Predictive 'Next Best Action' Recommendation Engine

You are a Personalization Strategy Lead.

Context

We want to implement a 'Next Best Action' (NBA) module in our post-purchase emails. For customers who buy the {{anchor_product}}, we want to suggest items from the {{complementary_category}} using {{data_points}} to refine the selection.

Task

1. Define the 'Affinity Logic': Why does a buyer of {{anchor_product}} need {{complementary_category}}?
2. Create three 'Personalization Rules' using {{data_points}} (e.g., 'If size=M and weather=cold, recommend X').
3. Write the 'Reason Why' copy for the recommendation (e.g., 'Because you bought X, you might love Y').
4. Design a 'Bundle' incentive that triggers if they add the recommended item within 24 hours of their first purchase.
5. Outline the data schema needed to track the success of these recommendations.

Constraints

• MUST NOT suggest products the customer has already purchased.
• MUST NOT recommend items with less than a 4-star rating.
• The 'Reason Why' copy MUST be dynamic, not static.

Output format

• Logic Architecture (If/Then Rules)
• Dynamic Copy Templates
• Data Requirements Table
• Performance KPIs

Quality bar

• Is the logic for {{complementary_category}} sound?
• Does it effectively utilize all {{data_points}}?
• Is the UX flow seamless for the customer?