Integration

Integration prerequisites and timeline.

Prerequisites

The following conditions must be met before onboarding can begin:

• Standard Integration
  The website must already be integrated into Fanplayr's standard page and order tracking flows either by installing one our existing shopping cart extensions or via a manual integration.
• Enhanced User Identification
  The website must also implement the Enhanced User Identification feature to ensure returning visitors are identified consistently so that their product browsing and purchase signals can be used to accurately feed prediction models.
• Capture Order Products
  The website must include full a complete list of products purchased in all order events captured. Products in the list must include the id, sku, name, price and qty fields.
• Consistent Product IDs and SKUs
  Ensure that product IDs and SKUs are tracked consistently between all page view, shopping cart and order tracking events in a way that is compatible with your desired product catalog and recommendations configuration. See, catalog levels for more information.
• Minimum data requirements
  For most recommendation models you will need at least 90 days of page view and order events to have been captured before the models can be trained. Having fewer events will limit the options available and affect the quality of the models generated. See, model requirements for more information.

Integration timeline

The time to go live with Fanplayr Product Recommendations is typically between 2-3 weeks depending on the size of your product catalog and the amount of traffic generated by your store. A typical onboarding timeline is as follows:

Step	Average timeframe
Project configuration Creation of Google AI project and connections to the Fanplayr platform.	2 days
Import initial data Backfilling at least 90 days of user events and importing initial product catalog.	5 days
Validate imported data Validation of initial data imported.	2 days
Configure event streaming Setup streaming of new user events and catalog updates.	3 days
Train initial AI models The initial training of each AI model is estimated to take 7-14 days depending on the size of the product catalog and the amount of traffic generated by your store. Models can be trained in parallel.	7-14 days
Validation of AI models General testing of trained models.	3 days

Post-integration you will be able to create widgets and set up segments to add product recommendations to your site.