Analyze A/B Test Results

Introduction

For this project, A/B test results of an e-commerce company were analyzed. The data was provided by Udacity

In the data provided, the company has developed a new web page in order to try and increase the number of users who "convert," meaning the number of users who decide to pay for the company's product. The objective of the project is to help the company understand if they should implement this new page, keep the old page, or perhaps run the experiment longer to make their decision.

Analysis

The dataset was analyzed using in three parts

Part I - Probability

Part II - A/B Test

Part III - Regression

Statistical analysis was conducted to test if there is a significant effect on the conversion rate. Link to analysis

Conclusion

The results of A/B test are not significant to conclude if one page is better than the other. We fail to reject the Null Hypothesis.

There is no significant difference in the mean conversion rate between the control group and treatment group. Hence, the e-commerce website would not benefit by changing to the new page and should continue using the old page.

Recommendation

The e-commerce company needs to maintain the old page as there's no evidence that the new page can perform better.