Borges Map How Big Data changes business strategy European Leasing & Consumer Credit Industry Annual Conventions Barcelona, October 9, 2014 Philip Evans
The world becomes its own map What the Google self-driving car sees
The world becomes its own map What the Google self-driving car sees Layers of modular, interoperable, frequently open-source software Successive probabilistic algorithmic approximation Machine learning, not 80s AI 50B IPenabled sensors by 2020 1-10 Trillion sensors of all types 2017-2025 The world s stock of data doubling every 2 years 99% digitized 50% has an IP address 2.2B broadband mobile connections by 2015 Smartphones the fastest-adopted technology ever except tablets
The world becomes its own map Smart parking
The world becomes its own map What the retailer sees and the customer
The world becomes its own map Data as infrastructure Côte d Ivoire Source: http://www.dailymail.co.uk/news/article-2170422/revealed-the-stunning-images-europe-wastes-energy-pollutes-light-africa- South-America-Asia-darkness.html
Data as infrastructure 2.5 billion call records in Côte d Ivoire Dec 2011-Apr 2012 Mobility network 1 Communication network 2 1 Trajectories of 50,000 randomly selected individuals between sub-prefectures over a 5 month period 2 Number and duration of calls between pairs of cell phone towers aggregated by hour Nodes color-coded by predominant language Source: Exploiting Cellular Data for Disease Containment and Information Campaigns Strategies in Country-Wide Epidemics A. Lima, M. De Domenico, V. Pejovic, and M. Musolesi
Data as infrastructure Optimizing the bus network in Abidjan, Côte d Ivoire Current bus network Off-network travel Travel volume to one destination Optimized network Four new routes proposed. No additional resources required 10% reduction in city-wide travel times Source: AllAboard: a system for exploring urban mobility and optimizing public transport using cellphone data Michele Berlingerio, Francesco Calabrese, Giusy Di Lorenzo, Rahul Nair, Fabio Pinelli, Marco Luca Sbodio IBM Research https://www.youtube.com/watch?v=9givvecxnee
The hardware: Google Data Center 2013 One of 81: Council Bluffs, Iowa
The software: Google Search A technology stack ¼ sec USERS WEB PAGES DNS (FRONT-END) SERVERS MAPREDUCE MAP REDUCE MAP REDUCE MAP REDUCE MAP REDUCE Adwords/ Adsense Server Ad Auction Algorithm Crawler PageRank Algorithm Spellcheck API Spell Algorithm Google Translate API Translate Algorithm BIG TABLE 1,000 servers /query INDEX AND DOCUMENT SERVERS
Top of the stack: Mashups Housingmaps.com: the first real estate search engine
Walmart, Tesco, Sainsbury s Data aggregation via loyalty cards Walmart (90s) Tesco Nectar Customer Basket Transaction data: SKU x POS x time Customer data: Identity (50 descriptors) x SKU x POS x time Customer data: Identity (50 descriptors) x SKU x POS x time SKU POS Store Retailer Aggregator Manufacturer
From value chains to stacks The typical layers Experiments, recombination (low cost of failure) Distributed user & developer innovation, swarming Social capital (reputation, reciprocity) embedded in communities Large, thin platforms, driven by network effects Coexistence of collaboration & competition Hierarchical projects Linear: spec plan execution completion Disciplined replicable processes Engineering, reengineering, end-to-end optimization Co-design of handoffs, interfaces Infrastructure: Build it, and they will come Risk-avoiding, robust, resilient (high cost of failure) Long investment horizons Capital-intensive Scale- & utilization-intensive
From value chains to stacks A self-replicating template INSTITUTIONS ORGANIZATION IT Experiments, recombination, Distributed user innovation Social capital, communities Platforms, network effects Collaboration + competition Peer communities trading user-generated content Open source & developer communities Social networks Enterprise 2.0 Curated platforms Contests, innovation jams Revenue/growth Scrum, agile Mashups, hacks Algorithmic iteration Advanced analytics A/B testing [APIs] Hierarchical projects Linear: spec to completion Disciplined replicable process Engineering, reengineering Co-design of handoffs Traditional oligopolists Vertically-integrated or in close vertical relationships Portfolio of parallel businesses Divisional, SBU Tight, vertical end-to-end management ROCE, ROI Projects focused on product/segment requirements (ERP, CRM, etc) Surveys, trials [Virtualization] Infrastructure Risk-avoiding, robust Long investment horizons Capital-intensive Scale- & utilization-intensive Utilities, co-ops Public investment Data repositories Open data Cloud services Functionallyconsolidated services Cost centers Efficiency metrics Data centers Networks Security Compliance Data base integrity
The elephant in the room Facebook s default privacy settings 2005-2010 2005 2006 2007 2009 2010 (Nov) (Dec) (Apr) Source: Matt McKeon http://mattmckeon.com/facebook-privacy/ Number of people (logarithmic) 1 40 1K 5M 1B
BCG THE BOSTON CONSULTING GROUP evans.philip@bcg.com