You can use the Recommendation Feeds dashboard to upload your custom recommendations. Following types of feeds are supported. 

  1. User to Product recommendations : Recommendations for each user can be imported using one of the user identifiers and one or more product_ids from the catalog. Popular examples include personalized promotions.
  2. Product to Product recommendations : Recommendations that go from one product to others. Popular examples include people who view or buy this product consider these other products next.

The import process for user to product recommendations is illustrated below and same applies to all types of imports.


1. Recommendations file (csv)

Prepare a csv with a user identifier (email or customer_id) and a list of recommended item ids.

Data Format

  • customer identifier (email or customer_id)
  • Item Id list (separated by a delimiter)

Sample Data for User to Product recommendations: 

email product_ids ITEM_0000 | ITEM_1111 ITEM_222 | ITEM_333 | ITEM_5555

Sample Data for Product to Product recommendations:

source_product_id product_ids
ITEM_123 ITEM_234 | ITEM_345
ITEM_234 ITEM_123 | ITEM_345 | ITEM_456

2. Create Recommendation

You can create a new recommendation content feed by clicking on the "+ Content Feed" button on the Recommendation Feed Index page


Enter a unique name for the recommendation feed


3. Import Recommendation File

Click on the import button to import new data


Map columns from .csv to standard Blueshift columns (email, customer_id and product_ids). Enter a delimiter for your item's list and finally click on "Import Recommendations"


You will receive an email notification when the upload process completes including error report. To use the imported recommendations you can go to the "predictive content" tab in the Recommendations editor.


Recurring Uploads

You can setup recurring Recommendation Content Feed upload jobs through our dashboard.

On the import page, please choose the Upload method as "S3 Upload".  The default adapter is "Blueshift Import". You can also choose "Custom Import" and provide your own s3 credentials and s3 file path. With default adapter you can drop any files in to the folder bsft-customers/<site-name>/import/recommendations/feeds/

  • Text
bucket: bsft-customers 

Please find a screen shot below with sample configuration. You can click on "Fetch Sample S3 File" to get some sample data and add attribute mappings and click on "Import Recommendations".


Import errors and possible causes

It's possible that you run into errors at the time of importing recommendations. Here are a few errors that you may run into and what could be the reasons behind those:

Firefly resource not found error=“{\“errors\“:[\“attribute_value_not_found\“]}”

This error occurs if an item in the product_ids list is not present in the catalog. 

Firefly resource not found error=“{\“errors\“:[\“product_id_not_found\“]}”

This error occurs when the source_product_id itself is not the catalog

Was this article helpful?
1 out of 1 found this helpful



Please sign in to leave a comment.