Data Driven Business Decisions
October 2011, ©2011
The appropriate use of quantitative methods lies at the core of successful decisions made by managers, researchers, and students in the field of business. Providing a framework for the development of sound judgment and the ability to utilize quantitative and qualitative approaches, Data Driven Business Decisions introduces readers to the important role that data plays in understanding business outcomes, addressing four general areas that managers need to know about: data handling and Microsoft Excel, uncertainty, the relationship between inputs and outputs, and complex decisions with trade-offs and uncertainty.
Grounded in the author's own classroom approach to business statistics, the book reveals how to use data to understand the drivers of business outcomes, which in turn allows for data-driven business decisions. A basic, non-mathematical foundation in statistics is provided, outlining for readers the tools needed to link data with business decisions; account for uncertainty in the actions of others and in patterns revealed by data; handle data in Excel; translate their analysis into simple business terms; and present results in simple tables and charts. The author discusses key data analytic frameworks, such as decision trees and multiple regression, and also explores additional topics, including:
Use of the Excel functions Solver and Goal Seek
Partial correlation and auto-correlation
Interactions and proportional variation in regression models
Seasonal adjustment and what it reveals
Basic portfolio theory as an introduction to correlations
Chapters are introduced with case studies that integrate simple ideas into the larger business context, and are followed by further details, raw data, and motivating insights. Algebraic notation is used only when necessary, and throughout the book, the author utilizes real-world examples from diverse areas such as market surveys, finance, economics, and business ethics. Excel add-ins StatproGo and TreePlan are showcased to demonstrate execution of the techniques, and a related website features extensive programming instructions as well as insights, data sets, and solutions to problems included in the material. The enclosed CD contains the complete book in electronic format, including all presented data, supplemental material on the discussed case files, and links to exercises and solutions.
Data Driven Business Decisions is an excellent book for MBA quantitative analysis courses or undergraduate general statistics courses. It also serves as a valuable reference for practicing MBAs and practitioners in the fields of statistics, business, and finance.
1.1 Setting out business data.
1.2 Different kinds of variables.
1.3 The idea of a distribution.
1.4 Typical performance (the mean).
1.5 Uncertainty in performance (standard deviation).
1.6 Changing units.
1.7 Shapes of distributions.
Chapter 2. What stands out and why? Who Wins? Data driven views of performance dynamics.
2.1 Different layouts of business data.
2.2 Comparing performance across several segments.
2.3 Complex comparisons - using pivotables.
2.4 Unusually high and low outcomes - z scores.
2.5 Choosing a sensible peer group.
2.6 Combining different performance measures.
Chapter 3. Dealing with uncertainty and chance.
3.1 Framing what could happen: outcomes and events.
3.2 How likely is it? Probability basics.
3.3 Market segments and behaviour: Using probability tables.
3.4 Example in health care: testing for a disease.
3.5 Changing your assessment with conditional probability.
3.6 How strong is the relationship? Measuring dependence.
3.7 Probability trees.
Chapter 4. Let the data change you views: Bayes Method.
4.1 Bayes Method in Pictures.
4.2 Bayes Method as an algorithm.
4.3 Example 1. A simple gambling game.
4.4 Example 2. Bayes in the courtroom.
4.5 Some typical business applications.
Chapter 5. Valuing an uncertain payoff.
5.1 What is a probability distribution?
5.2 Displaying a probability distribution.
5.3 The mean of a distribution.
5.4 Example: Fines and violations.
5.5 Why use the mean?
5.6 The standard deviation of a distribution.
5.7 Comparing two distributions.
5.8 Conditional distributions and means.
Chapter 6. Business problems that depends on knowing “how many”.
6.1 The binomial distribution.
6.2 Mean and standard deviation of the binomial.
6.3 The negative binomial distribution.
6.4 The Poisson distribution.
6.5 Some typical business applications.
Chapter 7. Business problems that depends on knowing “how much”.
