From BI to Predictive Analytics
0 likes1,858 views
The document discusses the evolution of business intelligence into predictive analytics, emphasizing the importance of decision management systems (DMS) for improving organizational decision-making processes. It outlines steps for implementing DMS, including discovering key decisions, building predictive models, and continuously improving decision performance. Various case studies illustrate the benefits of analytics in making informed decisions, reducing costs, and enhancing customer interactions.
From BI to Predictive Analytics
- 1. From Business Intelligence to Predictive Analytics James Taylor CEO
- 2. Your Presenter – James Taylor CEO of Decision Management Solutions Works with clients to improve their business by applying analytic technology to automate & improve decisions Spent the last 9 years championing Decision Management and developing Decision Management Systems ©2012 Decision Management Solutions 1
- 3. The Value of Business Intelligence Improving the quality of decisions ©2012 Decision Management Solutions 2
- 4. Patterns Inform Decision-making 10 5 0 Day 1 Day 2 Day 3 Day 4 Day 5 Day 6 Day 7 Day 8 Day 9 Day 10 ©2012 Decision Management Solutions 3
- 5. Time To Look Forward ©2012 Decision Management Solutions 4
- 6. The Growing Power of Analytics Processing power has increased and data storage costs have dropped The increased data that organizations have amassed Large and growing body of mathematical knowledge ©2012 Decision Management Solutions 5
- 7. From BI to Predictive Analytics ? © Decision Management Solutions, 2011 6
- 8. 3 Steps to Decision Management Discover Build Improve ©2012 Decision Management Solutions 7
- 10. Different Kinds of Decisions Strategic Decisions • Few in number, large impact • Should we acquire this company or exit this market? Tactical Decisions • Management and control, moderate impact • Should we re-organize this supply chain, change risk management approach? Operational Decisions • Day-to-day decisions that affect one transaction or customer • Best offer for this customer? Which supplier? How to handle this claim? ©2012 Decision Management Solutions 9
- 11. Operational Decisions Are Everywhere ©2012 Decision Management Solutions 10
- 12. Decisions implement strategy © Decision Management Solutions, 2012 11
- 13. Decisions are the focal point for risk Risk is not acquired in “big lumps” but one transaction at a time © Decision Management Solutions, 2012 12
- 14. Decisions maximize customer value © Decision Management Solutions, 2012 13
- 15. Decisions scale for large impact Strategic Decision Tactical Decision Operational Decision © Decision Management Solutions, 2012 14
- 16. Case study: Cable TV Business challenges Solution Benefits 1.2M households Predictive analytics to 13-18% cross-sell hit Many single-product predict churn, cross- rate on average households sell Up to 40% cross-sell Whole industry suffers Business rules use success rate for some from low loyalty and analytics and data to drive dynamic scripts Teams using the 20%+ customer churn scripts have more Increasing Embedded in call sales competition and center application to improve decision Reduced churn by 20- changing regulations 30% making ©2012 Decision Management Solutions 15
- 17. Advice: Begin with the Decision in mind Discover Build Improve Find the decisions that matter to your business and model them © Decision Management Solutions, 2012 16
- 19. Decision Management Systems deploy and apply predictive analytics ©2012 Decision Management Solutions 18
- 20. Analytics Must Drive Action Operational Systems Decision Management Systems link Decision analytics to operational systems Analytic Systems ©2012 Decision Management Solutions 19
- 21. Decision Management Systems Agile Analytic Adaptive ©2012 Decision Management Solutions 20
- 22. Analytics, Business And IT Business Decision © Decision Management Solutions, 2011 21
- 23. ©2012 Decision Management Solutions 22
- 24. Manage the rules of decisions Decision © Decision Management Solutions, 2012 23
- 25. What Can You Do With Business Rules? Automate Claims Personalize the experience Detect fraud Create loyalty Target Cross-Sells And more… ©2012 Decision Management Solutions 24
- 26. Case: Global Manufacturer Business challenges Solution Benefits Supplier onboarding Extract “Validate 50% reduction in time consuming and Supplier” decision supplier onboarding time manual Automate and manage All local variations Standard process using business rules supported across 175 countries— Genuinely global “Intelligent” self-service hundreds of local process applications exceptions 3,000 supplier updates a month
- 27. Three kinds of Predictive Analytics Risk Fraud Opportunity © Decision Management Solutions, 2011 26
- 28. Analytics Predict Risk How risky is this customer’s application for service… And how should we price it? ©2012 Decision Management Solutions 27
- 29. Analytics Predict Fraud How likely is this claim to be fraudulent…. and what should we do about it? ©2012 Decision Management Solutions 28
- 30. Analytics Predict Opportunity What represents the best opportunity to maximize loyalty and revenue? And when should we promote it? ©2012 Decision Management Solutions 29
