Churn Analytics

Customer Churn
 Customer Churn
• Customer churn is one of the most important metrics to be evaluated for a
growing business

• While it's not the happiest measure, it's a number that can give the
company the hard truth about its customer retention

• Customer churn is the percentage of customers that stopped using the

company's product or service during a certain time frame. Any
Company can calculate customer churn by dividing the number of
customers lost during that time period -- say a quarter -- by the number of
customers they had at the beginning of that time period

• For example, if they start their quarter with 400 customers and end with
380, their churn rate is 5% ((400-380)/400) because they lost 5% of their
customers
 Customer Churn Example
• The telecom industry customers can choose from a variety of service
providers and actively switch from one to the next. The
telecommunications business has an annual churn rate of 15-25 percent in
this highly competitive market

• Since most firms have a large number of customers, individualized

customer retention is challenging as they can not afford to devote much
time to each of them

• However, if a corporation could forecast which customers are likely to

leave ahead of time, it could focus customer retention efforts only on
these “high-risk” clients

• To reduce customer churn, telecom companies need to predict which

customers are at high risk of churn
 Why Studying Customer Churn Rate Important?

• It is a critical metric because it costs more to acquire new customers than it

does to retain existing customers

• An increase in customer retention leads to an increase in profit. Also, the

company can spend less on the operating costs of acquiring new customers
 How to Reduce Customer Churn

• Focus the attention on the best customers

• Analyze churn as it occurs

• Show the customers that you care

• Identify the variable that impacts more customer churn through the modelling
 Few Objectives

• What's the % of churn customers and customers that keep in with the active
services?

• Are there any patterns in churn customers based on gender?

• Are there any patterns/preferences in churn customers based on the type of

service provided?

• What are the most profitable service types?

• Which features and services are most profitable?

Classification Problems
 Classification Problems

• Classification problems are an important category of problems in

analytics in which the response variable (Y) takes a discrete value

• The primary objective is to predict the class of a customer (or

class probability) based on the values of explanatory variables or
predictors
 Classification Problems
• Classification is an important category of problems in which the
decision maker would like to classify the case/entity/customers into
two or more groups

• Examples of Classification Problems:

Customer profiling (customer segmentation)

Customer Churn
Credit Classification (low, high and medium risk)
Employee attrition
Fraud (classification of transaction to fraud/no-fraud)
Stress levels
Text Classification (Sentiment Analysis)
Outcome of any binomial and multinomial experiment
Logistic Regression
 Logistic Regression

• Logistic regression is a statistical method used for binary

classification problems

• It is used to predict the probability that a given input belongs to

one or two classes, typically denoted as 0 and 1

• Despite its name, logistic regression is not a regression

algorithm, it’s a classification algorithm
 Logistic Regression – Real Life
Applications

• Churn Detection: Customer will stay or leave

• Spam Detection: Email is spam or not

• Fraud Detection: The transaction is fraudulent or not

• Disease Prediction: Suffering from the condition or not

• Weather Prediction: Will it rain or not

 Why Logistic Regression?

• The Sigmoid function, also known as the logistic function, is a mathematical

function that maps any input value to a value between 0 and 1
• It has an S-shaped curve and is commonly used in logistic regression and
other machine-learning algorithms. The sigmoid function is defined as:
 Logistic Regression

• Logistic regression models estimate how the probability of an event may be

affected by one or more explanatory variables

• Logistic regression is a technique used for predicting “class probability”,

that is, the probability that the case belongs to a particular class
 Binomial & Multinomial Logistic Regression

• Binomial (or binary) logistic regression is a model in which the

dependent variable is dichotomous

• In the multinomial logistic regression model, the dependent variable can

take more than two values

• The independent variables may be of any type

 Formula for Logistic Regression

In logistic regression, the relationship between the input features and the probability
of belonging to the positive class is modelled using the following equation:
Logistic Regression with one Explanatory Variable

1
𝑃 (𝑌 =1∨ 𝑋=𝑥 )= −(𝛼+ 𝛽 𝑥 )
1+𝑒𝑥𝑝

β = 0 implies that P(Y|x) is same for each value of x

β > 0 implies that P(Y|x) is increases as the value of x increases
β < 0 implies that P(Y|x) is decreases as the value of x increases
 Logistic Regression Example

• Predicting Loan Approval

Suppose the bank official is working with a dataset containing information about loan
applicants, and the goal is to predict whether a loan should be approved or not based on
two features: the applicant's income and credit score

• x1: This feature represents the applicant's annual income

• x2 : This feature represents the applicant's credit score

The logistic regression model will assign a coefficient to each of these features to
determine their influence on the prediction
 Customer Churn Prediction
Required Data

• Demographic information about the clients (or accounts): gender, location, age, etc.

• Subscriptions/services related information: the products customers have subscribed to, adds-
on activated, activation and cancellation dates

• Payment information: How much do clients pay? What payment method do they use? Are they
paying regularly?

• Product usage information: login information, clicks information, minutes of interaction with
the product

• Information related to the interaction with customer support: chats or calls made by clients,
rating of support services, claims details

• Etc.
 Customer Churn Prediction
Steps
• Data Reading
• Exploratory Data Analysis
• Data Visualization
• Label Encoding
• Applying Logistic Regression Model
• Prediction
• Feature Importance
• Performance of the model
 Case Analysis

• Predicting Customer Churn at Qwe Inc.

• Telecom Sector

• Employee Churn
Customer Churn Prediction: Telecom Sector

• The Orange Telecom's Churn Dataset, consisting of customer activity data

(features), along with a churn label specifying whether a customer cancelled
the subscription, will be used to develop a predictive model
o Customers who left within the last month – the column is called Churn

o Services that each customer has signed up for – phone, multiple lines, internet, online security,
online backup, device protection, tech support, and streaming TV and movies

o Customer account information – how long they’ve been a customer, contract, payment method,
paperless billing, monthly charges, and total charges

o Demographic info about customers – gender, age range, and if they have partners and
dependents
 Employee Churn Prediction
• Analyze employee churn

• Find out why employees are leaving the company and predict who will leave the company

o satisfaction_level: It is the employee satisfaction point, which ranges from 0-1

o last_evaluation: It is evaluated performance by the employer, which also ranges from 0-1

o number_projects: How many projects are assigned to an employee?

o average_monthly_hours: On average, how many hours an employee works in a month?

o time_spent_company: time_spent_company means employee experience. The number of years spent by an

employee in the company

o work_accident: Whether an employee has had a work accident or not

o promotion_last_5years: Whether an employee has had a promotion in the last 5 years or not

o Departments: Employee's working department/division

o Salary: Salary level of the employee, such as low, medium and high

o left: Whether the employee has left the company or not