The Marketing Campaign

The company has been developing a new product, and upon the public release of the product, the marketing team wants to run a successful campaign where it is able to convince potential customers to buy the product. To run a successful campaign, the marketing team must determine to how best target the potential customers such that they can efficiently manage their budget and the company's time. To help determine the best approach, the following questions need to be answered:

• Should the marketing campaign commence at a larger scale?
• Which is more effective, advertisements or public service announcements?
• On which days do advertisments and PSAs have the biggest impact on buying the product?
• At which time of the day do advertisements and PSAs have the biggest impact on buying the product?
• Is there an amount of advertisements or PSAs that is most effective at convincing people to buy the product?

With the above questions in mind, the marketing team saw that the best approach was to run an A/B testing experiment, a randomized experimentation process where two or more versions of the campaign type are shown to different segments of people at the same time. As such, the following two test groups were created for the purpose of the A/B testing:

1. Ads
2. PSAs

The two groups represent whether the person in the study was only shown standard advertisements (Ads group) or whether they were shown public service annoucnements (PSAs group). Once the subjects were placed into their respective groups, the company began showing the content in the assigned messaging.

Objective

Analyze existing dataset from Kaggle that contains the results of an A/B testing experiment. This website will contain visualizations that help users such as marketing managers to understand the effectiveness of advertisements and public service announcements (psas). The marketing team and business leaders will be able to determine the effectiveness of the campaign with respect to the test groups (ads and psas), the optimal number of ads to share with users, and times which the ads/psas are most effective and targeting potential customers.

The data is obtained via the Marketing A/B Testing dataset from Kaggle.

Overview

In the A/B testing, the marketing team recorded, for each participant, the total number of ads or psas viewed, the day in which the participant viewed the highest number of ads or psas, the hour in which the participant viewed the highest number of ads or psas, and whether the participant ended up buying the product. In the graphs below, we get a brief overview into the total amount of messaging viewed by each group (ads or psas) as well as the conversion rate for each group. Here, conversion rate is defined as the percentage, within each group, of participants who ended up buying the product. It does not account for the amount of the product that was bought.

As shown in the graphs above, a much higher number of ads (approximately 14 million) were shown than psas (approximately 580,000). While there is a large discrepancy bewteen those two groups, there were still a high number of psas which were shown, allowing for the necessary analysis of potential public response to both ads and psas. Diving into the graph titled "Conversion % by Test Group", it is clear that there was a more positive response to the ads (2.55% conversion rate) compared to the psas (1.79% conversion rate). While this is less than a one percent difference, the amount of each that was shown suggests that the difference would hold if both ads and psas were shown to the public upon the release of the product. Furthermore, the marketing team can run a cost analysis (psas are free to show) of developing and showing the ads and psas to determine, based on the percentages indicated here, which of ads or psas (or both) should be shown.