32
APPLIED STATISTICS ON BUSINESS MANAGEMENT AT SPAIN. A CASE OF STATISTICAL ENGINEERING 1 Igor BARAHONA & Alex RIBA [email protected]

JSM 2012 statistical engineering study case barcelona spain

Embed Size (px)

Citation preview

Page 1: JSM 2012 statistical engineering study case barcelona spain

APPLIED STATISTICS ON BUSINESS MANAGEMENT AT SPAIN. A CASE OF

STATISTICAL ENGINEERING

1

Igor BARAHONA &

Alex RIBA

[email protected]

Page 2: JSM 2012 statistical engineering study case barcelona spain

2

INTRODUCTION

Motivation

What is it done?

THE METHODOLOGY

DISCUSSION

REFERENCES

Page 3: JSM 2012 statistical engineering study case barcelona spain

3

INTRODUCTION

Motivation

What is it done?

THE METHODOLOGY

DISCUSSION

REFERENCES

Page 4: JSM 2012 statistical engineering study case barcelona spain

MOTIVATION

To help Companies to improve their decision making by promoting the use of statistical tools

To understand how a set of 4 key-drivers are related with better analytic performance in companies.

To demonstrate how the Statistical Engineering is a powerful approach for the successful integration of several statistical tools.

To provide another documented case of Statistical Engineering, that was made using data from the real world.

The big motivation for this research is:

Page 5: JSM 2012 statistical engineering study case barcelona spain

5

INTRODUCTION

Motivation

What is it done?

THE METHODOLOGY

DISCUSSION

REFERENCES

Page 6: JSM 2012 statistical engineering study case barcelona spain

WHAT IS IT DONE?

We are proposing a 5-level scale to measure the LAAT at companies.

We are introducing 4 key-drivers for the expansion of the Level of Adoption of Analytical Tools (LAAT)

1. Data-Based Competitive Advantage. (DB-CA) 2. Management Support on Data Analysis. (MS-DA) 3. Systematic Thinking. (SYS) 4. Communication Outside the Company. (COM-OUT)

We are applying the concepts of Statistical Engineering to extract relevant information from the dataset

1. The starting point is a survey with 255 responses. 2. One questionnaire with 21 items was designed. 3. Seven statistical tools were used and integrated. 4. With this is clear that different statistical tools are

complementary rather than exclusive.

1. Statistical Ignorance. 2. Local Focus. 3. Statistical Aspirations. 4. Statistical Engineering. 5. Statistics as Competitive Advantage.

Statistical Methods and Tools

Statistical Engineering

Statistical Thinking

1

2

3

Porter (1998)

Davenport. & Harris (2007)

Deming (2000)

Checkland (1999)

Hoerl & Snee (2010)

McDonough III. (2000)

Davenport & Harris (2010)

Page 7: JSM 2012 statistical engineering study case barcelona spain

7

INTRODUCTION

Motivation

What is it done?

THE METHODOLOGY

DISCUSSION

REFERENCES

Page 8: JSM 2012 statistical engineering study case barcelona spain

Statistics ignorance

Local focus

Statistical aspirations

Statistical Engineering

Statistics as competitive advantage

DM Based on past

experience, judgements and

under uncertainty

GOALS Having data of

quality and improve datasets

Systems

None

DM

ASBM supports only specific tasks and

local impact

GOALS

Improving interaction between functional areas at

company

Systems

Local systems, e.g. Return over

Investment (ROI), Statistical Process

Control (SCP)

DM

The beginning of the ASBM as competitive

advantage.

GOALS

Maintain and improving the

emerging system by working on the Key

Drivers

Systems

Predictions and forecasts of all types,

measurement of intangibles as brand equity and human

capital.

DM

ASBM impact decision making at

strategic, tactical and operational levels

GOALS

strengthening the interaction between

three levels (strategic, tactical and operational)

Systems

Analysis of all types to understand

current and future results. Consolidate

the Business intelligent systems. To make from the

ASBM a competitive advantage

DM ASBM is an

important toolkit for maintaining

the leadership at the market

GOALS

Maintaining the leadership

through creating new and better ways to analyse

data

Systems

Innovation and leadership in the

market, but as well in the

creation metrics and indicators.

