Statistics for managers, Multiple regression analysis

16
Statistics For Managers Live-Project On Multiple Regression Analysis

Transcript of Statistics for managers, Multiple regression analysis

Page 1: Statistics for managers, Multiple regression analysis

Statistics For Managers

Live-Project

On

Multiple Regression Analysis

Page 2: Statistics for managers, Multiple regression analysis

Project Introduction

• Under this project, we collected sales data from an insurance brooking company “NetAmbit Insurance Brooking Pvt Ltd.”, for consecutive 30 months starting from July-2010 to December-2012.

• By using Graphical data representation technique, we made histogram and scattering diagrams for the available data.

• We identified independent variables, & thereafter used “Minitab” Statistical software tool to obtain Multiple regression equation.

Page 3: Statistics for managers, Multiple regression analysis

Data Overview

• Data Obtained, included:

• Number if Insurance Policies sold monthly

• Premium received (month-wise)

• Brokerage received (month-wise)

• Interval : July-2010 to December-2012

• In specified Interval, the company generated an approximate Revenue of 400 Crores.

Page 4: Statistics for managers, Multiple regression analysis

Collected Data (Brief) Month (MMM-YY) Insurance Policy (Unit

Sales)

July – 2010 235

August – 2010 2822

September – 2010 2618

October – 2010 4246

November – 2010 5532

December – 2010 6441

January – 2011 4978

February – 2011 6103

Mar – 2011 12787

April – 2011 11226

May – 2011 11426

June – 2011 13880

July – 2011 13640

August – 2011 13647

September – 2011 14413

October – 2011 13694

November – 2011 16191

December – 2011 20256

January – 2012 17962

February – 2012 18280

March – 2012 21810

April – 2012 12999

May – 2012 18855

June – 2012 19924

July – 2012 18683

August – 2012 19541

September – 2012 19941

October – 2012 19199

November – 2012 19844

December - 2012 22756

Page 5: Statistics for managers, Multiple regression analysis

Unit Policy Sales

0

5000

10000

15000

20000

25000

2010 2011 2012

jan

feb

march

april

may

june

july

aug

sept

oct

nov

dec

Started & Hired; Invested in Technology; JFM exponential increase; Approx flat in mid-yr; high in yr-end; drop after JFM

Page 6: Statistics for managers, Multiple regression analysis

Problem

Sales Figure shows a dip

during the month of April.

Why?

Page 7: Statistics for managers, Multiple regression analysis

Regression Line

0

5

10

15

20

25

30

35

0 5000 10000 15000 20000 25000

Highly ???????

Page 8: Statistics for managers, Multiple regression analysis

Depth Analysis:

Jan-12 to Dec-12

0

2

4

6

8

10

12

14

0 5000 10000 15000 20000 25000

April

Highly ???????

Page 9: Statistics for managers, Multiple regression analysis

Depth Analysis:

Jan-11 to Dec-11

0

2

4

6

8

10

12

14

0 5000 10000 15000 20000 25000

April

Highly ???????

Page 10: Statistics for managers, Multiple regression analysis

Depth Analysis:

Jul-10 to Dec-10

0

1

2

3

4

5

6

7

0 1000 2000 3000 4000 5000 6000 7000

April Not Applicable; Company started

operations from July

Highly ???????

Page 11: Statistics for managers, Multiple regression analysis

Cause of Problem Independent

Variable Name Cause

New Financial Year

As April marks the beginning of a new Financial year,

people do not generally feel a need to purchase life

insurance policy and therefore the sales show a downward

trend during the month of April

JFM ended

*(JFM = Jan + Feb + Mar)

Most of the people tend to purchase the policies by the

end the financial year, in the month of JAN-MAR for

Tax planning, which becomes a prime reason for low sales

in the month of April

Page 12: Statistics for managers, Multiple regression analysis

Premium paid during JFM months

Premium for policies like LIC, etc. have to be paid By the

end of a financial year which impacts the financial

position of the customers affecting their ability to buy new

policies during this period. Now the propensity to save is

more evident.

Impact of other expenses

People are burdened with other expenses in the beginning

of a new financial year which also affects the sale of life

insurance policies. Some of the expenses may be the

academic fees of children, interest on loans,

Independent Variable Name

Cause of Problem

Cause of Problem

Page 13: Statistics for managers, Multiple regression analysis

Limitation Faced

• We rationally came up with independent variable, as discussed before.

• But Multiple Regression Analysis using MINITAB, required discrete numeric values available in Companies data.

• So, we had to take following numbers to determine multiple regression equation:

• Number of Policies sold

• Premium Received

• Brokerage Received

Page 14: Statistics for managers, Multiple regression analysis

Multiple Regression Equation

Equation Obtained using MINITAB:

Premium = 405048 + (3735 x Policy count) + (3.45 x Brokerage)

Predictor Coef SE Coef T P

Constant 405048 217536 1.86 0.063

Policy count 3735.3 137.8 27.11 0.000

Brokrage In 3.4499 0.1123 30.73 0.000

S = 3498572 R-Sq = 94.2% R-Sq(adj) = 94.2%

Page 15: Statistics for managers, Multiple regression analysis

Conclusion and Solution

• Being Insurance brooking company, sales see a dip during April, because customers have spent enough money during JFM months and they tend to save money after that.

• Hence in this case, as a brooking company, I know that principal cost of booking a policy cannot be decreased by policy issuing firm, though my brokerage charges can be decreased by my side, so the grand total amount paid my customer is decreased, and as a result policy count will increase.

• This can be said as P-value for brokerage is 0.00, making this independent variable as a significant predictor of premium.

Page 16: Statistics for managers, Multiple regression analysis

Thank You !!!

A live-project presentation by

Abhitanjay Chaudhary (A0102214093);

Owais Ashraf (A0102214161);

Rajat Sharma (A0102214079);

Anant Prakash Gupta (A0102214086)