Define Product Recommendation Model

To be able to generate Product Recommendations, you need to define a Product Recommendation Model which handles the data mapping and input data scope.

Steps:

  1. Go to Optimization > Models and create a new model.

  2. Provide Name and Label for the new model.

  3. As Model Class select Product Recommendation 2.

  4. Using the User Group (Edit / View Details) fields define which user groups can edit or view the model.

  5. After you click Create, the Definition step opens.

Definition Step

This step defines mapping, and scope of input data.

  1. Select a Transaction Source – Data Source or Datamart with historical transactions to use for generating product recommendations.

  2. If needed, define a Transaction Filter to scope down transactions to use for generating product recommendations

  3. Define the following options:

    • Product Name – Select a field which contains product name.

    • Product ID – Select a field which contains a unique code for each product (should be the same code utilized in the Product Master table).

    • Transaction identifier – Value used for grouping multiple items purchased under one ID, such as basket ID. If not available, transaction date can serve this purpose.

    • Revenue Measure – Used to get the unit price and unit margin (often used to calculate the recommendation score).

    • Margin Measure – Select a field which contains profitability metric to be used for calculating recommendation score.

    • Quantity Measure – Used to get the unit price and unit margin (often used to calculate the recommendation score).

  4. Click Apply Settings. You will get a list of transactions both within and outside the specified scope.

  5. After you click Continue, you will proceed to the Configuration step.

Screenshot 2024-01-12 at 11.54.53.png

Â