displays it - the rating, the headline, the full text, the author and the date - plus the product's SKU and three flags shoppers attached to it: whether the reviewer recommended the product, whether the purchase was verified, and how many people voted the review helpful or unhelpful. Those last two are separate columns rather than one netted score, which is what keeps a divisive review distinguishable from one nobody voted on.
The sku column is the one that makes this joinable. It arrives filled whether you submitted a URL or a SKU, pulled out of the address, so a review export drops straight into a table keyed on SKU alongside price and stock rather than sitting in a folder of its own. And in the JSON the values come back typed - numbers for the rating and the vote counts, booleans for the two flags, an ISO date - so there is less cleaning before the first query.
One practical note up front: the review bodies are verbatim and contain line breaks, commas and double quotes, so quote your CSV properly or take the JSON. And while this page can describe the column types because the archived runs of this service contain real reviews, it deliberately states no rating scale, no distribution and no range for the vote counts - the archived queries all pointed at one product under the default sort, which settles shape and not values. Pull one product on the free tier and read your own first rows. Your first 500 are free and need no credit card, and after that it is $0.002 per row, flat.