- 31. Embed Predictive Analytics ? ? Decision ©2012 Decision Management Solutions 30
- 32. Case study: Specialty Insurer Business challenges Solution Benefits 12,000 claims a Business rules and Loss ratio expense month predictive analytics from 14% to 11% Reduce staff by 25% Automatically identify 32% higher in a recession and subrogation subrogation returns lower expenses opportunities $10M/year additional Reduce fraud and Increase Fast Track subrogation returns improve subrogation rate from 2% to 22%
- 33. Advice: Find decision-making rules Discover Build Improve Analyze and manage the business rules that underpin your operational decisions © Decision Management Solutions, 2012 32
- 34. Advice: Industrialize Predictive Analytics Discover Build Improve Decision Become efficient at building and embedding predictive analytic models © Decision Management Solutions, 2012 33
- 35. Continuously Improve Decisions
- 36. Measure decision performance © Decision Management Solutions, 2012 35
- 37. Improve for Increasing ROI ©2012 Decision Management Solutions 36
- 38. and for Adaptive Systems ©2012 Decision Management Solutions 37
- 39. Experiment To Learn And Adapt ©2012 Decision Management Solutions 38
- 40. Case: State dept of taxation Business challenges Solution Benefits Paper tax returns Single central Recovered millions of increased costs and taxpayer database dollars from dubious slowed responses Integrated system tax returns Information system Sophisticated real- Increased collection silos time predictive of unpaid taxes Manual fraud analytics Decreased number of detection and return questionable returns review Increased customer satisfaction
- 41. TAKEAWAYS
- 42. To Decision Management Systems © Decision Management Solutions, 2012 41
- 43. Successful Predictive Analytics Pervasive Used in every transaction At the point of contact/delivery In operational decision making Predictive From reporting to prediction and forecasting Data mining Predictive analytics and scoring Actionable Decisions being made, actions being taken Decision Management Systems Decision Support Systems ©2012 Decision Management Solutions 42
- 44. Decision Management Systems What if you could make your systems active participants in optimizing your business? What if your systems could act intelligently on their own? Decision Management Systems can do all that and more. This book shows how to integrate operational and analytic technologies to create more agile, analytic, and adaptive systems. Discount Code: TAYLOR4389 For more information about this new release, visit www.decisionmanagementsolutions.com/book © Decision Management Solutions, 2012 43
- 45. Questions
- 46. Thank You James Taylor, CEO james@decisionmanagementsolutions.com 45
Editor's Notes
- #2: To ensure that Decision Management Systems are analytic and adaptive you must embed the results of data mining and predictive analytics in them. In this webinar you will learn what can be discovered using data mining and predictive analytic techniques and how this can be applied to the decision-making embedded in Decision Management Systems. The role of analytics in predicting risk, fraud and opportunity and the importance of continuous improvement and learning will also be covered.
- #4: Find the decisions that matter to your business and understand themDon’t start with your data, start with the decisions you need to improve
- #5: This future focus for decisions contrasts with what we typically do with BIBI has historically focused on the past – like this chart of the last 9 days of salesThis works, for people, because they can see patterns and extrapolateYou, for instance, would have no difficulty in estimating day 10 as being around 9Enter Predictive Analytics
- #6: The interest and excitement around predictive analytics is sometimes described in terms of a move away from looking in the rear view mirror to looking forwardAnd doing so with software not with human intelligenceIf humans can extrapolate from the past, why is this necessary?As the road ahead starts to differ from the road behind, as we must decide more quickly or in real-time, and as we need systems to do more of the deciding
- #7: Data Mining and Predictive Analytics are increasingly importantSome companies are beginning to use data mining and predictive analytics as key elements of their strategy and more will do so over time.There are two main reasons data mining and predictive analytics have become more important recently:First there is the increase in data that most organizations have amassed and the growth in third-party data providers. Secondly processing power has increased and data storage costs have dropped, making this type of technology affordable to more businesses. Add to these trends the large and growing body of mathematical knowledge around the techniques and the stage is set for a massive expansion as witnessed the emergence of some mainstream books on the topic such as Competing on Analytics, Super Crunchers and Numerati. These books, and some more technical references, are listed in the bibliography under data mining.
