Data-Science Recommendations Release

Feature Release:

Recommendations leveraging machine learning algorithms are now enabled for specific soft categories. Please refer to user guides for Soft Categories here.

Best seller soft category is now driven by dynamic data based on historical sales (instead of being manually curated)

Recommendations have been enabled in the following onsite placements:

  • Best Sellers (soft category GID = 4)
  • Cart > Inspired by Your Cart (soft category GID = 6)

    • This placement is shown in the user’s cart
    • Inspired by Your Cart (also known as Cartbuilder) model looks at a given customer’s cart variants and other customer’s cart & order variants, and learns which variants are commonly bought together
  • Homepage > Top Picks For You (soft category GID = 5)

    • This placement is shown in the Homepage personalized module
    • Top Picks For You model looks at a user’s purchases, cart adds, or PDP views, and learns which variants the user is interested in and is able to create recommendations.
  • PDP > Recommended For You (soft category GID = 5)

    • This placement is shown in the “Recommended” tab of PDPs
    • Recommended for You model looks at a user’s purchases, cart adds, or PDP views, and learns which variants the user is interested in and is able to create recommendations (same model as Top Picks for you)
  • PDP > Customers Also Bought (soft category GID = 796)

    • This placement is shown in the “Recommended” tab of PDPs - user guide here
    • Customer Also Bought model looks at all previous orders and learns what items are commonly bought together.
  • PDP > You Might Also Consider (soft category GID = 797)

    • This placement is shown in the “Recommended” tab of PDPs
    • You Might Also Consider model looks at variant level features (i.e. brand name, faction, category) the model finds variant that are most similar to a given variant.