May 10, 2012 EDUCAUSE LIVE! Session leader: Jerry Grochow GOAL: Share results of research on ITs role in analytics programs and making them successful. General comments about analytics programs Review of findings to date Research sponsored in part by the International Institute for Analytics, Intel Corporation, and SAS Institute, Inc. Some materials copyright 2011 International Institute for Analytics 2 Questions to be answered: DEFINE: What is “analytics”? INTRODUCE: How do you start an analytics program? WHO: Whose job is it? CSF: What are the critical success factors? ISSUES: Other issues 3 Value-focused data analysis Predictive modeling, optimization – not just statistics Leading to data-driven decision-making A component of “business intelligence” Collection, management, reporting, analytics Characterized by research and experimentation DEFINE INTRODUCE WHO CSF ISSUES 4 EDUCAUSE definition: Analytics is the use of data, statistical analysis, and explanatory and predictive models to gain insights and act on complex issues. Holy Grail: “Dynamic real-time business optimization” DEFINE INTRODUCE WHO CSF ISSUES 5 Business question Terminology What happened in the past? Periodic (regular) reporting, ad hoc reporting, “dashboards” Tell me what happened that wasn’t expected Tell me when something unexpected happens Exception reporting Alerts (real-time exception reporting) • All of these are “reporting” but not really “analytics” activities DEFINE INTRODUCE WHO CSF ISSUES 6 Business question Terminology Tell me something I don’t know Data mining Why is this happening? Analysis, including statistical analysis, on-line analytical processing (OLAP, an older term), modeling What will happen in the future? Forecasting, predictive modeling, predictive analytics How can I improve what happens in the future? Show me graphically Optimization Visualization techniques I want to know now how to improve the “Dynamic real-time business future, based on what happened in the past optimization” based on predictive and everything I know about what is likely analytics – “prescriptive analytics” to happen in the future – and I want to know what steps to take. Take those steps automatically. “Embedded analytics” 7 Establish the value/importance of analytics Set specific business goals and strategy Develop a plan for analytics activities Staffing plan Data plan Technology plan Execute the first project Measure the value Communicate DEFINE INTRODUCE WHO CSF ISSUES 8 The value of analytics is often stated in terms of “understanding” the organization or the business or the customers But this translates into: DEFINE INTRODUCE WHO CSF ISSUES 9 Improved operations [Operational Analytics] Goal: Reduce costs Grow the existing business [Product Analytics] Goal: Increase revenues Improve outcomes of research or academics [Learning Analytics; Research Analytics] Goal: Improve outcomes of research or teaching Innovation Goal: Create new businesses or sources of revenue DEFINE INTRODUCE WHO CSF ISSUES 10 What are the goals of your analytics program? Don’t have an analytics program Haven’t determined goals If you do have program goals, what are they? ▪ ▪ ▪ ▪ ▪ ▪ General understanding Reduce costs of operations Improve outcomes of research or teaching Increase revenues from existing business Create new businesses or sources of revenues Other DEFINE INTRODUCE WHO CSF ISSUES 11 Funding Governance Data architecture Technology architecture Operational implementation and assessment Integration of analytics with operational systems Communication and education Evolution DEFINE INTRODUCE WHO CSF ISSUES 12 Under consideration “Visioning” Getting Started “Launching” Under Construction “Implementing” Mature “Transforming” DEFINE INTRODUCE WHO CSF ISSUES 13 People Process Governance Technology On-board Collaboration between IT and BUs; measuring value Steering Comm; data and analytics governance “SOP” Operational DW; ETL tools; analytics integrated into operational systems; measurement tools; evaluation program “Under Consideration” (Visioning) “Getting Started” (Launching) “Under Construction” (Implementing) “Mature” (Transforming) 14 Typically started by a business function Most successful programs come from a business/IT partnership “Creating a marriage…” Working together on people and process IT has to take responsibility for infrastructure IT can become a “champion” for analytics DEFINE INTRODUCE WHO CSF ISSUES 15 How should analytics activity be organized? One or multiple departments? Where should it report? To IT or not to IT… Does it matter? DEFINE INTRODUCE WHO CSF ISSUES 16 President / Provost / Chancellor Institutional Analytics Academic Hierarchy Administrative Hierarchy IT Research Hierarchy DEFINE INTRODUCE WHO CSF ISSUES 17 President / Provost / Chancellor Academic Hierarchy Institutional Analytics (?) Administrative Hierarchy