Cafe Rewards Promo Analysis

About the dataset.

Data that simulates the behavior of Cafe Rewards members over a 30-day period, including their transactions and responses to promotional offers. The data is contained in three files: one with details on each offer, another with demographic information on each customer, and a third with the activity for each customer during the period. The activities are divided into offer received, offer viewed, offer accepted, and transaction. For a transaction to be attributed to an offer, it must occur at the same time as when the offer was "completed" by the customer.

Project Objective.

Identify key Cafe Rewards customer segments and define offers that should used to target them effectively.

My Approach.

Data Cleaning and normalization.

Blank values were removed from the gender and income columns of the Customers table. They were all customers aged 118 which I considered to be outliers.

This proved to significantly impact the overall analysis, as the number of customers dropped from 17,000 to 14,825.

Analysis.

My customer analysis was focused on identifying key customer segments centred around age, gender, income and customer acquisition over time. Promotional offer analysis was focused on conversion rate, marketing channel performance, the effectiveness of the rewards system used, and which offer performed the best.

Insights.

• Web, Email, Mobile, Social channels group drove the most sales with the highest conversion rate. The web is an effective means of communicating with customers because the only channel group without the Web had the lowest sales and conversion rate.
• All Rewards were successful. The most successful was the $2 discount promo reward. The customers had to spend $5 to get a $1 reward.
• Both promos were successful but the discount promo drove the most sales with a higher conversion rate. Although the BOGO had a way lower conversion rate, the difference in sales was only $28K.
• Even though 57% of customers are male, female customers drove more sales with the promo.

Recommendations.

• The next promo should target primarily Seniors in the Medium and High income bracket by using Web, Email, Mobile and Social.
• A Discount promo should be offered in combination with a BOGO with less difficulty and less reward.
• Individual Marketing Channels' spend data should be collected so that we can more accurately assess the performance of each.