How To Automate Smart Decision Making with BI & Analytics

How To Automate Smart Decision Making with BI & Analytics Mark Smith – CEO & EVP Research ©2006, Ventana Research, Inc. Agenda  What is "decision management technology"  What kinds of decisions should not be automated?  What are rule-based systems and why are they advantageous?  What's are main the differences between BI, Analytics and Decision Management?  How are the above practices/technologies measured? 2 ©2006, Ventana Research, Inc. Excessive Focus on Understanding: Rear View, Historical Perspective 3 ©2006, Ventana Research, Inc. Performance Management Provides Emphasis on Improvement and Automation of Decisions Performance Management A methodology that examines the people, process, information and technology and a process that aligns business and technology to improve results. Performance Management Process Align – Communicate and provide guidance, objectives, goals and incentives. Optimize – Plan for future actions and decisions that can achieve the objectives and goals. Understand – Comprehend the status of people, organization and business processes. Ventana Research PerformanceCycle™ 4 ©2006, Ventana Research, Inc. Decision Management What is It? Use analysis of the past to modify decision strategies in the present to correctly manage and guide future activities  Improve business decisions through design, deployment and refinement of operational decision making procedures  Decision Management helps enterprise make better, more consistent business decisions – Maximize revenue, minimize costs, streamline business processes – Via Web, OLTP, call center, and batch applications – Agile evolution of decision making to meet new business challenges  Give data analysts, business policy makers, corporate portfolio managers, and it implementers tools to understand and use each others information 5 ©2006, Ventana Research, Inc. Decision Management Examples Customer management decisions  how to handle particular requests for services, how to treat customers in collections, whether to increase exposure to current customers through additional credit offers, etc. Supplier selection decisions  which supplier to use for a given order of parts based on cost, past delivery history, customer preferences, etc. Advertiser selection decisions  which advertiser should be selected for a given advertisement opportunity given existing contracts, type of opportunity, recently shown ads, etc. Audit selection decisions  which taxpayers should be selected for auditing given experience with previous audits, past filings, current policies, etc. 6 ©2006, Ventana Research, Inc. Examining Decision Cycles for Automation Information Acquisition 2- Understand Factors 1- Recognize Threat/Opp. 8- Review Action results Action Coordination Traditional BI use 3- Set Goals/ Objectives 7- Execute Action Plan 6- Decide on Action Plan Automated Decision Systems 5- Compare Solutions 4- Devise Solutions Goal Setting Decision Making Areas of facilitation of the decision making process 7 ©2006, Ventana Research, Inc. Decision Management and Automation Automated does not mean lights out  Decision review sometime or always required  Decision exception management  Rules engine maintenance and updates Automated Decisions Input Reviews Models/Rules Exceptions 8 ©2006, Ventana Research, Inc. Decision Management Framework Decision Management Applicationa decision Application or User requesting or Interface Repository Business Rule Engine Third Party Applications or Interfaces Development Environment Data Sources Predictive Models 9 ©2006, Ventana Research, Inc. Business Rules – Decision Management Rules are conditional statements of how to respond to data or events with a business transaction  If Ferrari is $30, then buy quickly Business rules take many forms  Policies, practices, and procedures you use to make operational decisions  Strategies based on historical data and business acumen  Decision Trees  Decision Tables Rules carry out actions suggested by modeling and optimization strategies Business rules are found in your company policy manuals, in lines of code in legacy systems, interdepartmental memos, and in the minds of your experienced employees 10 ©2006, Ventana Research, Inc. Business Rules – Decision Management Rules Engines  Evolution of Expert systems  Two step process – Create rules using experts – Deploy the rule set  What’s new – Graphical development and administration – Mature rule development environments – Robust platforms – Integration with complimentary technologies – Accessible to business managers/non-technologists – Industry experience with success and failure 11 ©2006, Ventana Research, Inc. Predictive Modeling – Decision Management Predictive Modeling  Applied