Predictive Analytics World San Francisco, February 19, 2009 The Unrealized Power of Data Andreas Weigend people & data © people & data | www.weigend.com Andreas S. Weigend, Ph.D. 韦思岸教授 | +1 650 906-5906 | +49 174 906-5906 | +86 138 1818 3800 © people & data | www.weigend.com Outline Q: Current bottleneck for you in your business? (Scarce vs abundant)? Historical perspective Business, Data and Communication Current trends From Transaction Economics to Relationship Economics The Customer Data Revolution: Shift in Customer Expectations Implications: From CRM to CMR Customer Managed Relationships Applications to business: Marketing 2.0 Why predictive analytics: Relevance How to do it: PHAME Problem – Hypotheses – Action – Metrics - Experiments 2 | +1 650 906-5906 | +49 174 906-5906 | +86 138 1818 3800 © people & data | www.weigend.com Business, Data, and Communication 1970’s “Experts” learn a language the computer understands 1bn people poking at stuff 100M people producing stuff Digitizing back office 10M people 1980’s Front office interacts with back office 100M people 2000’s Peer-production and collaboration Customers interact with customers Now Discovery in addition to search Serendipity: Discover what not searched People in addition to pages 1990’s Customers interact with firm Search: 1bn people poking at stuff Social commerce Mobile in addition to PC, and paper) Continuous partial attention Model current situation plus history Sensing 3 | +1 650 906-5906 | +49 174 906-5906 | +86 138 1818 3800 © people & data | www.weigend.com Amount of data Overall : About 100GB per person on the planet Doubling every 1-2 years Mainly user generated Example: Youtube 15 hours of video uploaded every minute Example: Flash 1bn installs 4 © people & data | www.weigend.com | +1 650 906-5906 | +49 174 906-5906 | +86 138 1818 3800 My behavior • IMMI Listening into your room every 30 seconds, for 10 seconds. 5 | +1 650 906-5906 | +49 174 906-5906 | +86 138 1818 3800 © people & data | www.weigend.com Current trends Market research Combine surveys with click data Assumption heavy Data rich model Relationships Interactions Transaction 6 | +1 650 906-5906 | +49 174 906-5906 | +86 138 1818 3800 © people & data | www.weigend.com The Customer Data Revolution 1. Sniffing the digital exhaust Mainly implicit data, some explicit data What is new? More data sources, esp. location data 2. Individuals talk about themselves Mainly explicit contributions 3. Individuals reveal relationships with others Directed, asymmetrical, multidimensional (not binary!) The Customer Data Revolution: Shifting expectations Attitude of individuals to their information Economics of data 7 © people & data | www.weigend.com | +1 650 906-5906 | +49 174 906-5906 | +86 138 1818 3800 Wishlist 8 | +1 650 906-5906 | +49 174 906-5906 | +86 138 1818 3800 © people & data | www.weigend.com Outline Historical perspective Business, Data and Communication Current trends From Transaction Economics to Relationship Economics The Customer Data Revolution: Shift in Customer Expectations Implications: From CRM to CMR Customer Managed Relationships Customer value E-Business Me-Business Who pays whom? Applications to business: Marketing 2.0 9 | +1 650 906-5906 | +49 174 906-5906 | +86 138 1818 3800 © people & data | www.weigend.com Marketing 2.0 Broadcast 1:1 Marketing? Social marketing Implications for predictive analytics: redefining CLV Intrinsic / individual External / network component Applications to business Amazon’s “Share the Love” 10 | +1 650 906-5906 | +49 174 906-5906 | +86 138 1818 3800 downcasting © people & data | www.weigend.com Conversations Conversation / Communication Between whom? Company Individuals 11 | +1 650 906-5906 | +49 174 906-5906 | +86 138 1818 3800 © people & data | www.weigend.com Leverage the social graph Example: New communications service US phone company with deep experience with targeted marketing Sophisticated segmentation models based on experience, intuition, and data e.g., demographic, geographic, loyalty data Hill, S., F. Provost., and C. Volinsky. Network-based Marketing: Identifying likely adopters via consumer networks. Statistical Science 21 (2) 256–276, 2006 . • • 4.82 (1.35%) 2.96 (0.83%) 1 0.4 (0.28%) Non-NN 1-21 (0.11%) NN 1-21 NN 22 NN not targeted Response increases by a factor of 4.82 by marketing to nearest neighbors (NN) From 0.28% based on segmentation, to 1.35% based on social graph 12 | +1 650 906-5906 | +49 174 906-5906 | +86 138 1818 3800 © people & data | www.weigend.com Recommendations 2.0 • People • Friends Data Specific people you know Clicks Purchases Viral marketing Peers Fans (G-star) Experts Forward, tell a friend Relationship Fashion bloggers Annotate Attention Search Intention Location Situation Product data 13 | +1 650 906-5906 | +49 174 906-5906 | +86 138 1818 3800 © people & data | www.weigend.com Outline Historical perspective Business, Data and Communication Current trends From Transaction Economics to Relationship Economics The Customer Data Revolution: Shift in Customer Expectations Implications: From CRM to CMR Customer Managed Relationships Applications to business: Marketing 2.0 Why predictive analytics: Relevance Respect How to do it: PHAME 14 © people & data | www.weigend.com | +1 650 906-5906 | +49 174 906-5906 | +86 138 1818 3800 You want to be PHAME-ous! • PHAME Problem Hypotheses Action Metrics Experiments 15 | +1 650 906-5906 | +49 174 906-5906 | +86 138 1818 3800 © people & data | www.weigend.com Summary Historical perspective Business, Data and Communication Current trends From Transaction Economics to Relationship Economics The Customer Data Revolution: Shift in Customer Expectations Implications: From CRM to CMR (Customer Managed Relationships) Applications to business: Marketing 2.0 Why predictive analytics: Relevance How to do it: PHAME Web: www.weigend.com Phone: +1 650 906-5906 16