Marketing Campaign Analysis using Exploratory Data and Predictive Analyses
An analysis of Atlas Bank marketing leads dataset using Exploratory Data Analysis (EDA) and Market Basket Analysis (Association Rule Mining) in Excel.
A term deposit is one of the arrays of products offered by Atlas Bank, a Portuguese financial institution. The product is a fixed-term investment that enables clients to deposit money into an account at Atlas Bank for an agreed interest rate.
To boost revenue, the bank uses direct marketing through phone calls to reach potential and existing customers to create awareness about and subscribe to a term deposit. The dataset consists of over 40,000 records describing the results of the most recent campaign and a previous campaign, which can be accessed here.
Problem Statement
Atlas Bank has a wide range of clients with varying socioeconomic backgrounds. This together with some economic variables influence customers’ decisions. To increase marketing ROI and maximise marketing efforts, Atlas Bank wants to identify clients that have the highest possibility of subscribing to their term deposits product. The aim of this analysis is to analyse the marketing campaign result dataset using predictive analytics classification techniques and exploratory data analysis to predict the clients with a high potential of converting to subscribers.
Methodology
- Dataset: Using Excel, the raw dataset was converted to a table and cleaned to check for duplicate and null values. Important data fields are identified and explored to establish connections.
- Analysis: Relevant insights were unearthed using exploratory data analysis and Market Basket Analysis in Excel. The results are then communicated using the appropriate visualization.
Exploratory Data Analysis
To identify patterns and trends, the first step is to explore the dataset using relevant exploratory techniques such as Descriptive Statistics, Pivot tables and Pivot charts.
Descriptive Statistics: This was conducted using the Descriptive Statistics tool in the Excel Analysis ToolPak to display the statistical summary of data fields such as age, call duration in seconds and the campaign which describes the number of contacts made to the client. This reveals that the average age is 40, the average call duration is 4 min 18 sec, and the average client contact is up to 3 times.
Education Analysis: The top three subscribers by level of education are university degrees, high school and professional courses.
Time Analysis: The analysis shows that Thursday is the best day to contact clients as the highest number of subscriptions happened on that day.
Comparative Analysis: The dataset provides the results of the previous and current campaigns. When compared, the differences can be found in two significant areas which are the variation in the percentage of subscribers as well as the months with the highest subscribers.
An analysis of subscribers by month in the current campaign indicates that the top months are May, July and August. The high number of subscribers in May could be attributed to the campaign stage which during that month has passed the awareness stage. While the result in July and August could be attributed to the preparation and saving up for the new academic session which starts in September. While for the previous campaign, November was part of its top three months. This is attributed to clients saving up for the end-of-the-year holiday season.
Predictive Analysis
Market Basket Analysis: To identify the patterns or variables that are associated with ‘Subscription’, the Market Basket Analysis, also known as the Association Rule Mining classification technique is applied.
To identify the variables that are highly associated with subscription, the 2-Way and 3-Way Lifts are conducted.
If the lift value is =1, it means there is no correlation
If the lift value is >1, it means there is a positive correlation
If the lift value is <1, it means there is a negative correlation
The 2-Way Lift analysis reveals that the top five categories of clients who are more likely to subscribe to a term deposit are — retired people, students, the unemployed, single marital status and university degree holders.
While the 3-Way Lift analysis was used to determine if relationships exist between two variables and the possibility of subscribing. This revealed that admin workers, the retired and students when paired with being married or single and possessing a professional course certificate or a university degree have a higher possibility of subscribing to term deposits.
Recommendation
- Potential Subscribers: To optimize marketing spend, efforts should be focused on admin workers, retired people, students and the unemployed and if they were to be paired with another variable, admin workers, retired people and students paired with being married, single and professional course certificate and university degree holders have the higher possibilities of subscribing.
- Campaign Period: Marketing activities should be intensified during the months preceding the new school session and the months preceding the December holiday as clients usually prepare ahead financially for these annual events.
- Interest Rate: Client decision is influenced by interest rate. From the analysis, a high-interest rate results in a high number of subscribers. Therefore, high-interest rates should be amplified as part of the product benefit.
- Marketing Tactics: Today’s marketing activities provide brands with more than one way to reach potential clients. Together with telemarketing, other marketing tactics such as content marketing, Meta ads campaign, Google ads, email marketing, search engine optimisation, social marketing, influencer marketing and referral marketing should be considered.
- Marketing Strategy: A campaign game plan should be developed that specifies how to reach the audience and the engagement plan. This would give the campaign an end-to-end holistic approach.
Conclusion
Using exploratory data analysis and market basket analysis (association mining) classification technique, clients that have a high potential of subscribing to a term deposit with Atlas Bank were identified. It was established that relationships exist across jobs, marital status, level of education and clients’ potential to subscribe to term deposits.
Track your progress. Get the Data Analytics Learning Tracker for FREE here.
Thank you for reading till the end. Please check out the project dashboard and GitHub repository. I would love to receive your feedback on this report, kindly share. You can reach out to me via my email — moriamadesolasulaimon@gmail.com and here is my LinkedIn.
