Last updated: Feb 06 2019
Catalog feed specification
This section provides an overview of requirements for the product catalog feed used when implementing product recommendations for Episerver personalization solutions. The feed is critical to the success of the Episerver personalization technology, and the quality of the feed affects the ease of integration, and recommendation quality.
Note: If your website supports more than one location, see the Episerver Catalog Feed Location document for details on how to handle this within the feed.
How it works
The product catalog feed combined with the tracking script, enables Episerver to track visitor interaction and serve suitable recommendations. The feed is a file including all the products in the product catalog. For each product, there is an image link, title, identifier, URL, price, currency, and optional product attributes.
The quality of the product recommendations depends on the feed. For example, if the feed is not set up correctly, the wrong images, prices, or out-of-stock products may appear in the recommendations.
The following image shows how the feed is used within the Episerver personalization technology.
The product feed
Exporting the product catalog
The product catalog must be regularly exported from Episerver Commerce, to be picked up by the product feed in Episerver Perform. This is done using an Episerver scheduled job, see Exporting your catalog to Episerver Perform.
Importing the product catalog
The product catalog is regularly imported through the product feed, combined with visitor track events, and processed by machine learning algorithms customized by the customer's merchandising rules, to produce product recommendations.
At the initial setup, a review of the feed is completed by Episerver, and you can request additional content and/or attributes. An understanding of the required strategy determines what to include in the feed; the strategy will be agreed by both parties before the feed is imported.
Note: Occasionally, after some user behavior is collected and the continuous optimization and improvements process begins, additional attributes may be requested. Specific attributes may present themselves as being necessary, meaning further revisions to the feed may be required.