IBM Next Best Action Tony Hocevar Business Analytics Growth Markets
Agenda Introduction to Next Best Action Demonstration Questions
Data is the new oil Ann Winblad Hummer-Winblad
The CMO s Dilemma: Analyze Effectively or Drown 73% of CEOs identify customer insights as the most critical investment area Global Chief Marketing Officer Study IBM Institute for Business Value 2011 Global Chief Executive Officer Study IBM Institute for Business Value 2012
As the voice of the customer in the C-suite and an officer of the company, the CMO has the responsibility and, increasingly, the credibility to lead a customer-obsessed transformation of the business strategy. Global Evolved CMO Online Survey Q3 2011 Forrester/Heidrick & Struggles Forrester Research Inc.
Seeing Customers as Individuals Video
Understanding Each Customer as an Individual Act Predict Analyze Interconnect Collect
Unless a decision has degenerated into work it is not a decision; it is at best a good intention. The Effective Executive, 1967 by Peter Drucker
The Old Marketing Approach I have an offer Who wants to buy? Direct Mail Product A let me find a customer to sell to Agent, IVR Product B F2F Product C Online, Email Product D Mobile, SMS Kiosk
The Next Best Action Approach I have a customer What do they need most? Product A Product B let me find a customer to sell to Customer Analytics service Agent, IVR Brilliant! offer Product C inform Online, Email Product D Governance, Prioritization & Optimization Mobile, SMS
Next Best Action aligns all channels Marketin g Customer Support Buying Patterns Recent Issues Next Best Action Interactio n Profile Financial Profile Channels Customer Accounts
IBM Next Best Action Build relationships, one interaction, one decision at a time Call Center Agent Next Best Action Real-time Business Analyst NEXT BEST ACTION Improve customer satisfaction Optimize revenue Increase lifetime value Marketing Support Cross-channel delivery of best action Customer Data Insights Data Miner s List Comprehensive view of a customer
IBM Next Best Action Demonstration Call Center Agent Marketing Business Analyst Support Data Miner
Dashboard visually shows customer situation At Risk Valuable Influential
Real-time sentiment analysis feeds dashboard
Action tab recommends action
Dashboard reflects higher customer sat
Behind the Scenes Business Analyst (optimized next best action) Marketing (multiple offers) Support (multiple offers) Data Miner (insights & predictions)
IBM Next Best Action Demonstration Call Center Agent Marketing Business Analyst Support Data Miner s List
s are aligned with segments
Campaign includes automated nurturing actions
Case Study Business Challenge Suspects it is wasting money on direct marketing campaigns Focusing on products rather than customer knowledge and behavior Abundance of data but not using it Solution Combine customer segment data with real P&L data to optimize its marketing spend to focus on programs that deliver the highest return on investment (ROI) IBM SPSS and Unica Outcome Increased marketing response rates by 3.1% by more accurately targeting offers to high-value customer segments Achieved a 20% reduction in mailing costs and a 17% reduction in printing costs due to the ability to target the most attractive segment for specific offers Tennessee s oldest and largest bank 5,500 employees 180 bank locations
IBM Next Best Action Demonstration Call Center Agent Marketing Business Analyst Response Propensity Customer Churn Risk Customer Lifetime Value Data Miner s List Support
Quick Quiz Johnny's mother had 3 children. The first child was named April; The second child was named May. What was the third child's name? Source: Forbes, 10 brainteasers to test your mental sharpness, Holly Green
Find Patterns Using Powerful Tools Most segmentation approaches only focus on attributes such as Length of Time as Customer Preferred Channel Preferred Product Categories Time until Repurchase in Key Categories Econometric: Real-estate & Unemployment Annual Spend Level Annual Age + Transactions Income + Geography Segmentation Recency + Frequency + Value Response to Media Breadth of Categories Shopped Return / Exchange Behavior Use of Service Programs Use of In- House Credit Card Participation in Loyalty Program
Find Patterns Using Powerful Tools Segmentation Association Anomaly Detection Social Network Analysis Identify what events occur together
Real-time sentiment analysis feeds dashboard
Structure the unstructured: Text Analytics What is Text Analytics? Puts text into categories to impose structure: positive, neutral, negative What are the benefits? Makes unstructured text quantifiable Can now be mined UNSTRUCTURED STRUCTURED
Modeler interface showing churn propensity calculation Churn Propensity
Case Study Business Challenge Protect existing revenue by reducing customer churn Improve client service without substantially increasing headcount Solution Use data mining to identify at-risk customers Provide insights to client services agents to focus their proactive outbound health check calls IBM SPSS Outcome Reduced customer churn from 1.9% to 1.4% in the first year Decreased the number of client services agents needed for same level of contact Build analytics competence within an inhouse team to quickly extend predictive analytics to new business areas Communications provider in USA Employs >4,000 Revenues >$1.5B
IBM Next Best Action Demonstration Next Best Action Call Center Agent Marketing Business Analyst Response Propensity Customer Churn Risk Customer Lifetime Value Data Miner s List Support
Quick Quiz: What do we know about a shopper if they buy the items below? Cotton balls Hand sanitizers Washcloths Unscented lotion Big purse Zinc supplements Calcium, magnesium supplements 87% chance of pregnancy Kashmir Hill, Forbes, 2/16/2012
Drive Real-time Decisions at the Point of Interaction Optimized decisions + + Business rules Optimization Predictive analytics
Decision models add rules and span depts
Optimization equation finds the best offer
Case Study Business Challenge Agents had to investigate all claims, both high-and low-risk All claims took a minimum of three days to settle Santam began to feel its good reputation for customer service suffer in the age where customers demand fast results Solution A solution that more effectively assessed risk Separated potentially fraudulent claims from lower risk ones in real-time SPSS Decision Management Outcome Reduced processing time by 90% for lowest-risk claims Saved more than US $2.5 million through early fraud detection and prevention in the first six months The largest short-term insurance company in South Africa
IBM Next Best Action Build relationships, one interaction, one decision at a time Call Center Agent Next Best Action Real-time Business Analyst NEXT BEST ACTION Improve customer satisfaction Optimize revenue Increase lifetime value Marketing Support Cross-channel delivery of best action Customer Data Insights Data Miner s List Comprehensive view of a customer
Next Best Action Logical Architecture Deploy Decision Models Reporting Sensors producing readings Internet Social Media Services Customer Conversations Capturing Customer's Activity Process Process Process Process APPLICATIONS Files Shared Staging Area CEP Master Information and System of Record INFORMATION WAREHOUSE Updates, Actions + Outcomes CUSTOMER ACTIVITY Update Sandbox CUSTOMER MASTER SANDBOX PRODUCT MASTER Decision Management Workbench Analytics Management Data Miner 10001 01010 00100 01010 10001 Business Analyst? Decision Model Usage Trigger Additional Information Real-time Decisions Action + Outcome Deploy Decision Model Context Action Outcomes Customer Human Operated Channel Automated Channel Channels Trigger Execute Action MARKETING PLATFORM DECISION ENGINE ENGINE ENTERPRISE SERVICE BUS (ESB) For thorough review read the IBM Red Guide: Smarter Analytics: Driving Customer Interactions with the IBM Next Best Action Solution
Let s Make a Trade Video