Coupang Category Recommendation API is a service that recommends the Coupang category (displayCategoryCode) that best fits the product information (product name, brand, attribute, etc.) input by the seller. It is a machine learning service that leverages category data of previously registered products.
Note that it is essential for the seller to enter accurate information in order to receive appropriate category recommendations.
It is up to the seller to decide whether or not to accept the category recommended by the API. Before making the decision, note that the fee charged is different by category and that Coupang does not take responsibility for any disadvantages incurred by inaccurate results produced by Category Recommendation API.
◼︎ Example of Category Recommendation API:
Product information : “[Yuhan] Yuhan Zen powder detergent 1kg container (sterilized bleach)x10pcs”
Coupang Category Recommendation API first identifies category1 that best matches product information, which is "HPC" in this case.
It then identifies category2, "detergent", and the following sub-categories to finally recommend “HPC>Detergent>Laundry Detergent>Bleach>Powder” for the relevant product.
Since the recommendation is generated solely based on the input information, it is crucial that the user enters in accurate and detailed product information to receive the most fitting recommendation.
For example, if the user oversimplifies product information to "[Yuhan] Yuhan Zen (sterilized bleach)" in the above example, the system will not be able to identify whether the product is powder or liquid since the relevant data was not provided, which may result in inaccurate recommendation.
The same can happen if the user enters in more than 1 keyword for a single product (e.g. “[Yuhan] Yuhan Zen liquid/powder (sterilized bleach)”). This can make it confusing for the system to identify the accurate product category.
Product information : “[Yuhan] Yuhan Zen powder detergent 1kg container (sterilized bleach)x10pcs”
Coupang Category Recommendation API first identifies category1 that best matches product information, which is "HPC" in this case.
It then identifies category2, "detergent", and the following sub-categories to finally recommend “HPC>Detergent>Laundry Detergent>Bleach>Powder” for the relevant product.
| Input Information | Recommendation |
|---|---|
| [Yuhan] Yuhan Zen powder detergent 1kg container (sterilized bleach)x10pcs | HPC>Detergent>Laundry Detergent>Bleach>Powder (5depth leaf catecode: 63955) |
| [Yuhan] Yuhan Zen liquid detergent 1.8L refill (sterilized bleach) | HPC>Detergent>Laundry Detergent>Bleach>Liquid (5depth leaf catecode: 63954) |
| [Yuhan] Yuhan Clorox power gel 1L (sterilization/odor treatment)x10개 | HPC>Detergent>Cleaning Detergent>Clorox/Sterilizer (4depth leaf catecode: 63922) |
For example, if the user oversimplifies product information to "[Yuhan] Yuhan Zen (sterilized bleach)" in the above example, the system will not be able to identify whether the product is powder or liquid since the relevant data was not provided, which may result in inaccurate recommendation.
The same can happen if the user enters in more than 1 keyword for a single product (e.g. “[Yuhan] Yuhan Zen liquid/powder (sterilized bleach)”). This can make it confusing for the system to identify the accurate product category.
To prevent such issues, please follow the guidelines below when using Coupang Category Recommendation API.
1. Enter the product name in a way that clearly describes the product characteristics.
Bad example: “Round t-shirt gn 95 aden green”
Fashion apparel products like the example above must include gender (women/men/unisex) data in the product name. Also, instead of 'round t-shirt', a more detailed description including attributes such as 'sweatshirt', 'hoodie' or 'long/short sleeve' will be more helpful in finding the appropriate category.
Good example: “Round t-shirt men long sleeve sweatshirt gn 95 aden green”
2. Be careful not to enter differing keywords in a single product name. You may receive an inaccurate recommendation if you input a wrong keyword.
Bad example: “Pet carrier pet toy pet apparel leash”
If it is difficult for a person to determine product category based on the input product name, it is likely that the Coupang Category Recommendation API won't be able to produce a fitting recommendation.
Make sure you do not register product name in deal format. The search keyword and product name must be clearly distinguished from one another in the product name.
Good example: “Pet puppy auto-leash”
Path
POST
/v2/providers/openapi/apis/api/v1/categorization/predict
Example Endpoint
https://api-gateway.coupang.com/v2/providers/openapi/apis/api/v1/categorization/predict
Request Parameters
Body Parameter
| Name | Required | Type | Description | ||||
|---|---|---|---|---|---|---|---|
| productName | O | String |
Product name
|
||||
| productDescription | String |
Detailed product description
|
|||||
| brand | String |
Brand
|
|||||
| attributes | Array |
Product attribute information (e.g. size, color, material, etc.)
|
|||||
| sellerSkuCode | String |
Seller sku code (supplier sku code)
|
|||||
Request Example
{
"productName": "코데즈컴바인 양트임싱글코트",
"productDescription": "모니터 해상도, 밝기, 컴퓨터 사양 등에 따라 실물과 약간의 색상차이가 있을 수 있습니다. 캐주얼하지만 큐티한디자인이 돋보이는 싱글코트에요 약간박시한핏이라 여유있고 편하게 스타일링하기 좋은 캐주얼 싱글코트입니다. 컬러:베이지,네이비 사이즈:FREE 실측(측정자,측정기준,제작과정에따라 다소차이가있을수있습니다) 단면기준 CM 단위 가슴단면:61 어깨:54 팔기장:55(어깨절개선기준) 총장:88",
"brand": "코데즈컴바인",
"attributes": {
"제품 소재": "모달:53.8 폴리:43.2 레이온:2.4 면:0.6",
"색상": "베이지,네이비",
"제조국": "한국"
},
"sellerSkuCode": "123123"
}Response Message
| Name | Type | Description | ||||
|---|---|---|---|---|---|---|
| code | Number |
HTTP Status Code
|
||||
| message | String |
HTTP Status message
|
||||
| data | Object |
|
||||
| autoCategorizationPredictionResultType | String |
Result type (SUCCESS, FAILURE, INSUFFICIENT_INFORMATION)
|
||||
| comment | String |
|
||||
| predictedCategoryId | String |
Recommended category ID (displayCategoryCode)
|
||||
| predictedCategoryName | String |
Recommended category name
|
||||
Response Example
{
"code": 200,
"message": "OK",
"data": {
"autoCategorizationPredictionResultType": "SUCCESS",
"predictedCategoryId": "63950",
"predictedCategoryName": "일반 섬유유연제",
"comment": null
}
}
Error Spec
| HTTP Status Code (Error Type) | Error Message | Solution |
|---|---|---|
| 400 (check request variable) |
Input product name should not be empty! |
Please make sure you have filled in the product name(productName). |
| 500 (server error) |
Internal Server Error |
Please make sure you are not missing the productName parameter input. |
| 500 (server error) | Error occurred when communicating with seller_intelligence domain service. vendorId = A00123456, productName = ******** .. | Error generated in the process of predicting recommendation category. Communicate again later or modify input value. |
| 500 (server error) | Fail to communicate with the downstream domain services (catalog prediction service) | Communicate again later or modify input value. |
| 500 (server error) | Fail to communicate with the listing domain to retrieve category name | Communicate again later or modify input value. |
| 500 (server error) | Receive error message from the downstream domain services (catalog prediction service) | Communicate again later or modify input value. |
URL API Name
GET_PRODUCT_AUTO_CATEGORY