1. Michael N. Ferrara, Jr. Sr. VP Field Operations February 17, 2009 Predictive Analytics for Fleet Safety

2. Over fifty years of combined experience in the transportation industry. Industry leader in predictive modeling for transportation safety. Very broad client base. Intelligent Intervention Systems. Minimize Effort, Maximize Results FleetRisk Advisors

3. A leveled approach Calibration, Planning and Customizing are the keys

4. What is it… Innovative and sophisticated use of statistical and quantitative data to build predictive models that drive human decisions, and fully automated decisions, to predict future events and dramatically improve performance. Predictive Analytics

5. Some perspective….. In business, as in baseball, the question isn’t whether or not you’ll jump into analytics; the question is when! Do you want to ride the analytics horse to success…..or follow it with a shovel? Rob Neyer, ESPN Analytics will be the gold rush of the next several years, and many companies will be assiduously panning their data for treasure. Michael Treacy, cofounder, GEN3 Partners

6. Every 30-90 days you knew which drivers were most likely to have an event, or type of event, with a 65- 85% accuracy rate before they happen? What if I told you that the driver does not have to go online, fill out any forms, take any tests or submit any personal data that would violate any company policy or privacy regulations? What if you knew whether it was a personal, professional or skill issue that was impacting the driver. What if there was an intelligent intervention system that could effectively manage the issue? What if……

7. Modeling Platform Analytics Engine Utilize your data for a 2 nd generation ROI! Vehicle Type Traffic Population Other Weather Fatigue

8. © 2008 QUALCOMM Incorporated. External presentation to (audience), prepared by Qualcomm’s (presenters name) – Month Day, 2008. 8 7 8 9 10 4 5 6 1 2 3

9. Parking & Backing 12345678910

10. Over The Road Carrier Trip Distance variability Net Payroll amount Age of Vehicles % of training done on time Gasoline Hauler Fatigue Score Hard Brakes Idle Time Monthly average speed Commercial Fleet Years Driving Experience Vehicle age Sales performance Zip code Predictive Data at Work Retention Service Failures Monthly payroll amount Take home percentage of pay Total miles prior 6 months Recruiting Avg. number prior employers Longest prior tenure Count of W/C claims

11. Real Results We love gain share opportunities! A Predictive Model’s Capabilities Accidents Severity TurnoverProductivity 23 % 81 % 49 % 12 %

12. The Target - Likelihood of drivers to have a worker comp claim in January Based on Decembers data. The Result – 3 out of the 8 worker comp claims were made by drivers identified in the worst 1%. A model built on 2 years of historical data: Workers Comp Summary

13. Change from: What happened? How many and where? What do we think the problem is? What do you think we should do about it? What’s the worst that can happen? Change to: Why is this happening? What does our data really say? What will happen next? How do we stop it from happening? What’s the best that can happen? Change Management

14. Intelligent Risk Management… That Drives Your Company Forward Thank You! Thank You! FleetRisk Advisors