Research Hierarchy IT Institutional Analytics (?) DEFINE INTRODUCE WHO CSF ISSUES 18 President / Provost / Chancellor Academic Hierarchy Administrative Hierarchy Research Hierarchy IT Institutional Analytics DEFINE INTRODUCE WHO CSF ISSUES 19 Where does the analytics organization report within your institution? [May select multiple] President/Chancellor/Senior Official Provost/Academic Leader CFO/CBO/Administrative Leader Leader of some academic unit Leader of some administrative unit (other than IT) Leader of IT Other DEFINE INTRODUCE WHO CSF ISSUES 20 What determines the likely success of an analytics program? How do you define “success”? Meeting goals [see above] Becoming an “analytical organization” DEFINE INTRODUCE WHO CSF ISSUES 21 Maturity as an “analytic competitor” Stage 1: “Major Barriers” Stage 2: “Local Activities” Stage 3: “Vision Not Yet Realized” Stage 4: “Almost There” Stage 5: “Analytical Competitor” DEFINE INTRODUCE WHO CSF ISSUES 22 Pressing business need Availability of data / data quality Executive leadership/sponsorship Committed, knowledgeable people Clearly defined objectives Focus on analytics that have value to the business Choosing the right first problem Communication/education DEFINE INTRODUCE WHO CSF ISSUES 23 Importance Initial Project Sustaining Program Committed, knowledgeable people: interested in, knowledgeable about analytics M H Executive leadership/sponsorship M H Clearly defined (and initially limited) objectives H M Choosing the right problem: find a pressing business need with high value Communication/education: about the value of analytics, about the importance of the problem being studied H M M H Using the right analytic techniques and software L H Availability and access to quality data L M Critical Success Factors 24 Which CSF’s are in place in your organization? Pressing business need Availability of data / data quality Executive leadership/sponsorship Committed, knowledgeable people Clearly defined objectives Focus on analytics that have value to the business Choosing the right first problem Communication/education DEFINE INTRODUCE WHO CSF ISSUES 25 Analytics Critical Success Factors Committed people 5 4 Technology 3 2 Educated management 1 0 Defined objectives Executive support Quality data DEFINE INTRODUCE WHO CSF ISSUES 26 Focusing excessively on one dimension of analytical capability (e.g. too much technology) Attempting to do everything at once Investing excessive resources on analytics that have minimal impact on the business Investing too much or too little in any analytical capability, compared with demand Choosing the wrong problem, not understanding the problem sufficiently, using the wrong analytical technique or the wrong analytical software Automating decision-based applications without carefully monitoring outcomes and external conditions to see whether assumptions need to be modified.” [Tom Davenport, Competing on Analytics, p. 129] DEFINE INTRODUCE WHO CSF ISSUES 27 Change management Introducing analytics isn’t so different from introducing other new management processes Assessment of implementation (how will you know when you are an “analytic organization”?) assessment of value of analytic program vs. goals Future technology challenges HPC, cloud, anywhere-anytime analysis Unstructured data,“big data” DEFINE INTRODUCE WHO CSF ISSUES 28 1. 2. 3. 4. Can you articulate how analytics will help the organization? Do you have good relationships with the business leaders whose groups will most benefit? Do you know what your peers are doing with analytics and how it is helping their organizations? Are you “passionate about analytics?” [Have you been to an analytics conference or symposium recently?] 29 1. 2. 3. 4. 5. 6. 7. Does the IT department understand the key analytics issues for the enterprise? Do you understand what the key analytics issues are for the IT department? Does your department have the most important skill sets necessary for success with analytics? Do you encourage experimentation? Is your development methodology flexible enough to accommodate analytics projects? How good is the organization’s data? Are definitions consistent? Is data “scrubbed”? Do you know who the vendors and integrators are in analytics IT and what they can do for your organization? Are you prepared for “big data,” high performance computing, realtime analytics – i.e. the future? 30 CIO “Readiness Assessment” -- How did you score? 1-4: Lots of work to do! 5-8: On the way! 9-11: You’re ready to get moving! 12: You know there are more questions! 31 Led by Jackie Bichsel (jbichsel@educause.edu) Survey underway Presentations at ECAR Symposium in Boulder (June 20) 32 Jerrold M. Grochow jerry@jerroldgrochow.com • Senior Consultant to colleges and universities and the organizations that serve them • Internet2 Interim Vice President for NET+ Services 33