Deploying analytics with a rules based infrastructure pawcon sf 2011
3 likes1,117 views
The document discusses the integration of analytics and business rules to enhance decision management, emphasizing that successful analytics must deliver business value. It highlights the challenges of operational analytics and proposes that a rules-based management system can provide a framework to automate and improve decision-making processes. The document also outlines several case studies demonstrating the application of business rules in different contexts to optimize analytics-driven decisions.
![Unmanageable business rules
public class Application {
private Customer customers[];
private Customer goldCustomers[];
...
public void checkOrder() {
for (int i = 0; i < numCustomers; i++) {
Customer aCustomer = customers[i];
if (aCustomer.checkIfGold()) {
numGoldCustomers++;
goldCustomers[numGoldCustomers] = aCustomer;
if (aCustomer.getCurrentOrder().getAmount() > 100000)
aCustomer.setSpecialDiscount (0.05);
}
}
}
©2011 Decision Management Solutions 20](https://image.slidesharecdn.com/deployinganalyticswitharulesbasedinfrastructurepawconsf2011-110316120709-phpapp01/85/Deploying-analytics-with-a-rules-based-infrastructure-pawcon-sf-2011-20-320.jpg)
Deploying analytics with a rules based infrastructure pawcon sf 2011
- 1. Deploying Analytics with a Rules-Based Infrastructure James Taylor, CEO
- 2. Your presenter CEO of Decision Management Solutions Decision Management Solutions works with clients to improve their business by applying analytics and analytic technology to automate and improve decisions Spent the last 8 years developing the concept of Decision Management 20 years experience in all aspects of software including time in FICO, PeopleSoft R&D, Ernst & Young ©2011 Decision Management Solutions 2
- 3. AGENDA 1 2 Zero value analytics are easy Operational analytics are hard(er) 3 Introducing business rules 4 5 6 Deploying Decision Wrap Up analytics with Management business rules
- 4. The one slide you need It is easy to have analytic success without creating business value It is especially easy to fail to deliver business value when focused on operational analytics Business rules and a business rules management system provide an ideal platform for analytics Decision Management ties analytics and business rules together in an effective framework ©2011 Decision Management Solutions 4
- 5. Zero value analytics are easy ©2011 Decision Management Solutions 5
- 6. “ The operation was a success… But the patient died ” ©2011 Decision Management Solutions 6
- 7. “ Making information more readily available is important, but making better decisions based on information is what pays ” the bills. ©2011 Decision Management Solutions 7
- 8. What is a decision? Data is gathered, considered, analyzed A choice or selection is made That results in a commitment to action ©2011 Decision Management Solutions 8
- 9. Operational analytics are hard(er)
- 10. Different kinds of decisions Type Strategy Tactics Operations Low Economic impact High ©2011 Decision Management Solutions 10
- 11. Analytic power in operational decisions How do I… prevent this customer from churning? convert this visitor? acquire this prospect? make this offer compelling to this person? identify this claim as fraudulent? correctly estimate the risk of this loan? It’s not about “aha” moments It’s about making better operational decisions ©2011 Decision Management Solutions 11
- 12. Operational decisions are different High Low High Volume Latency Variability Ensure Manage Personalize Compliance Risk Straight Unattended Self-Service Through Operation Processing After Smart (Enough) Systems, Prentice Hall 200 ©2011 Decision Management Solutions 12
- 13. Insights must drive action ? * ** * ** * * * * * * * * * ** * * ** * * * ** * * * * * * * * ** * * * * * * * ** * * * * ** * * * ©2011 Decision Management Solutions 13
- 14. Time to deploy models matters ©2011 Decision Management Solutions 14
- 15. The three legged stool Business ©2011 Decision Management Solutions 15
- 16. Case: Varolii Personalized, automated consumer communication SaaS Challenge: apply advanced analytics Analyze past behavior of consumers Drive recommendations to their clients Actionable and automatic Solution Identify key decisions Analytically derive new rules based on past success Integrate client rules with analytic rules ©2011 Decision Management Solutions 16
- 18. What are business rules? “… statements of the actions you should take when certain business conditions are true.” ©2011 Decision Management Solutions 18
- 19. Business rules drive decisions Decision Regulations Policy History Experience Legacy Applications ©2011 Decision Management Solutions 19
- 20. Unmanageable business rules public class Application { private Customer customers[]; private Customer goldCustomers[]; ... public void checkOrder() { for (int i = 0; i < numCustomers; i++) { Customer aCustomer = customers[i]; if (aCustomer.checkIfGold()) { numGoldCustomers++; goldCustomers[numGoldCustomers] = aCustomer; if (aCustomer.getCurrentOrder().getAmount() > 100000) aCustomer.setSpecialDiscount (0.05); } } } ©2011 Decision Management Solutions 20