1 2

3 4

5 This scale was used for classifying each surveyed company

Page 9: JSM 2012 statistical engineering study case barcelona spain

9

THE ROADMAP From literature review we defined 4 key-drivers

A questionnaire was designed based on the

key-drivers

It was sent to 6460 companies at Barcelona, Spain. 255 responses

received

The Statistical Engineering concept is applied on the

data analysis

“How to best utilize statistical concepts,

methods and tools and integrate them with

information technology to generate improved

results”

Conclusions and discussion

Hoerl & Snee (2010)

Page 10: JSM 2012 statistical engineering study case barcelona spain

10

COMMUNICATION OUTSIDE

COMPANY

DATA BASED. COMPETITIVE ADVANTAGE

SYSTEMATIC THINKING

MANAGEMENT SUPPORT ON

DATA ANALYSIS

APPLIED STATISTICS ON

BUSINESS MANAGEMENT

(ASBM)

Final conclusions.

Understanding project’s scope

Flowchart 1

Survey design and collect data

Operational Definition (DO) *

Principal Component Analysis

2

Applying the scale at companies

Bar Chart Box plots

3

Correspondence Analysis

Factorial Analysis Relationships between companies

4

Relationships between key-drivers

Correlation matrix

Logistic regression 5

This is a 5 steps methodology and is BASED ON STATISTICAL ENGINEERING concepts, as is shown in the following figure.

THE FLOWCHART

Page 11: JSM 2012 statistical engineering study case barcelona spain

This is the questionnaire s structure There are 5 sections and 21 ITEMS in the questionnaire, as it is shown in the following table:

11

section number of ITEMS

General information about the company 4

Data Based Competitive Advantage 5

Management Support Data Analysis 6

Systemacic Thinking 5

Comunication outside the company 1

Total 21

GETTING DATA FROM THE REAL WORLD

https://www.surveymonkey.com/s/ASBM

5-level Likert scale was used on the 17 ITEMS

Page 12: JSM 2012 statistical engineering study case barcelona spain

12

DATASET APPEARANCE

.

.

. . . .

.

.

. . . .

.

.

. . . .

.

.

. . . .

.

.

. . . .

.

.

.

Page 13: JSM 2012 statistical engineering study case barcelona spain

13

QUESTIONNAIRE DESIGN

MS-DA DB-CA SYS COM-OUT1 2 3 4

DB-CA2 .766 DB-CA3 .851 DB-CA4 .707 DB-CA5 .570 .614 MS-DA1 .837 MS-DA2 .753 MS-DA3 .635 .523 DB-CA1 .595 .584 MS-DA4 .831 MS-DA5 .644 .400 SYS1 .433 .595 SYS2 .754 SYS3 .739 SYS4 .630 .528COM-OUT .904MS-DA6 .561SYS5 .430 .519 .534

Rotated Component Matrixa

Component

Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization.

DB-CA. Data-Based Competitive Advantage

MS-DA. Management Support on Data Analysis

SYS. Systemic Vision of the business

COM-OUT. Communication Outside company. (clients and suppliers)

In order to support our conceptual model, the 17 items were clustered on the first 4 factors using the loadings as classification criteria

PCA gave us a quantitative foundation to support our conceptual model

Page 14: JSM 2012 statistical engineering study case barcelona spain

APPLYING THE SCALE

Level Freq Percent Cum. Freq

Cum. Percent

1 65 25.5 65 25.5 2 38 14.9 103 40.4 3 83 32.6 186 72.9 4 52 20.4 238 93.3 5 17 6.7 255 100.0

65

38

83

52

17

L1 L2 L3 L4 L5

25%

15% 33%

20%

7% L1 L2 L3 L4 L5

Companies at Level 3 are the biggest group

Communication outside the company is highest Key-Drivers

Page 15: JSM 2012 statistical engineering study case barcelona spain

15

FACTOR ANALYSIS

The 255 responses were discomposed and represented at the 2 biggest factors

Page 16: JSM 2012 statistical engineering study case barcelona spain

16

FACTOR ANALYSIS

Page 17: JSM 2012 statistical engineering study case barcelona spain

17

FACTOR ANALYSIS

Level 1 is close from Micro Size.

