Solution Background and Business Value
Personalized try-something-new recommendations introduce customers to products they haven’t purchased before but are likely to enjoy based on their buying habits and preferences. This strategy enhances customer experience by encouraging product discovery while also expanding sales into new product categories. By integrating these recommendations into push notifications, in-product placements, and email campaigns, businesses can:- Drive cross-selling opportunities by exposing customers to new product lines.
- Enhance customer satisfaction by keeping the shopping experience fresh.
- Increase engagement and retention by offering relevant and timely suggestions.
Data Requirements Graph Schema
To develop an effective Try-Something-New recommendation model, we need a structured dataset. While a small set of core tables is sufficient to start, Kumo AI enables us to improve the model by incorporating additional sources of data. Core Tables The three essential tables required for this solution are:-
Users Table
- Stores details about users for whom we want to generate recommendations.
-
Key attributes:
-
user_id
: Unique identifier (Primary Key). -
join_timestamp
: When the user joined. - Other optional features: age, location, and shopping behavior.
-
-
Items Table
- Stores information about the items available for recommendation.
-
Key attributes:
-
item_id
: Unique identifier (Primary Key). -
start_timestamp
/end_timestamp
: Availability period of the item. - Other optional features: price, category, color, and brand.
-
-
Transactions Table
- Stores user purchase history, which the model learns from.
-
Key attributes:
-
transaction_id
: Unique identifier (Primary Key). -
user_id
: Foreign Key linking to Users. -
item_id
: Foreign Key linking to Items. -
timestamp
: When the purchase was made. - Other optional features: total amount, payment method, and discount applied.
-
Predictive Query
The challenge in try-something-new recommendations is balancing item similarity and user behavior modeling while ensuring recommendations remain novel to the user. Training a model solely on new purchases would discard valuable purchasing patterns. Instead of training on only first-time purchases, we train a general item-to-user recommendation model and apply filters at prediction time to remove items the user has already purchased. This ensures:- The model learns broad user-item affinity.
- Users receive only new product recommendations.
- Predicts the top 50 distinct items a user is likely to buy.
- Looks at a future X-day window to determine potential purchases.