Predicting & Preventing Banking Customer Churn by Machine Learning

Overview

Customer retention is one of the primary goal of any subscription based industry. customers are free to choose from plenty of providers even within one product category. Several bad experiences — or even one — and a customer may quit. And if droves of unsatisfied customers churn at a clip, both material losses and damage to reputation would be enormous.

What to expect

In this article, I am going to show you how to make our hand dirty for building a model that is able to identifies our beloved customers with the intention to leave us in near by future.we are going to build model using machine learning cutting edge technologies. Although our model is not perfect, But I except that I will be able to inspire you from my work so that you can implement in your own organization

Download the datasets from here:

Download is available at github repository

Exploratory Data Analysis

Overall, Our end goal here is better understand our bank’s datasets through visualization and answer the three questions accordingly. We will build our model at end of the article

Getting some insight from data

Actually we should start by investing into the datasets that how data is actually distribute between the retained customer and churned customer. For this, I am going to draw a pie chart so that we can visualize easily

As we can see clearly that our data contains 20.4% churned customer and 79.6% retained customer.

Predictors

We are going look at gender that how our customer gender is related to churning of the customers

On average, women are more likely to churn as compare male. It is tribute to the statistical fact that women are mode conservative than men in terms of risk and reward.So bank may start looking towards there female customer.So that they can decrease churning rate significantly.

Is Active Member

This is the most critical variable, as it gives us indicator that whether or not a customer is availing services of the bank. It will help us to make conclusion that that who are beloved customer and who are going to churn the bank

Non-active members were more likely than active ones to leave a bank. Moreover, non-active members also seemed to be less financially responsible as their credit scores were lower.

This is the time to answer first interesting question.

Which gender has highest churning rate and are they active member or inactive member?

From the above plot it it clear that The proportion of female customers churning is also greater than that of male customers. Unsurprisingly the inactive members have a greater churn. Worryingly is that the overall proportion of inactive members is quite high suggesting that the bank may need a program implemented to turn this group to active customers as this will definitely have a positive impact on the customer churn.

Has Credit Card

It is necessary to know that one who uses banking services like credit card. What are the chances of retaining of them.

Giving shot to second question

Which customer has higher churning rate one who uses our services like credit card or one who not uses services frequently?

Interestingly, majority of the customers that churned are those with credit cards. Given that majority of the customers have credit cards could prove this to be just a coincidence.This remind me that

Correlation does not imply causation

Age

It seems from box plot that younger customer are less likely to leave a bank as they are still not as educated or have the same wealth of choice that middle-aged people have, given that they have not had the time to build on their credit score yet.

Credit Score

A credit score is primarily based on a credit report, information typically sourced from credit bureaus. Lenders, such as banks and credit card companies, use credit scores to evaluate the potential risk posed by lending money to consumers and to mitigate losses due to bad debt.

Tenure

Tenure is the period or duration for which the loan amount is sanctioned. Personal loans, car loans, education loans have shorter tenures as compared to home loans. Some banks and financial institutions extend the loan tenure for an extra fee or a slight increase in interest rates.

Going to answer third question

Is there any relation between age and tenure time?

The older customers are churning at more than the younger ones alluding to a difference in service preference in the age categories. The bank may need to review their target market or review the strategy for retention between the different age groups With regard to the tenure, the clients on either extreme end (spent little time with the bank or a lot of time with the bank) are more likely to churn compared to those that are of average tenure.

If you are inspired to get your hand dirty for building your model then then feel free to clone my github repository

Reflection

Churn rate is a health indicator for subscription-based companies. The ability to identify customers that aren’t happy with provided solutions allows businesses to learn about product or pricing plan weak points, operation issues, as well as customer preferences and expectations to proactively reduce reasons for churn.

It’s important to define data sources and observation period to have a full picture of the history of customer interaction. Selection of the most significant features for a model would influence its predictive performance: The more qualitative the datasets, the more precise forecasts are.

Companies with a large customer base and numerous offerings would benefit from customer segmentation. The number and choice of ML models may also depend on segmentation results. Data scientists also need to monitor deployed models, and revise and adapt features to maintain the desired level of prediction accuracy.