How do I analyze Coupang and Olive Young reviews in English (without reading Korean)?
Short answer: export the reviews with a Coupang (쿠팡) or Olive Young (올리브영) review scraper, then label each one by complaint type, sentiment and purchase motive with English keys — and count the labels instead of translating every review. Korean Review Classifier does the labelling step for any review scraper's dataset. It costs $0.50 per 1,000 reviews, with no prompts and no LLM key. It does not scrape the stores itself.
Korean reviews are long, full of slang, and often mix several complaints in one line. A team that does not read Korean usually ends up with a machine translation it still has to read row by row. Labels with probabilities can be filtered and charted instead.
What you get
Rows from the classifier's public example run (2026-10-04), with one custom label, "mentions the smell". English in parentheses is added for readers — the tool returns the original Korean text.
| text (English added) | complaint (probability) | sentiment | motive | mentions the smell |
|---|---|---|---|---|
| 한 달도 안 돼서 충전이 안 됩니다. 교환 요청했는데 답이 없어요. (Stopped charging in under a month. Asked for an exchange, no reply.) | quality_defect (0.95), customer_service (0.92) | negative | unknown | false |
| 배송이 일주일이나 걸렸고 박스가 다 찢어져서 왔어요. (Delivery took a week and the box arrived torn.) | packaging (0.97), delivery (0.96) | negative | unknown | false |
| 향이 너무 강해서 두통이 와요. 가격 대비 별로입니다. (The scent is so strong it gives me a headache. Not worth the price.) | other (0.76), price_value (0.74), skin_trouble (0.70) | negative | unknown | true |
| 항상 쓰던 거라 이번에도 재구매했어요. 할인할 때 사니 좋네요! (My usual product, bought it again. Nice to get it on sale!) | none (0.98) | positive | repurchase | false |
Complaint types are delivery, quality/defect, size/fit, price/value, customer service, packaging, no effect, skin trouble and other — each with a probability, and each with a Korean name too (quality_defect / 품질·불량), so the same table works for a Korean team. Purchase motive covers price/discount, reviews and reputation, brand, gift and repurchase.
How to set it up (5 minutes)
- Export the reviews. Run a review scraper on Apify Store. Field detection was checked against real output of, for example, Coupang Review Scraper (zen-studio), Olive Young Review Scraper (cocomoon) for the Korean store, and Olive Young Review Scraper (solidcode) for Olive Young Global.
- Open the classifier on Apify: Korean Review Classifier.
- Pass the scraper's dataset. The text, rating and ID fields are detected automatically. Add up to 10 yes/no labels of your own, in English or Korean:
To try it without a dataset, the form default classifies two pasted reviews for about $0.001.{ "datasetId": "YOUR_REVIEW_DATASET_ID", "extraTextFields": ["title"], "customLabels": ["mentions the smell", "says they will buy again"] } - Filter and count. Sort by complaint type to see why buyers leave 1-star reviews, or compare the complaint mix of your product against a competitor's.
- Schedule it with Apify Schedules to classify a scraper's new reviews every day.
For scale: 1,000 reviews cost $0.50 and 10,000 cost $5.00; Apify's free $5 monthly credit covers about 10,000 reviews. About 100 reviews take 5 seconds. Rows without review text, such as the summary rows some scrapers add, are skipped and not charged.
You can also ask Claude, Cursor or Claude Code through the Apify MCP server — "Classify dataset abc123 from my Coupang scraper run and summarize the complaints." Setup is on the tool page.
How accurate is it?
On 100 hand-labelled Korean reviews from Coupang, Naver Shopping and Naver Place (47 of them with 1–4 stars), complaint type matched 88% of the time and sentiment 96%. It is built and tested for Korean; English reviews also work. Automated labels can be wrong — check a sample before a big decision.
Other ways, and their limits
- Machine-translate and read. Works for a few dozen reviews. Past that, you are still reading every row, and translation flattens the slang that carries the complaint.
- Count star ratings. Fast, but a 4-star review can still mention skin trouble, and stars do not say whether the problem was delivery or the product.
- Prompt a general chatbot. It can summarize a pasted batch, but you maintain the prompt and turn free text back into a table. This tool returns the same labels with probabilities every run.
Independent tool — not affiliated with Coupang, CJ Olive Young, NAVER Corp. or the authors of the scrapers above. More Naver, AliExpress and K-beauty data tools: leoworks.kr/tools.