Version 2 catalogs (latest)

Details recommendation models and requirements for version 2 catalogs.

Google AI models types

The training data requirements listed for each of the Google AI models is a guideline for the typical data needed before training can begin. These data requirements may very depending on the training objectives and quality of the data available.

Others You May Like (OYML)

The "Others you may like" recommendation predicts the next product that a user is most likely to engage or convert with. The prediction is based on the viewing history of the user and the candidate product's relevance to a current specified product.

Supported pages

• Product page. See Product Page View event.

• Add-to-cart page. See Add To Cart event.

• Shopping cart page. See Shopping Cart Page View event.

Training data requirements

• At least 10,000 Product Page View events in the last 90 days.

• AND at least 100 unique catalog items for Product Page View events in the last 90 days.

• AND one of the following:

  • At least 7 days of Product Page View events in the last 90 days AND at least 10 occurrences per catalog item on average (with a 90 days window) of Product Page View events

  • OR at least 60 days of Product Page View events in the last 90 days.

Frequently Bought Together (FBT)

The "Frequently bought together" recommendation predicts items frequently bought together for a specific product within the same shopping session. If a list of products is being viewed, then it predicts items frequently bought with that product list.

This recommendation is useful when the user has indicated an intent to purchase a particular product (or list of products) already, and you are looking to recommend complements (as opposed to substitutes). This recommendation is commonly displayed on the "add to cart" page, or on the "shopping cart" or "registry" pages (for shopping cart expansion).

Supported pages

• Product page. See Product Page View event.

• Add-to-cart page. See Add To Cart event.

• Shopping cart page. See Shopping Cart Page View event.

• Order confirmation page. See Purchase Complete event.

Training data requirements

• At least 10 occurrences per catalog item on average (with a 1 year window of Purchase events) OR at least 90 days of Purchase events in the last 1 year.

• AND at least 100 unique catalog items for the Purchase event in the last 1 year.

• AND at least 1,000 Purchase events in the last 1 year.

Recommended For You (RFY)

The "Recommended for you" recommendation predicts the next product that a user is most likely to engage with or purchase, based on the shopping or viewing history of that user. This recommendation is typically used on the home page.

Supported pages

• All pages that generate an AI model event.

Training data requirements

• At least 10,000 Product Page View events in the last 90 days.

• AND at least 100 unique catalog items for the Product Page View event in the last 90 days.

• AND one of the following:

  • At least 7 days of Product Page View events in the last 90 days AND at least 10 occurrences per catalog item on average (with a 90 days window) of Product Page View events.

  • OR at least 60 days of Product Page View events in the last 90 days.

• AND at least 7 days of Home Page View events in the last 90 days.

• AND at least 10,000 Home Page View events in the last 90 days.

Recently Viewed (RV)

Returns up to 20 items recently viewed by the user, with most recently viewed items listed first.

The "Recently viewed" recommendation is not actually a recommendation. It provides a list of products the user has recently interacted with, with the most recent items first.

Supported pages

• All pages that generate an AI model event.

Training data requirements

• At least one Product page view event must have been captured for the user.

Google AI model events

AI models require specific user event details to provide context for queries to the prediction models. This section documents the different events supported by Fanplayr. In general, only the following Fanplayr page types (via the pageType property) produce AI model events:

• Home Page

• Product Page

• Category Page

• Search Page

• Shopping Cart Page

• Order Confirmation Page (tracked via the Order Confirmation snippet)

The following are a list of AI model events with details of when they are generated by Fanplayr:

Home Page View

This event occurs when the user views the homepage.

Requirements

• pageType property must be "home".

Product Page View

User views a product detail page.

Requirements

• pageType property must be "prod".
• productId property must be set to a valid product in the catalog.

Category Page View

User views a page which contains multiple product listings such as a product category list or collection.

Requirements

• pageType property must be "cat".
• categoryId and/or categoryName properties must be set.

