Optimizing Product Recommendations A/B Testing

A/B Testing using Python to determine the most effective marketing campaigns.

Case Study

In a recent e-commerce A/B test, we investigated the impact of product recommendations on purchase behaviour. By comparing conversion rates between users receiving phone and screen guard recommendations (control group) and those receiving screen guard and case cover recommendations (test group), we uncovered a significant difference, highlighting the influential role of case covers in driving purchases.

Dataset Description

Our data set consists of 868 observations which include:

• Customer_ID: Unique identifier for each customer. A sequential number identifies each customer.
• Recommendation_name: Name of the product recommended to the customer.
• Recommendation_date: Date when the recommendation was made to the customer.
• Suggestion_type: Indicates whether the recommendation was made with a phone or a cover.
• Purchase_flag: Binary variable indicating whether the customer made a purchase (1) or not (0) in response to the recommendation.

Libraries

• Pandas
• Matplotlib
• Scipy