Recommend products and categories - Bloomreach Experience - Open Source CMS

Recommend products and categories

Deprecated
This API is going to be deprecated by the end of 2020. New customers will be a part of the Pathways and Recommendations platform. To know more about Pathways and Recommendations, see this page. Please contact your Bloomreach representative for more information.

Example request

Here's an example REST API call for a JFY query:

Request

GET http://staging-core.dxpapi.com/api/v1/core/?
account_id=<Bloomreach Provided Account ID>
&auth_key=jazzhands
&domain_key=example_com
&request_id=2779692816548
&_br_uid_2=uid%3D9683404349848%3A_uid%3D1060194806196%3Av%3D11.5%3Ats%3D1366069652967%3Ahc%3D109
&url=http://www.example.com/index.html?JFY
&ref_url=http://www.example.com/home
&request_type=jfy
&content_type=jfy,dynamic_categories
&fl=thumb_image,url,title,brand,pid,sale_price,sale_price_type,price

Example response

Response snippet

{
 "dynamic_cats_response": {
     "title":"dynamic_cats"
      "docs": [ // list of suggestions
            { // suggestion 1
             "docs": [ // list of products in suggestion 1
                 {
                  title: "example product",
                  price: 34.9,
                  url: "http://www.example.com/91000048803",
                  pid: "91000048803",
                  thumb_image:"http://example/6838/1000048803-930-MAIN-170x269.jpg",
                  sale_price: 15,
                  sale_price_type: "",
                  brand: "Almost Famous"
               },
               {
                  // second product in suggestion 1
               },],
             "master_cat": "name of master category",
             "title": product by almost famous
            },
            {
             //suggestion 2
             .....
            }
       ] // end of list of all suggestions
    "personalized result": "true"
  },
  
 "response": { // jfy response (Recently browsed across devices)
       "docs": [
           {
               "thumb_image": "http://www.example.com/productpics/thumb/screen.jpg",
               "title": "Samsung Galaxy Anti-Glare Screen Protector Film",
               "url": "http://www.example.com/p-166880",
               "brand": "Samsung",
               "pid": "166880",
               "sale_price": 7.99,
               "sale_price_type": "",
               "price": 7.99
           },
           {
              // product 2
           },
           {
              // Product 3
           }
       ],
       "title": "jfy"
   }
  "personalized result": "true"
}

Suggest categories

The Dynamic Categories service can suggest up to five Dynamic Categories. Each of these categories contains as many as 30 products with ranking optimized by Commerce Search and Categories based on various signals such as product generated revenue, conversions, views, and others.

Dynamic Categories are derived by examining the user's browsing history for product categories, brands, and on-sale statuses (boolean). The tuples (category, brand, sale) with the highest number of user activities become the user's personalized Dynamic Categories. These categories are tuned to user activities. The recommended categories and products are sensitive to changes in the category tree, product availability, price, and product performance. As the user browses more products, the recommendations become richer and more relevant.

Dynamic Categories are stored in dynamic_cats_response.docs.

Response snippet

"dynamic_cats_response": {
     "title":"dynamic_cats"
      "docs": [ // list of suggestions
            { // suggestion 1
             "docs": [ // list of products in suggestion 1
                 {
                  title: "example product",
                  price: 34.9,
                  url: "http://www.example.com/91000048803",
                  pid: "91000048803",
                  thumb_image:"http://example/6838/1000048803-930-MAIN-170x269.jpg",
                  sale_price: 15,
                  sale_price_type: "",
                  brand: "Almost Famous"
               },

You can specify excluded categories from Dynamic Categories. For example, you might decide to exclude sexual wellness categories from Dynamic Categories. Putting these categories on a blacklist prevents them from being automatically suggested when a user browses for other drugstore items. To blacklist items from Product Recommendations, you need to contact your Bloomreach Technical Manager.

Suggest products

JFY begins generating personalized content after a user does three qualified activities, such as opening a product page. The Product Recommendations service can suggest as many as 20 products. These suggestions are based on the user's browsing history. The service compares the user's history with that of other users, identifying products with similar browsing patterns from similar users. As the browses more products, the recommendations are automatically updated. The more the user browses, the more refined the recommendations become.

The example response uses JSON output with JFY plus dynamic_cats_reponse. The user browsed at least three products, which is the threshold for showing Product Recommendations.

Product Recommendations are stored in response.docs.

Response snippet

"response": { // jfy response (Recently browsed across devices)
       "docs": [
           {
               "thumb_image": "http://www.example.com/productpics/thumb/screen.jpg",
               "title": "Samsung Galaxy Anti-Glare Screen Protector Film",
               "url": "http://www.example.com/p-166880",
               "brand": "Samsung",
               "pid": "166880",
               "sale_price": 7.99,
               "sale_price_type": "",
               "price": 7.99
           },

What do JFY widgets display for new site visitors?

JFY widgets personalize suggestions and recommendations based on individual site visitor activity. If Bloomreach doesn't have enough information about a site visitor to generate personalized suggestions and recommendation, then your JFY widgets display bestseller products. For information about deploying the bestseller API on your site, see Bestseller API. 

How do I check the quality of my JFY recommendations?

Here's how you can QA your JFY widgets:

  1. Start by opening your site.
    For best results, open your site in an incognito window. This way, you can be sure that your results are due to your activities during your test rather than prior activities.
  2. Open a page that displays a JFY widget. Let's call this page your test page.
    Observe the products in the widgets. You're going to compare these products with the products displayed at the end of the test.
  3. Run a few searches and open at least five product pages per search.
    This step simulates site visitor behavior so that Bloomreach can personalize the quality of results to individual visitors on your site.
  4. Return to your test page. That's the page where you first noted the product recommendations in the JFY widget.
    Observe the set of products in the JFY widget. You're checking to see that the product recommendations are changed according to the search behavior you simulated during this test.

What do I do if I don't like the results of my JFY test?

First, make sure that you simulated enough site visitor behavior. You need to run a few searches, not just one. For each set of search results, you need to open at least five product pages. 

If your JFY widgets are still not displaying recommendations that are influenced by your activities, then contact your Bloomreach representative for help.

Did you find this page helpful?
How could this documentation serve you better?
On this page
    Did you find this page helpful?
    How could this documentation serve you better?