Product Analytics: Applied Data Science Techniques for Actionable Consumer Insights (Pearson Business Analytics Series)

This guide shows how to combine data science with social science to gain unprecedented insight into customer behavior, so you can change it. Joanne Rodrigues-Craig bridges the gap between predictive data science and statistical techniques that reveal why important things happen -- why customers buy more, or why they immediately leave your site -- so you can get more behaviors you want and less you don’t. Drawing on extensive enterprise experience and deep knowledge of demographics and sociology, Rodrigues-Craig shows how to create better theories and metrics, so you can accelerate the process of gaining insight, altering behavior, and earning business value. You’ll learn how to: Develop complex, testable theories for understanding individual and social behavior in web products Think like a social scientist and contextualize individual behavior in today’s social environments Build more effective metrics and KPIs for any web product or system Conduct more informative and actionable A/B tests Explore causal effects, reflecting a deeper understanding of the differences between correlation and causation Alter user behavior in a complex web product Understand how relevant human behaviors develop, and the prerequisites for changing them Choose the right statistical techniques for common tasks such as multistate and uplift modeling Use advanced statistical techniques to model multidimensional systems Do all of this in R (with sample code available in a separate code manual) Cover Page About This eBook Half Title Page Title Page Copyright Page Dedications Contents Preface Acknowledgments About the Author I: Qualitative Methodology 1. Data in Action: A Model of a Dinner Party 1.1 The User Data Disruption 1.2 A Model of a Dinner Party 1.3 What’s Unique about User Data? 1.4 Why Does Causation Matter? 1.5 Actionable Insights 2. Building a Theory of the Social Universe 2.1 Building a Theory 2.2 Conceptualization and Measurement 2.3 Theories from a Web Product 2.4 Actionable Insights 3. The Coveted Goalpost: How to Change Human Behavior 3.1 Understanding Actionable Insight 3.2 It’s All about Changing “Your” Behavior 3.3 A Theory about Human Behavioral Change 3.4 Change in a Web Product 3.5 What Are Realistic Expectations for Behavioral Change? 3.6 Actionable Insights II: Basic Statistical Methods 4. Distributions in User Analytics 4.1 Why Are Metrics Important? 4.2 Actionable Insights 5. Retained? Metric Creation and Interpretation 5.1 Period, Age, and Cohort 5.2 Metric Development 5.3 Actionable Insights 6. Why Are My Users Leaving? The Ins and Outs of A/B Testing 6.1 An A/B Test 6.2 The Curious Case of Free Weekly Events 6.3 But It’s Correlated ... 6.4 Why Randomness? 6.5 The Nuts and Bolts of an A/B Test 6.6 Pitfalls in A/B testing 6.7 Actionable Insights III: Predictive Methods 7. Modeling the User Space: k-Means and PCA 7.1 What Is a Model? 7.2 Clustering Techniques 7.3 Actionable Insights 8. Predicting User Behavior: Regression, Decision Trees, and Support Vector Machines 8.1 Predictive Inference 8.2 Much Ado about Prediction? 8.3 Predictive Modeling 8.4 Validation of Supervised Learning Models 8.5 Actionable Insights Appendix 9. Forecasting Population Changes in Product: Demographic Projections 9.1 Why Should We Spend Time on the Product Life Cycle? 9.2 Birth, Death, and the Full Life Cycle 9.3 Different Models of Retention 9.4 The Art of Population Prediction 9.5 Actionable Insights IV: Causal Inference Methods 10. In Pursuit of the Experiment: Natural Experiments and Difference-in-Difference Modeling 10.1 Why Causal Inference? 10.2 Causal Inference versus Prediction 10.3 When A/B Testing Doesn’t Work 10.4 Nuts and Bolts of Causal Inference from Real-World Data 10.5 Actionable Insights 11. In Pursuit of the Experiment, Continued 11.1 Regression Discontinuity 11.2 Estimating the Causal Effect of Gaining a Badge 11.3 Interrupted Time Series 11.4 Seasonality Decomposition 11.5 Actionable Insights 12. Developing Heuristics in Practice 12.1 Determining Causation from Real-World Data 12.2 Statistical Matching 12.3 Problems with Propensity Score Matching 12.4 Matching as a Heuristic 12.5 The Best Guess 12.6 Final Thoughts 12.7 Actionable Insights 13. Uplift Modeling 13.1 What Is Uplift? 13.2 Why Uplift? 13.3 Understanding Uplift 13.4 Prediction and Uplift 13.5 Difficulties with Uplift 13.6 Actionable Insights V: Basic, Predictive, and Causal Inference Methods in R 14. Metrics in R 14.1 Why R? 14.2 R Fundamentals: A Very Basic Introduction to R and Its Setup 14.3 Sampling from Distributions in R 14.4 Summary Statistics 14.5 Q-Q Plot 14.6 Calculating Variance and Higher Moments 14.7 Histograms and Binning 14.8 Bivariate Distribution and Correlation 14.9 Parity Progression Ratios 14.10 Summary 15. A/B Testing, Predictive Modeling, and Population Projection in R 15.1 A/B Testing in R 15.2 Clustering 15.3 Predictive Modeling 15.4 Population Projection 15.5 Actionable Insights 16. Regression Discontinuity, Matching, and Uplift in R 16.1 Difference-in-Difference Modeling 16.2 Regression Discontinuity and Time-Series Modeling 16.3 Statistical Matching 16.4 Uplift Modeling 16.5 Actionable Insights Appendix Conclusion Bibliography Index Code Snippets