Organizations do not need a Big Data Strategy; they need a Business Strategy that incorporates Big Data

Transcription

1 Organizations do not need a Big Data Strategy; they need a Business Strategy that incorporates Big Data BILL SCHMARZO, CTO, DELL EMC GLOBAL SERVICES UNIVERSITY SAN FRANCISCO, SCHOOL OF MANAGEMENT EXECUTIVE FELLOW 1

2 BIG DATA BUSINESS MODEL MATURITY INDEX Measures degree to which organizations have integrated data and analytics into their business models Key Business Processes BUSINESS METAMORPHOSIS BUSINESS MONITORING Economic Drivers BUSINESS OPTIMIZATION BUSINESS INSIGHTS Prescriptive Recommendations INSIGHTS MONETIZATION SOURCE: BILL SCHMARZO, CTO, DELL GLOBAL SERVICES 2

3 ACTIONABLE BIG DATA BUSINESS STRATEGY BUSINESS INITIATIVE STAKEHOLDERS KEY DECISIONS QUESTIONS DATA ARCHITECTURE & BIG DATA TECHNOLOGIES 3

4 BIG DATA IS CREATING A CONNECTED WORLD 4

5 Your Customers Are Leaving Their Digital Fingerprints All Over Internet 5

6 Every Two Days We Create As Much Information As We Did From The Dawn Of Civilization Until Eric Schmidt, Google CEO A Billion People Visit Facebook Every Day - Business Insider By 2018, Leading Enterprises Will Support 1,000 10,000 TIMES More Customer Touch Points - IDC 6

7 INTERNET OF THINGS BECOMING 30 Billion CONNECTED DEVICES 44 Zettabytes OF DATA Source: Gartner Group, Billion CONNECTED PEOPLE Millions OF NEW BUSINESSES 7

8 THE INTERNET OF ONE 8

9 CREATING IOT VISION: TRANSITIONING FROM CONNECTED TO SMART 9

10 BIG DATA IS SMART SCHOOLS 10

11 BIG DATA IS SMART FACTORIES Copyright Dell EMC Inc. Corporation. All rights reserved. 11

12 BIG DATA IS SMART HOSPITALS 12

13 BIG DATA IS SMART CITIES 13

14 BUT WHAT IS SMART? SMART IS THE SUM OF THE (OPTIMIZED) DECISIONS IN SUPPORT OF THE ENTITY S BUSINESS OBJECTIVES Smart City would include the following Decisions: Roads Bike lanes Pedestrian lanes Traffic flow Parking Road maintenance Building permits Events management Pollution Sustainability Malls Office Parks Crime Parks Schools Internet of Things: Connected Does Not Equal Smart 14

15 WHAT IS DATA SCIENCE AND WHY IS IT IMPORTANT? Copyright EMC Corporation. Dell Inc. All rights reserved. 15

16 WHAT IS DATA SCIENCE? DATA SCIENCE: IDENTIFYING THOSE VARIABLES AND METRICS THAT MIGHT BE BETTER PREDICTORS OF PERFORMANCE 3" A's$vs.$Yankees$Cost$Per$Win$($M)$ 2.5" 2" 1.5" 1" 0.5" 0" Athle-cs" Yankees" 16

17 BUSINESS INTELLIGENCE PROCESS Step 1: Pre-build data schema (schema-on-load) Step 2: Define question to be answered (queries) Step 3: Use Business Intelligence (BI) tool s graphical user interface (GUI) to construct query) Step 4: BI tool creates SQL Step 5: SQL is run against data warehouse to create report DW 17

18 DATA SCIENCE PROCESS Historical Weather Forecast CDC Google Trends Physician Notes Local Events Step 1: Define Hypothesis to test or Prediction to be made Kronos Epic Lawson Step 2: Gather data and more data (Data Lake: SQL + Hadoop) Step 3: Prepare data; Build schema (schema-on-query) REPEAT Step 4: Visualize the data (Tableau, DataRPM, ggplot2, ) Step 5: Build analytic models (SAS, R, MADlib, Mahout, ) Step 6: Evaluate model results (probabilities, confidence levels) 18

19 BI AND DATA SCIENCE DIFFERENCES Business Intelligence Retrospective Reports, Dashboards, KPI s Descriptive Questions What Happened? Data Science Forward-looking view Patterns, Correlations, Coefficients Predictive & Prescriptive Decisions What Is Likely To Happen? Summarized Data Aggregated (Comparative) Analysis Granular Data Individual (Behavioral) Analysis Questions are Informative Decisions are Actionable 19

20 INTELLECTUAL CAPITAL OF THE MODERN ORGANIZATION 20

21 DETERMINING ECONOMIC VALUE OF DATA Data is an unusual currency. Most currencies exhibit a one-toone transactional relationship. For example, the quantifiable value of a dollar is considered to be finite - it can only be used to buy one item or service at a time, or a person can only do one paid job at a time. But measuring the value of data is not constrained by transactional limitations. In fact, data currency exhibits a network effect, where data can be used at the same time across multiple use cases thereby increasing its value to the organization. This makes data a powerful currency in which to invest. Source: Determining the Economic Value of Data 21

22 SOURCES OF INTELLECTUAL CAPITAL IN THE MODERN ORGANIZATION Data detailed business and operational transactions (at the level of the individual) coupled with internal unstructured data sources and growing wealth of external data sources Analytics - combinations of enriched data transformed into analytic results such as scores, association rules, propensities, segments, recommendations, usage patterns and behavioral insights Use Cases - clusters of (optimized) decisions in support of an organization s key business initiatives 22

