Deep Data Analytics for New Product Development - Deep Data Analytics for New Product Development -

New product failures Design failures Pricing failures Messaging failures An NPD process The heart of the NPD process Market research Business analytics Summary Notes Ideation: What do you do?Sources for ideas Traditional approaches A modern approach Big Data – external and internal Text data and text analysis Documents, corpus, and corpora Organizing text data Text processing Creating a searchable database Call center logs and warranty claims analysis Sentiment analysis and opinion mining Market research: voice of the customer (VOC)Competitive assessment: the role of CEA Contextual design Machine learning methods Managing ideas and predictive analytics Software Summary Appendix Matrix decomposition Singular value decomposition (SVD)Spectral and singular value decompositions Develop: How do you do it?Product design optimization Conjoint analysis for product optimization Conjoint framework Conjoint design for new products A new product design example Conjoint design Some problems with conjoint analysis Optimal attribute levels Software Kansei engineering for product optimization Study designs Combining conjoint and Kansei analyses Early-stage pricing van Westendorp price sensitivity meter Summary Appendix 3.A Brief overview of the chi-square statistic Appendix 3.B Brief overview of correspondence analysis Appendix 3.C Very brief overview of ordinary least squares analysis Brief overview of principal components analysis Principal components regression analysis Brief overview of partial least squares analysis Test: Will it work and sell?Discrete choice analysis Product configuration vs. competitive offerings Discrete choice background – high-level view Test market hands-on analysis Live trial tests with customers Market segmentation TURF analysis Software Summary Appendix TURF calculations Launch I: What is the marketing mix?Messaging/claims analysis Stages of message analysis Message creation Message testing Price finalization Granger–Gabor analysis Price segmentation Pricing in a social network Placing the new product Software Summary Launch II: How much will sell?Predicting vs. forecasting Forecasting responsibility Time series and forecasting background Data issues Data availability Training and testing data sets Forecasting methods based on data availability Naive methods Sophisticated forecasting methods Data requirements Forecast error analysis Software Summary Appendix Time series definition Backshift and differencing operators Random walk model and naive forecast Random walk with drift Constant mean model The ARIMA family of models Track: Did you succeed?Transactions analysis Business intelligence vs. business analytics Business intelligence dashboards The limits of business intelligence dashboards Case study Case study data sources Case study data analysis Predictive modeling New product forecast error analysis Additional external data – text once more Sentiment analysis and opinion mining Sentiment methodology overview Software Summary Appendix Demonstration of linearization using log transformation Demonstration of variance stabilization using log transformation Constant elasticity models Total revenue elasticity Effects tests F-ratios Resources: Making it work The role and importance of organizational collaboration Analytical talent Technology skill sets Data scientists, statisticians, and machine learning experts Constant training Software issues Downplaying spreadsheets Open source software Commercial software SQL: A must-know language Overall software recommendation Jupyter/Jupyter Lab Bibliography