Shopping Cart Page View

This event occurs when the user views the shopping cart.

Requirements

• pageType property must be "cart".

Search Page View

This event occurs when the user views a product search results page.

Requirements

• pageType property must be "srch".
• At least one of the following must be met:

  • The page includes category details via the categoryId and/or categoryName properties.
  • The page includes search query text via the searchQuery property.

Add To Cart

This event occurs when Fanplayr can detect a new product being added to the cart or an increase in product quantity.

Requirements

• User must have added a new product to their shopping cart or increased the quantity of a product.

Purchase Complete

This event occurs when the user completes the checkout process. This typically occurs on a store's order confirmation page.

Requirements

• An order must be tracked via Fanplayr's order tracking snippet.

• The tracked order details must include an order identifier, revenue details and at least one product.

• The "Track page view after order" campaign option must be enabled to allow segmentation and widgets to re-evaluate after an order is placed.

Ranking models types

The following models return a list of top ranked products by a particular metric based on up to the past 30 days of events.

There are four product ranking models to choose from:

Ranked by Views (RBV)

Returns the top ranked products by number of times viewed.

Ranked by Carts (RBC)

Returns the top ranked products by number of times added to cart.

Ranked by Purchases (RBP)

Returns the top ranked products by number of times purchased.

Ranked by Revenue (RBR)

Returns the top ranked products by total revenue.

Ranking model filters

The following optional filters can be applied to limit the products queried from each of the Ranking models above:

• Number of days The number of days of history to use when querying for the top ranked products, starting from today. Valid values are between 1 and 30, where "1" will query up to the past 24 hours of data and "2" will query up to the past 48 hours of data etc.

• Start price Limits the results to only products whose price is greater than or equal to the start price.

• End price Limits the results to only products whose price is less than or equal to the end price.

• Category Limits the results to only products of the specified category name. This can also be configured to use the current page category.

Other Fanplayr models

Recently Viewed

Returns up to 20 items recently viewed by the user in the last 30 days, with most recently viewed items listed first.

The "Recently viewed" recommendation is not actually a recommendation. It provides a list of products the user has recently interacted with, with the most recent items first.

Supported pages

• All pages.

Data requirements

• The user must have viewed at least one product in the last 30 days.

Options

Option	Description
Minimum Products	The minimum number of products needed by the display. The display will only execute if the model returns enough products. This value can be between 5 and 1440 (1 day).
Maximum Products	The maximum number of products to be queried.

Product Activity (Views)

This model is not actually a recommendation but rather returns the number of times the current product has been recently viewed by users on the website.

Supported pages

• Product pages.

Options

Option	Description
Recency Time (minutes)	This value is the window of time to query product activity in minutes. For example, a 60 minute window will calculate the number of times the current product has been viewed in the last hour. This value can be between 5 and 1440 (1 day).
Minimum Activity Threshold	This value is the minimum number of recent views needed to show the widget using this product display. Set to 0 to always show the widget.

Product Activity (Carts)

This model is not actually a recommendation but rather returns the number of times the current product has been recently added to the shopping cart by users on the website.

Supported pages

• Product pages.

Options

Option	Description
Recency Time (minutes)	This value is the window of time to query product activity in minutes. For example, a 60 minute window will calculate the number of times the current product has been added to cart in the last hour. This value can be between 5 and 1440 (1 day).
Minimum Activity Threshold	This value is the minimum number of recent add to cart events needed to show the widget using this product display. Set to 0 to always show the widget.

Product Activity (Purchases)

This model is not actually a recommendation but rather returns the number of times the current product has been recently purchased by users on the website.

Supported pages

• Product pages.

Options

Option	Description
Recency Time (minutes)	This value is the window of time to query product activity in minutes. For example, a 60 minute window will calculate the number of times the current product has been purchased in the last hour.
Minimum Activity Threshold	This value is the minimum number of recent purchases needed to show the widget using this product display. Set to 0 to always show the widget.