Analytics Vidhya JOB-A-THON - August 2022

Predict CTR of an Email Campaign

Can you predict the Click Through Rate (CTR) of an email campaign?

Problem Statement

Most organizations today rely on email campaigns for effective communication with users. Email communication is one of the popular ways to pitch products to users and build trustworthy relationships with them.

Email campaigns contain different types of CTA (Call To Action). The ultimate goal of email campaigns is to maximize the Click Through Rate (CTR).

CTR is a measure of success for email campaigns. The higher the click rate, the better your email marketing campaign is. CTR is calculated by the no. of users who clicked on at least one of the CTA divided by the total no. of users the email was delivered to.

CTR = No. of users who clicked on at least one of the CTA / No. of emails delivered CTR depends on multiple factors like design, content, personalization, etc.

How do you design the email content effectively? What should your subject line look like? What should be the length of the email? Do you need images in your email template?

As a part of the Data Science team, in this hackathon, you will build a smart system to predict the CTR for email campaigns and therefore identify the critical factors that will help the marketing team to maximize the CTR.

Objective

Your task at hand is to build a machine learning-based approach to predict the CTR of an email campaign.

About the Dataset

You are provided with the information of past email campaigns containing the email attributes like subject and body length, no. of CTA, date and time of an email, type of the audience, whether its a personalized email or not, etc and the target variable indicating the CTR of the email campaign.

Data Dictionary

You are provided with 3 files - train.csv, test.csv and sample_submission.csv

Train and Test Set

Train and Test set contains different sets of email campaigns containing information about the email campaign. Train set includes the target variable click_rate and you need to predict the click_rate of an email campaign in the test set.

Variable	Description
`campaign_id`	Unique identifier of a campaign
`sender`	Sender of an e-mail
`subject_len`	No. of characters in a subject
`body_len`	No. of characters in an email body
`mean_paragraph_len`	Average no. of characters in paragraph of an email
`day_of_week`	Day on which email is sent
`is_weekend`	Boolean flag indicating if an email is sent on weekend or not
`times_of_day`	Times of day when email is sent: Morning, Noon, Evening
`category`	Category of the product an email is related to
`product`	Type of the product an email is related to
`no_of_CTA`	No. of Call To Actions in an email
`mean_CTA_len`	Average no. of characters in a CTA
`is_image`	No. of images in an email
`is_personalised`	Boolean flag indicating if an email is personalized to the user or not
`is_quote`	No. of quotes in an email
`is_timer`	Boolean flag indicating if an email contains a timer or not
`is_emoticons`	No. of emoticons in an email
`is_discount`	Boolean flag indicating if an email contains a discount or not
`is_price`	Boolean flag indicating if an email contains price or not
`is_urgency`	Boolean flag indicating if an email contains urgency or not
`target_audience`	Cluster label of the target audience
`click_rate (Target Variable)`	Click rate of an email campaign

Submission File Format

sample_submission.csv contains 2 variables - campaign id and click_rate

Variable	Description
`campaign_id`	Unique identifier of a campaign
`click_rate (Target Variable)`	Click rate of an email campaign

Evaluation metric

The evaluation metric for this hackathon would be r2_score.