- 21. Manageable business rules If customer is GoldCustomer and Home_Equity_Loan_Value is more than $100,000 then college_loan_discount = 0.5% If member has greater than 3 prescriptions and prescription’s renewal_date is less than 30 days in the future then set reminder=“email” If patient’s age is less than 18 and member’s coverage is “standard” and member’s number_of_claims does not exceed 4 then set patient’s coverage to “standard” Smart (Enough) Systems, Prentice Hall June 2007. Fig 4.3 ©2011 Decision Management Solutions 21
- 22. A Business Rules Management System Validation Testing and Verification Decision Deployment Production Rule Service Application Repository Rule Engine Design Rule Tools Management Applications Operational Database After Smart (Enough) Systems, Prentice Hall June 2007. Fig 6.6 ©2011 Decision Management Solutions 22
- 23. Case: Health Management Personalized health recommendations Challenge: multiple sources of tailoring Medical research Data mining of participant and outcome information Best practices in personal health Solution Replace Java code with JBoss Drools Implement best practices as decision tables Decision trees from analytic results, medical research Implement as additional decision tables ©2011 Decision Management Solutions 23
- 25. Business rules and analytics Broader set of data for business rules to act on Association rules as business rules Decision trees as business rules Predictive (risk) scorecards as business rules ©2011 Decision Management Solutions 25
- 26. Integrate operational and analytic Operational Systems Business Rules Predictive Analytics Analytic Systems ©2011 Decision Management Solutions 26
- 27. Association rules speak for themselves If basket contains Hats AND basket contains Socks THEN offer category is Active Accessories Screenshots courtesy of KXEN ©2011 Decision Management Solutions 27
- 28. Deploying a decision tree Screenshots courtesy of IBM ©2011 Decision Management Solutions 28
- 29. Scorecards are a powerful tool Years Under Contract Reason Codes 1 0 Explaining results 2 5 Transparency More than 2 10 Number of Contract Changes It is really clear how a score card got its result 0 0 1 5 Compliance More than 1 10 Easy to enforce rules about Value Rating of Current Plan use of specific attributes Poor 0 Simplicity Good 10 Easy to use and explain Excellent 20 Easy to implement Score 30 Although not necessarily easy to build ©2011 Decision Management Solutions 29
- 30. Deploying a scorecard Screenshots courtesy of FICO™ ©2011 Decision Management Solutions 30
- 31. The power of business rules Visible, business friendly analytic implementation Avoiding the mistrust of a “black box” Platform for all three groups to share All three legs can participate and collaborate Time to deploy A BRMS handles much of the complexity Support for defining actions Wrap into decisions ©2011 Decision Management Solutions 31
- 32. Integration options Native model execution Generate code or SQL Let the rules call the models when they need them Models as rules Manual or automatic import of models Create rules and rule artifacts that are executable Database scoring Traditional Separate services Let the rules call scoring services ©2011 Decision Management Solutions 32
- 33. Cautions PMML variations still exist Make sure you understand limitations and issues Variable creation and PMML PMML 4.0 supports variable creation Most tools do not export variable definitions Matching data Operational and analytic data are not always the same From a flat analytic data set to object models Once a model is in rules it can be edited…. ©2011 Decision Management Solutions 33
- 34. Case: Major medical insurer Dental Claims Processing Challenge: operationalize fraud models Legacy claims system uses fixed business logic Analytics models predict provider fraud Only currently applied after the fact – pay and chase Solution Add a rules-based decision service to review claims Add rules to define new variables Make analytics visible and reviewable by experts Easily add judgment as well as analytics ©2011 Decision Management Solutions 34
- 35. From analytics to decision management
- 36. Don’t start by focusing on the data Better Analytic decision insight Derived information Available data ©2011 Decision Management Solutions 36
- 37. Start by focusing on the decision Better decision Analytic insight Derived information Available data ©2011 Decision Management Solutions 37
- 38. Decision Service Deploys Analytics ©2011 Decision Management Solutions 38
- 39. Case: Fiserv Core banking systems for mid-sized banks Challenge: create value-add analytic offering Core functionality perceived as commodity Analytics delivers unique value Customers value (but don’t understand) analytics Solution Identify key decisions Build rules-based, cross-channel decision services Automate analytic model creation and deployment Empower customers to “own” these decisions ©2011 Decision Management Solutions 39
- 40. Wrap Up
- 41. The one slide you need It is easy to have analytic success without creating business value It is especially easy to fail to deliver business value when focused on operational analytics Business rules and a business rules management system provide an ideal platform for analytics Decision Management ties analytics and business rules together in an effective framework ©2011 Decision Management Solutions 41
- 42. Action Plan Identify your decisions before analytics Adopt business rules to implement analytics Bring business, analytic and IT people together ©2011 Decision Management Solutions 42
- 43. Thank you! James Taylor, CEO james@decisionmanagementsolutions.com www.decisionmangementsolutions.com