23 INTELLECTUAL CAPITAL RUBIK S CUBE CHALLENGE How does organization align intellectual capital (data, analytics, use cases) to promote sharing and re-use? How does the organization determine the value of its data (monetization of data in use )? How does the organization avoid data silo s and unmanaged data proliferation? How does the organization avoid the disillusionment of orphaned analytics? How do you re-tool the organization to establish technical and cultural environment for collaborative value creation? 23

24 DATA SOURCE x USE CASE VALUATION Data is the fuel of the modern, intelligent organization an asset to be gathered, enriched and re-used across multiple Use Cases Data Sources Increase Store Traffic via Local Events Increase Store Traffic via Loyalty Program Increase Shopping Bag Revenue Increase Corporate Events Increase Noncorporate Events Increase Promotional Effectiveness Improve NPI Effectiveness Point of Sales Market Baskets Store Demographics Local Competition Store Manager Demo Consumer Comments Social Media Weather Local Events Traffic 24

25 CAPTURE ANALYTIC FOR RE-USE Analytic Profiles are structures that standardize the collection and reapplication of analytics about Business Entities across multiple use cases Bill Schmarzo Chipotle Analytic Profile NCE Score Var Trend 25

26 ANALYTIC PROFILES: CUSTOMER Create scores that support the decisions that support each use case, and store those scores in the appropriate Analytic Profile Bill Schmarzo Chipotle Analytic Profile NCE Score Var Trend Demographic segments Behavioral segments Loyalty Index Frequency Index Recency Index Use Case #1: Campaign Effectiveness Use Case #2: Customer Loyalty Use Case #3: Increase Customer Store Visits Use Case #4: Predict At-risk Customers 26

27 ANALYTIC PROFILES: CUSTOMER Over time, more data is applied and the analytics stored in the Analytic Profiles grow and get refined across multiple use cases Traditional Data Purchases Product Preferences Add-on Preferences Drink Preferences Visit Frequency Visit Recency Visit Monetary Market Basket Group Size Coupons Consumer Comments Store Manager Notes Bill Schmarzo Chipotle Analytic Profile NCE Score Var Trend Demographic segments Behavioral segments Loyalty Index Frequency Index Recency Index Lifetime Value Calc Event Propensity Promotion Propensity Advocacy Propensity Non-traditional Data Social Media Posts Home Value Employment history Job Change Frequency Job Change Recency Industry certifications Industry awards Social Media Connections Education degrees Rank of college College donations Volunteer activities Parking tickets Attrition Propensity

28 ANALYTIC PROFILE x USE CASE VALUATION Analytic Profiles capture and codify insights about Business Entities that support multiple Use Cases Analytic Profiles (Business Entities) Increase Store Traffic via Local Marketing Increase Store Traffic via Loyalty Program Increase Shopping Bag Revenue Increase Corporate Events Increase Noncorporate Events Increase Promotional Effectiveness Improve NPI Effectiveness Customers Products Stores Managers Local Events Competitors Suppliers 28

29 DATA LAKE PROMOTES DATA & ANALYTICS COLLABORATIVE VALUE CREATION Optimize Store Remodeling Increase Store Traffic via Local Marketing Increase Store Traffic via Loyalty Program Improve Manager Retention Improve Customer Retention Data Lake Increase Shopping Bag Revenue Increase Corporate Events Improve NPI Effectiveness Increase Promotional Effectiveness Increase Noncorporate Events 29

30 DATA LAKE REFERENCE ARCHITECTURE HADOOP DISTRIBUTION DATA SCIENCE DATA EXPLORATION DATA ENRICHMENT DATA VISUALIZATION ANALYTIC MODELING DATA CURATOR INGEST PLATFORM MANAGER ADMINISTRATION ANALYTICS CATALOG DATA CATALOG DATA GOVERNOR LINEAGE INDEX PRIVATE CLOUD COMPUTE QUALITY ENRICH HYPERSCALE DATA LAKE REAL-TIME STREAMING SECURITY 30

31 BIG DATA SUMMARY 31

32 PRIORITIZE BUSINESS INITIATIVES 2 3 week process Big Data Vision Workshop Identify big data analytics business use cases Start with business initiative / challenge Drive Business IT organizational alignment Brainstorm and assess potential data sources Build illustrative analytics (data science) Identify, flesh out and prioritize use cases week process Proof of Value Deploy analytics sandbox to quantify the business case For prioritized use case: Build business case (high-level ROI) Prove analytic feasibility (analytic lift) Create UEX mockup Install big data technology Build big data architecture Deliver Go Live recommendations Repeat the process for identified business cases 6-9 month process Applied Analytics Implementation Integrate analytics into operational and management processes; facilitate changes across people, process, data and technology 32

33 Q & A BILL SCHMARZO DELL EMC GLOBAL SERVICES, CTO, BIG DATA PRACTICE Executive Fellow, University San Francisco School of Management Blog: To Achieve Big Data s Potential, Get It Into The Boardroom Big Data Business Model Maturity Index Big Data For Competitive Differentiation History lesson on economic-driven business transformation User Experience: the new king of the business How I ve Learned To Stop Worrying And Love The Data Lake Thinking Like A Data Scientist Contact Information William.Schmarzo@dell.com 33