Configure Catalog and Order Feeds for B2C Commerce Einstein Deployment

Configure, schedule, and deploy catalog and order data feeds to Einstein.

Before you can use Einstein features, you must first configure catalog and order feeds for B2C Commerce Einstein. Data about products in the product catalog and about shopper orders is transmitted using these feeds. The feeds export important properties about your product catalog and orders to B2C Commerce Einstein.

Einstein employs this data to train the machine learning models used to generate results for the B2C Commerce Einstein features that you choose to activate on your storefront. The data is also fed into the Einstein Configurator tool. Use Einstein Configurator to configure business rules that fine-tune how the system generates recommendations.

Cofiguring these feeds controls how often catalog and order data is exported.

Note Configure catalog and order feeds only on production instances.

To configure catalog and order feeds:

1. In Business Manager, click App Launcher , and then select Administration | Operations | Einstein Status Dashboard.
2. Using the dropdown menu to the right of each site, select Configure.
   

   Note Unless it’s blank, leave the Host field alone and skip the next step.
3. If the Host field is blank, specify the Host URL for the current site.
   For example, www.myCompanyName.com.

   

   Note When specifying the host URL, don’t include 'https://'.
4. In the Deployment Configuration section, select Region.
   Select a region only one time per site. This setting specifies a region corresponding to your primary business location. The system uses this information to determine where to physically store and process predictive data. Make sure region setting is configured in Production and Staging instance.
5. Use the following table to define specific actions:

   Action	Select …
   Generate product recommendations on product details pages for pages that are currently out of stock.	Out-of-Stock Products Note This option enables Einstein to store out-of-stock product catalog information, to continue recognizing out of stock product IDs viewed by shoppers, and to honor any applied recommender rules.
   Storefront catalogs that leverage product type: Variation Group.	Variation Group Products
   Product catalogs that contain data for products beyond the default locale.	Multi-Locale

6. To export orders created after a specific date, specify a value in the Export Orders Created After field.
   

   Note When specifying this value, provide either the date your site went live on B2C Commerce or two years, whichever is the smaller value.
7. Define the maximum number of orders per run.
   This value is the anticipated number of orders you expect on a peak day. The default value is 10,000. As a reference, for a setting of 100,000 orders, the order feed processing time is approximately 30 minutes.

   

   Note When configuring maximum orders per run for new sites with large order histories, we recommend that you increase the number of orders to exceed 10,000. This setting enables the order feed to import orders from the most recent date instead of taking many days to catch up to the current day. If the order feed isn’t up to date, it's possible that Commerce Insights won't generate. Commerce Insights displays shopper basket visualization for the last 7 days.
8. (Optional) To run the catalog and order feeds immediately, click Save.
9. Set the Schedule Catalog and Order Feed switch to On.
   If the switch is inactive, catalog and order feeds aren’t scheduled for the site.
10. Define the catalog feed schedule.

    Field	Recommendations
    Start Date and Time	To make sure that Einstein receives your most current product information or recent product launches, schedule jobs to run shortly after replication from staging to production completes. To view staging to production replication job completion time go to Staging Instance | App Launcher | Administration | Replication | Data Replication Processes | view End Time for Type: Transfer & Publication.
    Time Unit	Day
    Frequency	1 Note We recommend a daily catalog feed to enable Einstein to remain up to date with any changes. Einstein exports the entire catalog only on the first run. To keep daily runtimes shorter, subsequent runs contain only deltas.

11. Define the order feed schedule.

    Field	Recommendations
    Start Date and Time	Schedule jobs for off-peak ordering times.
    Time Unit	Day
    Frequency	1 Note Because it uses orders from the last 7 days, configuring a daily order feed is important for keeping the Commerce Insights feature up to date. Order feeds are also used in the Product Affinity Algorithm for the Product Recommendations strategy.

12. Click Save.
    Business Manager saves the configuration.
13. Wait at least 24–48 hours for deployment to complete.
14. Log a ticket with Salesforce Support and request to be add an initial Site Admin to the Configurator.
15. After you are added as the Site Admin, open the Einstein Configurator and navigate to Admin | User Management. Add additional users as needed.
    Users must have an Account manager account in the relevant organization, but don't need a specific Account Manager role.