Level 4 is close from Middle Size

Page 18: JSM 2012 statistical engineering study case barcelona spain

18

FACTOR ANALYSIS

Services Companies are more suitable to be analytical oriented

Products Companies are more related with level 1

and Micro size

Page 19: JSM 2012 statistical engineering study case barcelona spain

19

FACTOR ANALYSIS

Middle size companies are closer to “better and different” strategies.

There is a group for Micro-size, Products, Level 1

and No Competitive Advantage

Page 20: JSM 2012 statistical engineering study case barcelona spain

20

CORRELATION MATRIZ ANALYSIS

COMMUNICATION OUTSIDE

COMPANY

DB. COMPETITIVE ADVANTAGE

SYSTEMATIC THINKING

MANAGEMENT SUPPORT. DA

C.M allows us to understand and quantify relationships between the Key Drivers

0.702

0.648

0.300

Pearson Correlation Coefficients

DBCA MSDA SYS COMOUT

DBCA. Data Based Competitive Advantage 1.000 0.70243 0.69484 0.05246

MSDA. Management support data analysis 1.000 0.64852 -0.03397

SYS. Systematic Thinking 1.000 0.30036

COMOUT. Communication Outside Company 1.000

0.695

Page 21: JSM 2012 statistical engineering study case barcelona spain

21

To predict of a set of 255 Spanish companies, either a company has analytics aspirations or not. (Level=>4)

Level 4 is the starting point of the use of data and statistics as a distinctive competence in the industry

RESPONSE VARIABLE:

0=Y

1=Y

If the company does not has analytical aspirations. (Level<4) If the company has analytical aspirations. (Level>=4)

LOGISTIC REGRESSION

NO ANALYTICAL

ASPIRATIONS. (LEVEL 1 , 2AND 3)

ANALYTICAL ASPIRATIONS

(LEVEL 4 AND 5) TOTAL

186 69 255 73% 27% 100%

PREDICTORS

G1 Understanding the benefits of Statistics

G2 Statistics builds the Comp. Adv

G3 There is one mission and vision

G4 Communication with clients and suppliers

The predictors were taken from the

questionnaire ITEMS

Page 22: JSM 2012 statistical engineering study case barcelona spain

22

PROPORTIONAL ODDS

)(432101 ijkllkji GGGGP

PLn εβββββ +++++=

THE MODEL

have p-values less than 0.05, indicating that there is sufficient evidence that the coefficients are not zero using an alfa level of 95%

The goodness-of-tests, with p-value equal to 1.000. Indicate that there is insufficient evidence to claim that the model does not fit the data adequately.

1. UNDERSTANDING THE BENEFITS OF APPLIED STATISTICS BUSINESS.

2. BUILDING A COMPETITIVE ADVANTAGE BY DATA ANALYSIS.

3. ESTABLISHING A MISSION AND VISION STATEMENTS FOR THE COMPANY

4. STIMULATING COMMUNICATION OUTSIDE COMPANY.

Coefficients for these variables are not cero.

Logistic Regression Table Odds 95% CI Predictor Coef SE Coef Z P Ratio Lower Upper Constant -17.8045 3.13596 -5.68 0.000 DB_CA1 1.65439 0.313537 5.28 0.000 5.23 2.83 9.67 DB_CA3 0.723906 0.271505 2.67 0.008 2.06 1.21 3.51 SYS2 1.12321 0.273354 4.11 0.000 3.07 1.80 5.25 COM_OUT 1.54055 0.382019 4.03 0.000 4.67 2.21 9.87

Goodness-of-Fit Tests Method Chi-Square DF P Pearson 105.652 111 0.625 Deviance 72.350 111 0.998 Hosmer-Lemeshow 4.405 8 0.819

Page 23: JSM 2012 statistical engineering study case barcelona spain

23

INTRODUCTION

Motivation

What is it done?

