Published on March 20, 2014
What is your #BigData Strategy? @mpiskorski Associate Professor of Strategy Harvard Business School
2© 2014 Mikołaj Jan Piskorski Big Data: Large benefits Big data opportunity – Volume – Velocity Profitability higher by 5% points – Variety
3© 2014 Mikołaj Jan Piskorski Big Data: But huge strategy challenge 0 5 10 15 20 25 30 35 40 Top execs don't know how to turn data into business value Management has no bandwidth Business lines don't know how to turn data into business value Ability to get the data Culture does not emphasize data sharing Ownership of data is unclear Perceived costs outweigh perceived benefits No case for change Percentage of respondents
4© 2014 Mikołaj Jan Piskorski Big Data: Strategy challenge 0 2 4 6 8 10 Financial management Operations and production Strategy and business development Sales and marketing Customer service Product research General management Risk management Customer experience Brand management Workforce management Times more likely to implement 0 5 10 15 20 25 30 35 40 Capture information Aggregate information Analyze information Disseminate insights Percentageofrespondents Top performers Low performers
5© 2014 Mikołaj Jan Piskorski Big Data: Strategy challenge Big Data Opportunity Big Data Strategy Big Data Advantage
6© 2014 Mikołaj Jan Piskorski Big Data strategy creates new products or increases profitability of existing ones by integrating various datasets and applying analytic models to create better decision support tools Big Data: Strategy statement
7© 2014 Mikołaj Jan Piskorski Big Data: Four elements of Big Data strategy Business impact Data integration Analytic models Decision tools
8© 2014 Mikołaj Jan Piskorski Four elements of Big Data strategy: Example from an insurance company
9© 2014 Mikołaj Jan Piskorski Big Data: Four elements of Big Data strategy Business impact Data integration Analytic models Decision tools
10© 2014 Mikołaj Jan Piskorski Business impact: Building new products to create data ecosystems Nike’s devices collect data Phone/Computer Online social platform Twitter/Facebook
11© 2014 Mikołaj Jan Piskorski Business impact: Building new products from old products with Big Data General Electric invested $250,000,000/year over 4 years into industrial Internet – For example, turbines now have 100 physical sensors and 300 virtual sensors to help customers identify problems before they occur and improve fuel efficiency – A turbine generates over 500GB of data per day which General Electric uses to build its next generation turbines – 1% efficiency increase = $20,000,000,000 per year in savings for customers… Even if GE only captures 10% of these savings, ROI is very high!
12© 2014 Mikołaj Jan Piskorski Business impact: Building new products to create data ecosystems Data ecosystems Data shared with consumer Consumer uses product more Customer relationship management Time to buy new product Consumer shares data with friends and strangers Lower acquisition costs Create new data products Aggregate data for new product development
13© 2014 Mikołaj Jan Piskorski Four elements of Big Data strategy Business impact Data integration Analytic models Decision tools
14© 2014 Mikołaj Jan Piskorski Data integration: Merging various data for business impact Facebook has emails and phone numbers of its users Your CRM has emails and phone numbers of your customers Custom audience can be exposed to advertising on Facebook or Twitter or LinkedIn Use your own loyalty program to see if those who were exposed reacted (great for offline purchase) No wasted advertising means higher ROI
15© 2014 Mikołaj Jan Piskorski Data integration: Merging unstructured data in real time Bank card used here My phone is here Joint venture between a European bank and a big European mobile phone operator Huge value to everyone involved, but… – Can the operator sell the data? – Who should pay for this?
16© 2014 Mikołaj Jan Piskorski Four elements of Big Data strategy Business impact Data integration Analytic models Decision tools
17© 2014 Mikołaj Jan Piskorski Analytic models for Big Data: Typical HR analytics Business Goals Most organizations Hiring Manual interpretation of CV Person to person interview process Performance Largely standardized training emphasizing skills Retention Managed during annual review React when employee announces departure Org design Static organizational designs
18© 2014 Mikołaj Jan Piskorski Analytic models for Big Data: Google People Operations analytics Business Goals At Google More effective hiring Algorithms read CVs to compare those accepted to those rejected and invites some applicants back Higher performance Project Oxygen identified eight characteristics of great leaders which had little to do with skills Now Google teaches, measures and rewards those Higher retention Algorithms that predict when someone will leave allow Google to be proactive to stop that Better org design “What if” scenarios to anticipate who should work with whom and on what kinds of projects • “All people decisions at Google are based on data and analytics.” • “We bring the same rigor to people-decisions that we do to engineering decisions.”
