Walmart Reviews Scraper

Walmart product reviews → CSV / JSON / Excel in 2 minutes. Bypasses the 100-review default UI cap on walmart.com/reviews/product/<id>. 17 fields per review including image URLs and seller metadata.

For: BI dashboards, competitor monitoring, sentiment analysis, product-launch QA.

Output schema (17 fields)

Field	Type	Notes
productId	string	Walmart numeric product ID (extracted from URL or accepted directly)
page	integer	Pagination page on which this review was found
sourceUrl	string	Exact URL fetched (page + sort included)
reviewId	string | null	Walmart's stable review identifier
stars	integer	1 – 5
title	string | null	Review headline
body	string | null	Review body text
author	string | null	userNickname — display name (Walmart obfuscates real names)
date	string | null	Submission date (US-format M/D/YYYY as Walmart returns it)
verifiedPurchase	boolean | null	Derived from userBadges; null when Walmart omits the flag
helpfulCount	integer | null	positiveFeedback — number of "Helpful" votes
unhelpfulCount	integer | null	negativeFeedback — number of "Not Helpful" votes
recommended	boolean | null	Reviewer recommends product (Walmart-prompted yes/no)
imageUrls	string[]	Photos attached to this review (empty array when none)
fulfilledBy	string | null	E.g. "Walmart" / "Marketplace seller" — fulfilment side of purchase
sellerName	string | null	Seller of record at time of purchase
scrapedAt	integer	Unix epoch seconds — when this row was collected

Inputs

• productInputs (array, required) — list of full Walmart product URLs OR bare numeric product IDs. Mixed input is OK. Up to 50 products per run.
• maxReviewsPerProduct (integer, default 200, max 5000) — hard cap. If a product has fewer reviews than the cap, the scraper stops naturally; it does NOT pad.
• sortBy (enum, default most-recent) — most-recent | most-helpful | rating-high | rating-low.
• useProxy (boolean, default true) — strongly recommended ON.
• proxyConfiguration (object, default RESIDENTIAL) — Apify proxy group selector.
• maxConcurrency (integer, default 2) — parallel products. Reviews within a single product still paginate sequentially.
• requestDelayMs (integer, default 1500ms) — delay between sequential page fetches within a product.

Honest limitations

These are real, deliberate trade-offs — not bugs:

• Walmart aggressively rate-limits direct datacenter IPs. useProxy=true (RESIDENTIAL) is the default for a reason. Datacenter / no-proxy runs will see 403 / 429 within a few hundred requests.
• Reviews-per-page is fixed at 10 by Walmart. A 200-review product = 20 sequential page fetches = ~30 seconds at default 1500ms delay.
• Cap behavior. If a product has 47 reviews and maxReviewsPerProduct=200, you get 47 — no padding, no synthetic rows.
• verifiedPurchase is best-effort. Walmart sometimes omits the flag on older reviews. Missing → null.
• No auth, no seller dashboard reviews. Public review list only — what you'd see at walmart.com/reviews/product/<id> without logging in.
• 404 products are skipped. Deactivated, regional, or ID-typoed entries log a warning and continue with the next product.
• HTTP 403 / 429 aborts the current product (not the whole run) and continues. Retry the failed IDs in a separate run — repeated 4xx in the same session usually means proxy IP burnout.

How it works

The scraper fetches walmart.com/reviews/product/<id>?page=<n>&sort=<sort> and parses the embedded __NEXT_DATA__ JSON blob — not the rendered DOM. This is more stable across UI redesigns: as long as the Next.js page state is shipped with the HTML, the parser keeps working.

Canonical extraction path (verified 2026-04 against walmart.com): props.pageProps.initialData.data.reviews.customerReviews. If a future Walmart redesign moves the array, the parser falls back to a tree-walk that pattern-matches review records by their unique key set (reviewId + reviewText + rating) — so the scraper degrades gracefully rather than returning empty.

Cost framing (Apify pricing)

• Per product, full pagination to ~100 reviews = ~10 page fetches × ~3 KB each = ~30 KB transfer.
• Walmart pages are HTML-heavy (~150–250 KB each rendered, but the embedded JSON is what matters for parsing).
• Default RESIDENTIAL proxy traffic is the dominant cost driver — budget accordingly.

Use cases

• Competitor monitoring: track competitor product reviews daily, alert on rating drops or new negative themes.
• Product-launch QA: scrape your own product's reviews post-launch, build a sentiment dashboard.
• Market research: collect reviews across a category for thematic analysis (LDA, embeddings, GPT clustering).
• BI integrations: drop the dataset into BigQuery / Snowflake / DuckDB for ad-hoc analysis.

Related actors in this portfolio

Tool	Adds
Trustpilot Reviews	Cross-platform review parity (Trustpilot vs Walmart)
Reddit Discussion	Off-platform sentiment (where buyers complain about your product)
Google News	News coverage signals correlated with rating spikes

---

Need a custom data pipeline?

I build custom scrapers, ETL pipelines, and data-feed integrations. Pilot scope examples: a tailored parser for a specific Walmart category, a daily cron with Slack alerts, or a multi-marketplace aggregator.

📧 Email: spinov001@gmail.com 🌐 Portfolio: blog.spinov.online · apify.com/knotless_cadence 💬 Tips & tutorials: t.me/scraping_ai

Disclosure: I maintain Apify actors related to this topic; links above point to my Apify Store profile (commercial). I am not affiliated with Walmart Inc.