statistics and other data mining approaches – Regression/Neural Networks – Classification, Segmentation  2 step process – Create the model, use the model  What’s new – Improved, graphical user interfaces – Automation of model creation – Integration with rules engines – Industry experience with success and failure 12 ©2006, Ventana Research, Inc. Decision Management - Maturity Curve Decision Optimization Automated access Automated processing Routed Decisions Predictive Models Manual analysis Manual decisions Predictive Modeling Automated access Automated processing Decision Testing Predictive Models Manual analysis Manual decisions Adaptive Control Decision Refinement Automated processing Leveraged Expertise Decision Testing Automated Decisions Manual Tuning Leveraged Expertise Automated Decisions Manual Tuning Business Rules Automation No Optimization Automated access 13 ©2006, Ventana Research, Inc. Decision Management Application Needs E-Commerce now a driving factor in real-time transactions  $80B annual run rate (Q4CY2004) (US Census Bureau)  +29% from Q3CY2004  2% of total US sales Retailing Merchant Survey  Among Top 5 key selling tools – Cross-sell/Up-sell, Promotions, Sales/Specials Commerce Fraud  2004 On-line commerce fraud losses $2.6B (CyberSource survey)  2004 On-line losses growth 25%-30% growth expected  Manual review increased Speed of e-commerce  Interactive transaction rates now 10-20 seconds for leading sites (Keynote) Skills and turnover of workforce  10%-30% turnover (Call Center Mgmt Review) 14 ©2006, Ventana Research, Inc. Best Practices in Identification of Decision Management High decision volume  Manual oversight is not technically or economically feasible.  Potential for respective cost savings or incremental revenue Inconsistent workforce decision skills  High employee turnover  Low position compensation Need for self-service  Business demand a presence on the web  Services require instant decision making on pricing, packaging, delivery and other facets.  Web-site visitors expect instant response. Past indicates the future  Decision input/outcome relationship can be measured and modeled  Decisions consistent over time Internet is a key business enabler  Customers, and employees expect fast, high-value interactions  Comparable performance, personalization and accommodation are a critical success factor for your business. 15 ©2006, Ventana Research, Inc. Converging Technologies from Business Intelligence to Decision Management Tools & Applications Dashboard & Scorecard Monitoring & Alerting Analytics & Location Reporting & Visualization Rules & Workflow Data Mining & Modeling Platform and Server Technology Information Management & Master Data Mgt Integration Management (Application, Data & Info.) 1995-2004 Query & Analysis Platforms Reporting 16 2006Modeling & Analytics Rules & Workflow Master Data Management ©2006, Ventana Research, Inc. Decision Management Enables Performance Management Inconsistent Decision Making Decision Mgt Benefits Encoded Decision Process Limited Decision Competence Business Performance Understanding Management Facilitated Decisions Automated Decisions Shrinking Response Time Improved ROI Decision Risk 17 ©2006, Ventana Research, Inc. Summary Decision Management is part of Performance Management and Expands your Business Intelligence and Analytics Efforts Value       Faster response to customers and partners Greater consistency of decision making Lower cost of decision making Overall decision outcome improvement Knowledge workers available for higher value tasks IT costs lowered as application design are simplified 18 ©2006, Ventana Research, Inc. Ventana Research Analyst Summit Connections 2006 – San Francisco in June  June 11-14 in San Francisco, California at Fairmont  Sunday Night Welcome Reception, Monday and Tuesday, and Wednesday Conference Workshops in morning  Content: Performance Management  3 Tracks: Finance, Operations, IT  Keynotes:  Winning SunTrust Racing Team  Scott Adams – Author of Dilbert 19 ©2006, Ventana Research, Inc. Conclusion – Contact Information Ventana Research Mark Smith – CEO & EVP Research 650-931-0888 mark.smith@ventanaresearch.com www.ventanaresearch.com 20 ©2006, Ventana Research, Inc. About Ventana Research We are the leading Performance Management research and advisory services firm. We help business and IT be more efficient and effective by providing expert insights and guidance. Our research and advisory knowledge provides direct value for assessment and recommendations Consulting Advisory Our research and experience is leveraged throughout a year for our clients to maximize investments Research Our research is the foundation that provides valuable insight to our knowledge of market 21 ©2006, Ventana Research, Inc.