19© 2014 Mikołaj Jan Piskorski Four elements of Big Data strategy Business impact Data integration Analytic models Decision tools
20© 2014 Mikołaj Jan Piskorski Decision tools for Big Data strategy: Caesars Casinos (previously known as Harrah’s Casinos) Leader in customer loyalty through analytics – Total Rewards program became industry standard Customers never come back if – Lose when playing slot machines – Have to stand in lines to get to food Can prevent if customer receives offer or gets expedited – Must do before customer decides to leave Slot machines deliver information real time analysis free food offer Real time video analysis staff informed expedite top players to get food
21© 2014 Mikołaj Jan Piskorski Decision tools for Big Data strategy: Bank of America Big data • Record voice conversations with customers • Transcribe them Analytical models • Natural language processing to understand meaning • Identify patterns of dissatisfaction • Correlate with “customer journey” through other channels Decision tools • Next time customer calls, the agent can offer incentives to stay… • … or new products to prevent defection • Can also be used to target customers proactively Over $500 million in revenue that would be otherwise lost
22© 2014 Mikołaj Jan Piskorski Big Data strategy process Business impact Data integration Analytic models Decision tools
23© 2014 Mikołaj Jan Piskorski Big Data strategy process: Strategy development over time Business impact data integration * Start today… * Pick an average performing business * Identify sources of competitive disadvantage for that business * What kinds of data + analytic models + decisions tools will alleviate those Data integration business impact * ½ way through the previous process challenge them to figure out what new data integrations will get business results. Implement if good ideas * Set up another team and challenge them to figure out new data integrations for business results. Implement if good ideas New products data business * Give the winning team the challenge to develop new products that will collect data for business impact * Note that this team will need a lot of support for new product development * In some cases you might need to set up a separate division for them
24© 2014 Mikołaj Jan Piskorski Big Data strategy process: Data integration business impact What is your unique data? – LinkedIn = unique career data product can help you choose career – Google search = data on what people don’t know identify concerns – Netflix = data on what movies people like can make your own hit – What will your unique data be? How can your data be merged with outside data? – Location – Sentiment data – Purchase history – Weather
25© 2014 Mikołaj Jan Piskorski Big Data: Four elements of Big Data strategy Business impact Data integration Analytic models Decision tools
26© 2014 Mikołaj Jan Piskorski Implementing Big Data strategy People Org structure Culture
27© 2014 Mikołaj Jan Piskorski People: Data scientists Data Scientists Hacker • Codes • Understands technology Analyst • Statistics • Unstructured data Scientist • Makes sound decisions • Improvises Advisor • Strong communication skills Business • Knows business strategy
28© 2014 Mikołaj Jan Piskorski People: Data scientists Data Scientists Hacker • Python, Hive, Pig • Hadoop, MapReduce • Hackathons Analyst • Machine Learning • Natural Language Processing Scientist • No need for PhD • But experimental physics helps • Do it yourselfer! Advisor • Translate data and science into business goals • Tell story with data Business • Customer insights • Understands competitive reaction
29© 2014 Mikołaj Jan Piskorski People: Data scientists Hacker + Analyst + Scientist + Advisor + Business Hacker + Analyst Hacker + Scientist Scientist + Advisor + Business Analyst + Advisor + Business Team approach ok, but must have overlapping skills Where do I find them? – Junior university scientists – More EU universities begin to offer data science programs – Get IBM, Accenture, Deloitte to train them for you How do I retain them? – Train, train, train and – give them really ambitious projects… – … or expect to lose them within 18 months
30© 2014 Mikołaj Jan Piskorski Organizational structure: Where does your BigData team sit and who leads it? Center of excellence or embedded in divisions? – Depends on the underlying strategy Business impact data: Embedded in business unit Data business impact: Center of excellence Product data business impact: Matrix approach Leadership – What do you call them in top executive functions? Senior VP of Big Data, Social Design, and Marketing: Intuit SVP of Analytics, Insight and Loyalty: Charles Schwab Chief Data Officer: Bank of America
31© 2014 Mikołaj Jan Piskorski Organizational structure: Where does your BigData team sit and who leads it?
32© 2014 Mikołaj Jan Piskorski Culture Belief that technology disrupts business models Top management commitment to – internal innovation it is OK to fail – external innovation best inventions do not have to be invented here Data settles disputes in the organization – “what do the data say?” – “where did these data come from?” Open access – many people have access to data and can experiment with it
Small deck used during Use Case roundtable at JiveWorld 2014. On each on the 14 ta...
This 30 minute presentation was given at the 2014 Rochester Young Professionals En...
Creating a successful big data strategy means cutting through the information noise and finding useful business insights. This big data strategy ...
WHITE PAPER INTEGRATE FOR INSIGHT Enterprises must learn to understand how best to leverage big data soon, since the amount of data being generated
Bob Gourley, Pam - Data Divination: Big Data Strategies jetzt kaufen. ISBN: 9781305115088, Fremdsprachige Bücher - Netzwerke, Protokolle & APIs
Four Big Data strategies 1. Performance Management. Performance management involves understanding the meaning of big data in company databases using pre ...
From the headlines to the C-suite, big data embodied the height of strategy-as-fashion in 2013.
The Big Data Strategy Conference is coming back to Lithuania. The list of technologies used to solve Big Data problems is ever expanding, giving rise to ...
Big data strategy and roadmap for your enterprise, to improve your business performance by maximizing the value of your data.
Die Datenberge in Unternehmen wachsen. Mit der richtigen Big-Data-Strategie geht es von den Rohdaten zur verwertbaren Erkenntnis in 5 Schritten zur Lösung.
Die Komponenten einer Big-Data-Strategie. Xavier Verhaeghe, Vice President, Technology Solutions, Oracle Europe, Middle East & Africa @XVerhaeghe
Big Data: Technologien, Strategien & Trends Schulungen: alle oeffentlichen, Inhouse-Seminare oder Firmenseminare werden von hochqualifizierten Referenten ...