- #8: So how do you get from BI to PA?If BI and analytics are about improving decisions then Predictive Analytics must improve how we make decisionsBut what kind of decisions can they help with? And how do we adopt and use predictive analytics? What are the steps we must take
- #9: Discover and Model DecisionsDesign and Implement Decision ServicesMonitor and Improve Decisions
- #10: Focus on the day to day decisions that drive operational success.Operational decisions are everywhereThey implement strategyThey affect customersThey are the focal point for risk and opportunityThey multiply for large scale impact
- #12: What parts will the engineer need to repair this problem?What offer should we make when this person uses the ATM?Is this credit card fraudulent?
- #13: Strategy mattersBut it must be made real and executed onNot enough to simply say “we will improve customer retention” – to define a strategic intent and measuresMust figure out the tactical approaches to decision-making that will be required and make day to day decisions that will make the strategy happen
- #14: Risk is not acquired in big lumps but one bad loan, one fraudulent transaction at a time
- #15: Being customer centric means focusing on each customer and maximizing the value of interactionsMake decisions about a single customer, maximizing the value of that decision for next best action or retention or cross-sell
- #18: Find the decisions that matter to your business and understand themDon’t start with your data, start with the decisions you need to improve
- #19: Build independent decision-making componentsManage detailed dataManage the rules of decisions–rules from policies, from dataEmbed predictive analytic modelsBring all three groups together
- #20: The power of predictive analytics is their ability to turn uncertainty about the future into usable probability
- #21: [twitter]#decisionmgt systems link analytic systems to operational systems[/twitter]
- #22: [twitter]#decisionmgt systems are agile, analytics and adaptive[/twitter]AgileChanging CircumstancesComplianceProcess ImprovementAnalyticManaging RiskReducing FraudTargeting and RetainingFocusing ResourcesAdaptiveFinding New ApproachesTesting and LearningManaging Trade-offs
- #26: Changing ExpectationsReal-Time ResponsivenessGlobal Customers Expect Global ServiceSelf-ServiceThe 24/7 WorldChanging ScaleBig DataEfficiencyTransaction VolumesChanging InteractionsMobile interactionsSocial interactionsDistributed interactionsChanging Organizations
- #27: Payment methods Bank details Tax numbersWithholding taxBusiness partner roles DUNS numbersVendor returns
- #28: Risk – credit risk, delivery risk, price risk. Some upside if get right, big downside if get wrongFraud – good fraud decisions really have no effect but bad ones are a loss e.g. credit card fraud or claims fraudOpportunity – not much of a downside but a degree of upside e.g. cross-sell or up-sell
- #29: Story
- #30: Story
- #31: Story
- #33: Focus on decision-making – the rules, the measures, who makes which decision, how do you tell good ones from bad ones
- #34: Don’t just predict things, embed those predictions in operational systems
- #35: Use your data to adapt your response to evolving problems and opportunities.Measure decision performance so you can improve it Good decision making approaches and good outcomes are distinctUse performance management to monitor finance operations and decision makingDecisions change continuously so Decision Management Systems adaptExperimentation helps Decision Management Systems stay effective and become more effective
- #36: [twitter]#decisionmgt systems test and learn, improving over time[/twitter]Measure decision performance so you can improve it Good decision making approaches and good outcomes are distinctUse performance management to monitor finance operations and decision making
- #39: Decisions change continuously so Decision Management Systems adaptExperimentation helps Decision Management Systems stay effective and become more effective
- #42: [twitter]3 steps to #decisionmgt systems – discover, build, improve[/twitter]
- #43: In a predictive enterprise, analytics are applied systematically to improve operational decisions… (slide 8)Predictive analytics can then take full advantage of all data and know-how and apply intelligence at every transaction.
- #44: [twitter]Buy the book to learn more about #decisionmgt systems http://bit.ly/n4p25H [/twitter]