THE METHODOLOGY

DISCUSSION

REFERENCES

Page 24: JSM 2012 statistical engineering study case barcelona spain

24

STATISTICAL ENGINEERING

A set of 7 statistical tools were applied in this research

Different statistical tools can be successfully integrated, in order to extract relevant information from a unique problem

With this, It was demonstrated that several statistical tools can be complementary rather than exclusive

The three previous points are the core philosophy of the Statistical Engineering.

DISCUSSION

Hoerl & Snee (2010)

Page 25: JSM 2012 statistical engineering study case barcelona spain

25

INTRODUCTION

Motivation

What is it done?

THE METHODOLOGY

DISCUSSION

REFERENCES

Page 26: JSM 2012 statistical engineering study case barcelona spain

26

REFERENCES

Banks, D. "Is Industrial Statistics Out of Control?," Statistical Science, (8:4), 1993 , pp. 402-409

Checkland, P. (1999). "Systems Thinking, Systems Practice: Includes a 30-Year Retrospective". Wiley; 1 edition, New York USA

Cronbach, L. J. (1951). "Coefficient alpha and the internal structure of tests" Psychometrika. 16, 297-334.

Davenport, T, and Harris, J. (2010) "Analytics at Work: Smarter Decisions, Better Results". Harvard Business School Press , Boston USA

"Davenport, T. Harris, J. (2007). ""Competing on analytics the new science of winning"". Harvard Business School

Press , Boston USA"

Deming, W.E. (2000). "Out of the Crisis". The MIT Press, Boston USA

Ghemawat, P (2007). "Redefining Global Strategy: Crossing Borders in a World Where Differences Still Matter". Harvard Business School Press, Boston USA

Hoerl, R.W and Snee R.D. "Closing the gap: Statistical Engineering can bridge statistical thinking with methods and tools" , May 2010, pp. 52-55, Quality Press

Hoerl, R.W and Snee R.D. "Statistical Thinking and Methods in Quality Improvement: A Look to the Future", Quality Engineering, (22), 2010, pp 119-129

Hoerl, R.W and Snee R.D. (2001). " Statistical Thinking: Improving Business Performance". Duxbury Press; 1er edition, CA USA

Poon, P., and C.Wagner. "Critical success factors revisited: success and failure cases of information systems for senior executives," Decision Support Systems, (30:4), 2001, pp. 393-418

Porter, M. (1998). "Competitive advantage : creating and sustaining superior performance". Free Press, New Yorw USA

Roberts, H. "Applications in Business and Economic Statistics: Some Personal Views", Statistical Science, (5:4), 1990, pp. 399-402

Stainberg, D.M., "The Future of industrial statistics: A panel discussion", Technometrics, (50:2), 2008, pp103-127

Wang, Y.R., Kon, B.H and Madnick, S.E, "Data Quality Requirements Analysis and Modeling", the Ninth International Conference of Data Engineering, Vienna, Austria, April 1993, pp 670-677

Wang, R.Y and Strong, D. M. "Beyond accuracy: What data quality means to data consumers", Journal of Management Information Systems, (12:4), Spring 1996, pp 5

Yeo,K. "Systems thinking and project management — time to reunite," Int.J.Project Manage, (11:2), 1993, pp 111-117

Page 27: JSM 2012 statistical engineering study case barcelona spain

27

THE END OF PRESENTATION

Page 28: JSM 2012 statistical engineering study case barcelona spain

28

Page 29: JSM 2012 statistical engineering study case barcelona spain

29

Page 30: JSM 2012 statistical engineering study case barcelona spain

30

Page 31: JSM 2012 statistical engineering study case barcelona spain

31

Page 32: JSM 2012 statistical engineering study case barcelona spain

32