7.1 The normal distribution.
7.2 Calculating normal probabilities in Excel.
7.3 Combining normal variables.
7.4 Comparing normal distributions.
7.5 The standard normal distribution.
7.6 Example: Dealing with uncertain demand.
7.7 Dealing with proportional variation.
Chapter 8. Making complex decisions with trees.
8.1 Elements of decision trees.
8.2 Solving the decision tree.
8.3 Multistage Decision trees.
8.4 Valuing a decision option.
8.5 The cost of uncertainty.
Chapter 9. Data, estimation and statistical reliability.
9.1 Describing the past and the future.
9.2 How was the data generated?
9.3 The law of large numbers.
9.4 The variability of the average.
9.5 The standard error of the mean.
9.6 The normal limit theorem.
9.7 Samples and populations.
Chapter 10. Managing mean performance.
10.1 Benchmarking mean performance.
10.2 The statistical size of a deviation.
10.3 Decision making, hypothesis testing and P-values.
10.4 Confidence intervals.
10.5 One and two sided tests.
10.6 Using StatproGo.
10.7 Why standard deviation matters.
10.8 Assessing detection power.
Chapter 11. Are these customers different? Did the intervention work? Looking at changes in mean performance.
11.1 How variable is a difference?
11.2 Describing changes in mean performance.
11.3 Example: Is product placement worth it?
11.4 Comparing two means with StatproGo.
11.5 Different standard deviations.
11.6 Analysing matched pairs data.
Chapter 12. What is my brand recognition? Will it sell? Analysing counts and proportions.
12.1 How accurate is a percentage?
12.2 Tests and intervals for proportions.
12.3 Assessing changes in proportions.
12.4 Comparing proportions with StatproGo.
12.5 Alternative methods.
Chapter 13. Using the relationship between shares to build a portfolio.
13.1 How to measure financial growth.
13.2 Risk and return - both matter.
13.3 Correlation and industry structure.
13.4 The riskiness of a portfolio.
13.5 Balancing risk and return.
13.6 Controlling risk with TB's.
Chapter 14. Investigating relationship between business variables.
14.1 Measuring association with correlation.
14.2 Looking at complex relationships.
14.3 Interpreting correlation.
14.5 Untangling relationships with partial correlations.
Chapter 15. Describing the effect of a business input: Linear regression.
15.1 Linear relationships.
15.2 The line of best fit.
15.3 Computing the least squares line.
15.4 The regression model.
15.5 How reliable is the regression line?
Chapter 16. The reliability of regression based decisions.
16.1 Business prediction - three types of questions.
16.2 Estimating the effect of a change.
16.3 Estimating the trend mean.
16.5 Prediction errors and what they tell you.
Chapter 17. Multi-causal relationship and multiple regression.
17.1 Multi-linear relationships.
17.2 Multiple regression.
17.3 Model assessment.
17.4 Prediction and trend estimation.
Chapter 18. Product features, non-linear relationships and market segments.
18.1 Accounting for yes-no features.
18.2 Quadratic relationships.
18.3 Quadratic regression.
18.4 Allowing for segments and groups.
18.5 Automatic model selection.
Chapter 19. Analysing data that is collected regularly over time.
19.1 Measuring growth and seasonality.
19.2 How is the growth rate changing?
19.3 Seasonal adjustment.
19.4 Delayed effects.
19.5 Predicting the future (using auto-regression).
Chapter 20. Extending regression models – the sky is the limit.
20.1 Effects that depends on other inputs - interactions.
20.2 Effects that have proportional impacts.
20.3 Case study: How effective are catalog mail-outs?
20.4 More on time series.
CHRIS J. LLOYD, PhD, is Associate Dean of Research and Professor of Business Statistics in the Melbourne Business School at The University of Melbourne, Australia. Professor Lloyd has extensive international academic and consulting experience in the fields of statistics, data analysis, and market research within both academic and business environments. He has written more than 100 research articles in the areas of categorical data and is the author of Statistical Analysis of Categorical Data, also published by Wiley.