User Based

The User-Based algorithm recommends items by finding users similar to the current user. It generates recommendations based on user behavior and product popularity.

In this algorithm, product recommendations are based on the behavior of similar users (users with similar index scores who have viewed, purchased, or added the same or similar category products to their cart) to the current user. The algorithm recommends products that similar users have encountered in the past but that the current user has not visited. The user-based algorithm also takes the user-product-rating matrix as another input. For each product a user visits, the algorithm calculates a rating based on the number of visits, purchases, and add-to-carts within the last 30 days. This algorithm can be used on any page.

Set coherent=true to build each user's recommendations from their strongest interest: the category or brand that dominates their recommended items. Products outside this interest are removed and replaced with top sellers from the same category or brand, so the response stays focused on one interest. Result quality depends on how complete your catalog attributes are, such as category and brand.

Endpoint

GET https://recommendation.api.useinsider.com/v2/user-based

Visit our Postman collection to test this request.

Query Parameters

Parameter	Sample Value	Description	Data Type	Required
partnerName	mybrand	Partner Identifier assigned by Insider One. You can use PartnerID as well.	String	Yes
locale	us_US	Locale of requested product catalog	String	Yes
platform	web	Requested platform. Web comes by default.	Enum	No
userId	a1b2c3d4	User identifier that is assigned by Insider One	String	Yes
currency	USD	Requested currency of the products. If no value is set, the default currency in your settings is used.	String	No
size	50	Required number of items in response. Valid values are 0 to 100.	Integer	No
categoryList	[“Clothes”, “Skirts”]	Category filter of the products	Array (of string)	No
filter		Smart Recommender filtering. There can be more than one filter parameter.	String	No
details	true	Adds details to the products of the response	Boolean	No
shuffle	false	Shuffles the products of the response	Boolean	No
coherent	true	Builds the recommendations from the user's strongest interest (the dominant category or brand) and replaces products outside it with top sellers from the same category or brand. Accepts true or false; the default is false.	Boolean	No
getGroupProducts	false	Shows variant products under the products of the response	Boolean	No
groupProductsFields		Adds these fields to the variant products’ details	String	No
excludeVariants	true	Exclude variants from the response	Boolean	No
excludeViewDay	30	After how many days should viewed products be excluded	Integer	No (Can be used only with userId)
excludeViewItem	100	How many viewed products should be excluded	Integer	No (Can be used only with userId)
excludePurchaseDay	30	After how many days should purchased products be excluded	Integer	No (Can be used only with userId)
excludePurchaseItem	100	How many purchased products should be excluded	Integer	No (Can be used only with userId)
hp	false	Makes affinities affect products of the response	Boolean	No
dayLimit	2	If FMT is published_time, it adds a day limit filter	Integer	No
productId	ABC123CBA	Current product ID	String	No

Sample Request

The sample below displays a request to User Based, a personalized recommendation algorithm that finds users similar to the current user.

https://recommendation.api.useinsider.com/v2/user-based?partnerName={Partner_Name}&locale={Locale}&currency=TRY&userId={User ID}

The sample below requests closely related items for the same user:

https://recommendation.api.useinsider.com/v2/user-based?partnerName={Partner_Name}&locale={Locale}&currency=TRY&userId={User ID}&coherent=true

Sample Response

{
        "success": true,
        "total": 10,
        "types": {
        "mvop": 10
        },
        "data": [
        "QAZ-7890",
        "XYZ-1234",
        "QAZ-7899",
        "XYZ-1233",
        "QAZ-7898",
        "XYZ-1243",
        "QAZ-7891",
        "XYZ-1223",
        "QAZ-7892",
        "XYZ-1342"
        ]
        }

Fallback Algorithms

If the products from User Based are not enough to fill the response data, some fallback algorithms below fill it:

View-to-view of the last visited product
Most viewed of the category
Most viewed of the category without excluding the right-most item in the categoryList
Most viewed of the Partner

When you request recommendations through a saved recommendation strategy, the strategy's Recommend Closely Related Items setting applies. The coherent parameter is for direct User-Based requests.