It's mid-November in Ho Chi Minh City. A Vietnamese drop-shipper is watching a TikTok-trending phone stand explode on Shopee.vn — 11.11 traffic, six sellers offering the same SKU, and the top listing is sitting at 4.8 stars in Taiwan but 3.1 stars in Malaysia. That's an arbitrage opportunity if the MY problem is logistics, and a hard pass if the problem is the product. The dashboard won't tell you which. The reviews will. This guide walks you through using the ExportComments Shopee Reviews exporter to pull the full review history for any product, across all 9 Shopee marketplaces, into Excel, CSV, or JSON.

Why export Shopee reviews

Shopee is the largest review pool in Southeast Asia and a fast-growing one in Latin America. Buyers leave reviews in seven languages across nine marketplaces, attach photos and videos generously, and Shopee's seller-reply rate is one of the explicit signals the platform's recommendation algorithm rewards. Once the data is offline, the moves you can make on it are the moves your competitors aren't:

  • Cross-border arbitrage — same SKU on .tw and .com.my; pivot the average rating by marketplace and find the gap. A 4.8 in Taiwan and a 3.1 in Malaysia almost always points to a logistics or localization fix you can win on.
  • Video review extraction — Shopee allows video reviews and they're the highest-trust buyer signal on the platform. Pull the videos column, sample the URLs, and you have unfiltered, real-buyer footage no marketing team can fake.
  • Variation pivot — pivot on variation. A 4.6 overall hides a 3.0 average for the "Black, Type-C, EU plug" variant your store actually sells.
  • Seller-reply audit — Shopee's algorithm rewards reply rate. Filter to rows where seller_reply_text is empty and you have a list of negative reviews you haven't answered yet — every one of them is a free algorithm boost.
  • 11.11 / 9.9 / 12.12 sale review-bomb pattern — pull the review feed weekly around mega-sales, watch for the spike of low-star reviews from buyers complaining about delayed shipping. The pattern is consistent enough that you can prep buyer-comms templates ahead of it.
  • Multi-marketplace brand monitoring — pull the same brand SKUs across .vn, .co.id, .tw, .co.th, .sg, .com.my, .ph, .com.br, and .com.mx in a single bulk run.

How to export Shopee reviews — step by step

Step 1: Grab the Shopee product URL

Open the product page on any Shopee TLD — for example https://shopee.vn/product-i.123456789.987654321 on Shopee Vietnam, or https://shopee.com.my/... on Malaysia. Any canonical product URL works. The TLD tells the exporter which marketplace to query and the IDs at the end address the listing.

Step 2: Paste the URL into the exporter

Open the Shopee Reviews exporter and paste the URL. Comparing the same SKU across multiple marketplaces, or auditing your full Shopee catalog? Switch to bulk mode and paste one URL per line. Bulk runs return one Excel file per URL, bundled in a single ZIP at the end of the job — each marketplace or product stays cleanly separated, which is exactly what you want for cross-border comparison.

Step 3: Pick a format

Excel (.xlsx), CSV, or JSON. Excel for fast pivots and charts. CSV for BI imports. JSON for notebooks, clustering, or LLM theme summarization on the review text.

Step 4: Start the export

Hit Export. The job runs server-side and paginates through the Shopee review feed for that product, capturing the buyer name, the rating, the review text, the language code, the variation, the photo URLs, the video URLs, the seller's reply, and the helpful count. Larger products with thousands of reviews take a couple of minutes — the file lands in your dashboard and your inbox when the run finishes.

Step 5: Open the file

Open the .xlsx in Excel, Numbers, or Google Sheets. Each row is one review, with the columns described next.

Inside the export — what fields you get

Each row is a single Shopee review. You get columns for:

  • buyer_name — the display name on the review.
  • rating — the 1–5 star score.
  • review_text — the full review body in the original language.
  • language — the original review language code.
  • variation — the variation the buyer chose (size, color, plug type, etc.).
  • photos — list of buyer-submitted photo URLs.
  • videos — list of buyer-submitted video URLs (Shopee allows video reviews).
  • seller_reply_text — the seller's reply, when one exists.
  • seller_reply_at — when the seller replied.
  • helpful_count — community helpful-vote count.
  • permalink — direct link to the individual review.
  • created_at — review date.

Common workflows

  • Cross-border arbitrage flag — bulk-pull the same SKU on .tw, .com.my, .vn, and .ph. Pivot the average rating by marketplace. A 4.8 in Taiwan and a 3.1 in Malaysia almost always splits cleanly into a logistics fix or a localization fix once you read the language column. Either one is a winnable problem.
  • Video review extraction — sort to rating = 5 and skim the videos column. Real buyer unboxings are the highest-trust social proof on Shopee. The exporter doesn't download the binaries — it gives you the URLs so you can sample what to ask for permission to reuse.
  • Variation pivot — pivot on variation. The seller's overall 4.6 hides which variant your store sells; the average for "Black, Type-C, EU plug" might be 3.0 because the EU plug is loose. Same product, completely different fix.
  • Seller-reply audit — filter to rows where seller_reply_text is empty and rating <= 3. Every row is a missed algorithm signal. Reply rate is one of the things Shopee's recommendation engine rewards explicitly, and the lift on negative reviews is the biggest because it's also where future buyers look first.
  • 11.11 / 9.9 / 12.12 sale monitoring — schedule a weekly export across the mega-sale window and watch for the predictable spike of low-star reviews complaining about delayed shipping. The pattern is consistent enough that you can prep canned buyer-comms responses ahead of the wave instead of getting buried by it.
  • Trending-product validation for dropshippers — when a product trends on TikTok, Vietnamese and Indonesian drop-shippers race to the Shopee listings. Pull the reviews of the top three sellers in your marketplace, filter to rating <= 2, and you'll know in five minutes which seller is shipping the actual product and which one is shipping the fake.

Plan limits and API access

The Free tier returns up to 100 reviews per export — enough to evaluate the format on a single listing. Personal scales to 5,000 results per export, Premium to 50,000, and Business to 250,000 — enough to capture the full review history of even the deepest Shopee listings. The same job is available through the REST API and via webhooks for scheduled or pipeline-triggered runs. See pricing for the full breakdown.

FAQ

  • Which Shopee marketplaces are supported?
    All 9 — .vn, .co.id, .tw, .co.th, .sg, .com.my, .ph, .com.br, and .com.mx. Marketplace detection is automatic from the URL.
  • Do I get video review URLs?
    Yes. The videos column carries the URLs of every buyer-submitted video on each review. The exporter doesn't download the binary files — open the URLs to inspect or to ask buyers for permission to reuse.
  • Can I see the variation the buyer ordered?
    Yes. The variation column carries the buyer's choice — color, size, plug type, spec — so you can pivot reviews per variant and isolate the one dragging your average down.
  • Are seller replies included?
    Yes. seller_reply_text and seller_reply_at capture the reply text and timestamp. Filter to empty seller_reply_text on negative reviews to find the easiest algorithm wins.
  • Can I schedule a weekly export across mega-sales?
    Yes, on Premium and Business. Pair scheduled exports with webhook delivery to push new reviews into Slack or BigQuery during 9.9, 11.11, and 12.12 — useful for catching the predictable shipping-complaint wave before it tanks your average.
  • What if I want the same SKU across multiple marketplaces?
    Use bulk mode. Paste one Shopee URL per marketplace and the run returns one file per URL packaged in a single ZIP, so each marketplace's data stays cleanly separated for cross-border comparison.