Marketing companies often use A/B testing to evaluate the effectiveness of different campaign strategies. This involves showing different versions of a variable (e.g., a web page or banner) to different segments of an audience simultaneously to determine which version has the greatest impact.
In our A/B test, most participants see ads (experimental group), while a smaller segment sees a Public Service Announcement (PSA) or nothing (control group). Our analysis aims to answer two key questions:
- Will the campaign be successful?
- How much of the success can be attributed to the ads?
Using this dataset, we will analyze group performance, assess ad effectiveness, estimate potential revenue, and determine if the differences between the groups are statistically significant. In this guide, we will explore how to perform exploratory data analysis (EDA) and hypothesis testing on A/B test data using Python.
Code Link: Kaggle Notebook, Original data Source: Kaggle Marketing A/B Testing Data