Scribd Logo
Upload

Welcome to Scribd!

  • 19 views

    SPSSProject PPT

    Uploaded by

    mba23shriyapandey
    The document discusses analyzing customer churn in the telecom industry using spreadsheet software. It describes building a logistic regression model to assess factors leading to customer attrition and suggest which factors the firm should focus on to improve customer retention. The model uses data on 42 variables from 1000 customers to determine relationships between variables and churn. It finds that customers are less likely to churn for equipment rental, voice mail, and internet services, while calling cards and multiple lines see higher churn, indicating where the company needs focus.

    Copyright:

    © All Rights Reserved
    Download as PPTX, PDF, TXT or read online from Scribd

    SPSSProject PPT

    Uploaded by

    mba23shriyapandey
    19 views15 pages
    The document discusses analyzing customer churn in the telecom industry using spreadsheet software. It describes building a logistic regression model to assess factors leading to customer attrition and suggest which factors the firm should focus on to improve customer retention. The model uses data on 42 variables from 1000 customers to determine relationships between variables and churn. It finds that customers are less likely to churn for equipment rental, voice mail, and internet services, while calling cards and multiple lines see higher churn, indicating where the company needs focus.

    Original Title

    SPSSProjectPPT

    Copyright

    © © All Rights Reserved

    Available Formats

    PPTX, PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    The document discusses analyzing customer churn in the telecom industry using spreadsheet software. It describes building a logistic regression model to assess factors leading to customer attrition and suggest which factors the firm should focus on to improve customer retention. The model uses data on 42 variables from 1000 customers to determine relationships between variables and churn. It finds that customers are less likely to churn for equipment rental, voice mail, and internet services, while calling cards and multiple lines see higher churn, indicating where the company needs focus.

    Copyright:

    © All Rights Reserved
    Download as PPTX, PDF, TXT or read online from Scribd
    Download as pptx, pdf, or txt
    19 views15 pages

    SPSSProject PPT

    Uploaded by

    mba23shriyapandey
    The document discusses analyzing customer churn in the telecom industry using spreadsheet software. It describes building a logistic regression model to assess factors leading to customer attrition and suggest which factors the firm should focus on to improve customer retention. The model uses data on 42 variables from 1000 customers to determine relationships between variables and churn. It finds that customers are less likely to churn for equipment rental, voice mail, and internet services, while calling cards and multiple lines see higher churn, indicating where the company needs focus.

    Copyright:

    © All Rights Reserved
    Download as PPTX, PDF, TXT or read online from Scribd
    Download as pptx, pdf, or txt
    You are on page 1of 15

    DATA ANALYSIS FOR CONSUMER

    CHURN IN THE TELECOM


    INDUSTRY
    USING SPREADSHEET SOFTWARE

    Aditya Kumar 2022MBA003


    Anjali Singh 2022MBA008
    Aryan Roy2022MBA010
    Deepank kumar Ujjainia 2022MBA015
    Jivesh 2022MBA022
    Rahul Singh 2022MBA042
    Mayank Deep Gwal 2022MBA030
    Priya Pant 2022MBA038
    Problem Description

    Customer churn is very frequent in telecommunications sector, owing to the existence of highly
    competitive market.

    it becomes imperative for companies to determine the propensity for their customers to churn
    and thereby put efforts to retain them.

    Churn analytics offers useful tools for forecasting client churn and defining its fundamental
    causes of it.

    Predictive analytics models can be used to forecast customer churn and can produce a list of
    customers who are most likely to leave, allowing a retention program to focus on them first.
    Problem Description

    Dependent/Criterion Variable - Dichotomous variable labelled as “churn” in the


    dataset

    Predictor Variables - Different services being used: toll free service, equipment rental,
    calling card service, wireless service.

    Regression model applied to the dataset is Logistic Regression which assess factors
    leading to this customer attrition and suggest which factors the firm should be
    focusing on to improve customer retention.
    Data Description
    Data Description
    Data Description

    • Data Set of a telecom Industry

    • We have consumer parameters based on 42 variables

    • Establishing a relationship B/W various variables and consumer churn

    • Factors higher in frequency but have little effect on critical decisions

    • Factors relatively lower in frequency but high impact on churn


    Variable Description

    There are 3 types of Variables


    Nominal : A nominal variable can be defined as
    a categorical variable in which the categories cannot be
    ordered. Ex – Region, Martial
    Scale : A scale variable is a measurement variable — a
    variable that has a numeric value. Ex- Tenure, Age
    Ordinal : An ordinal variable is a variable that can be
    assigned a rank. This rank can be used to determine
    the order in which the variables are arranged.
    Ex- Level of education
    Variable Description
    Dependent/Criterion Variable - Dependent variables are variables that change if there is a change in another variable of
    interest.

    In our report, dichotomous variable labelled as “churn” in the dataset

    Independent/ Predictor Variables - Independent variables are variables that do not change as the result of some other
    variable of interest

    Different services being used: toll free service, equipment rental, calling card service, wireless service.
    Model & Result Discussion
    Binary Logistic Regression Model:
    • Strength of a relationship between one dependent and independent variable(s)
    • To assess the impact of predictor variables on the outcome variable
    • Determine the reason-result relationship of the independent variable(s) with the dependent variable

    Case Processing Summary indicates the No signifies = 0 (People who haven’t


    sample size. Here, the number of switched) and whereas Yes signifies = 1
    respondent entries is 1000. (People who have switched)
    Model & Result Discussion

    Describes the baseline model. This will serve as a baseline later for
    comparing the model with our predictor variable included. The
    predictions of this baseline model are made purely on whichever
    category occurred most often in our dataset. In our dataset, the
    model always guesses ‘no’ because more participants did not opt
    out of the service being provided. The overall percentage row tells
    us that this approach to prediction is correct 72.6% of the time.
    Model & Result Discussion
    The goodness-of-fit helps in determining whether the model adequately describes the data.

    Used to check that the new model (with predictor


    variables included) is an improvement over the
    baseline model. The significance values are p < .001,
    which indicates the accuracy of the model improves
    when we add our criterion variables.

    Hesmer & Lemeshow’s Test of the goodness of fit


    suggests the model is a good fit to the data as
    p=0.688 (>.05). There is no difference between
    the observed and predicted model.
    Model & Result Discussion

    The model adequately fits the data. There is no


    significant difference between observed and
    predicted models. Both values are approximately
    equal.
    Model & Result Discussion
    Provides an indication of how well the model is able to
    predict the correct category once the predictor is added
    to the study. Our model is now correctly classifying the
    outcome for 77.2% of the cases compared to 72.6% in
    the null model.

    This table shows the relation between the predictor and


    the outcome. Depending on the odds ratio, we can
    determine whether the probability of falling into the
    target group is greater/equal/less than the probability
    of falling into the non-target group category.
    Observation & Conclusion

    • For equipment rental, voice mail, and internet services, we can say that the odds of the customer not switching are
    higher. This indicates that people are satisfied with these service offerings.

    • For other services such as calling cards and multiple lines, customers are more likely to churn, indicating that the
    company needs to focus on these offerings.

    • Considering the confidence interval: Values greater than 1 mean that as the predictor variable increase, so do the odds of
    not churning.

    • Churn Rates are essential for a business perspective to know their customer demand with respect to the service they are
    providing. Customer strategy according to demand can then be made in line with the profit of the business.
    Thank You!

    You might also like