Slide Show Presentation Presented by Maureen Stolberg, CIPM. Predict before you purchase. Using Prediction Models to Reduce Defective Inventory Levels. 

The Market Today. Defective cars often result when there are tampered odometers, mechanical issues the dealer is not able to address, issues with getting the vehicle title from the seller, or some other unforeseen problem. This is because many auctions do not allow the buyer to inspect the vehicle prior to purchase. Defective Vehicle Resale: 1.2 million Cars 3% Market Share Wholesale Auction Car Resale: 9.9 million Cars 25% market Share Used Car Resale Industry: 40 million Cars Total Market Share Based on these conditions, approximately 12% or 1.2 million vehicles sold at auto auctions are classified as “defective”. 

Why does this matter? Defective car purchases can be costly and hurt the bottom line. Defective cars are declared unsellable and must be sold at salvage value. The average loss for each defective car purchased is $3000.00 Total industry loss exceeds by more than $3.6 Billion Dollars. 

Being a car expert isn’t enough to identify defective cars. 10.0% 10.9% 10.9% 11.8% 11.9% 14.7% 15.5% 16.0% 27.3% 33.3% 35.5% Toyota Hyundai Honda Kia Mazda Subaru Mini Nissan Acura Infiniti Lexus Defective rates of US Consumer Report-J.D. Power’s Top Most Reliable Cars Lexus achieved the highest defective rate across all 31 brands, despite having remained at the top of JD Power’s Most Reliable list for 6 straight years. Lexus 

Cosmetic Changes can distract and mislead buyers. Many times the most important features for determining whether a car is defective are not available. 0.35 0.88 0.93 0.99 Vehicle Age Color unknown Transmission unknown Wheel Type unknown Variable Importance 

The Solution. Predict before you purchase. Auto dealerships can reduce the likelihood of making bad, costly purchases, at wholesale auctions by using machine learning methods to predict which cars have a higher risk of being classified as defective before an auction takes place. 1.2MM .2MM Traditional Purchases Data-Driven Purchases Reduce defective inventory levels by 87% 

The Solution. Predict before you purchase. Maximize inventory profit potential. Make data-driven purchase decisions. Know ahead of time which cars offer the highest profit potential and which cars pose the greatest risk for potential loss. $2,453 $2,050 $1,826 $1,627 $1,515 VOLVO SUBARU TOYOTA SCION HONDA $(520) $(896) $(1,307) $(1,964) $(2,326) PLYMOUTH OLDSMOBILE INFINITI MINI LEXUS AVERAGE PROFIT POTENTIAL AVERAGE LOSS POTENTIAL 

The Solution. Predict before you purchase. Maximize inventory profit potential. Avoid costly mistakes. Auto dealerships trying to provide the best inventory selection possible can avoid making bad purchases and lower the overall risk of passing on a defective car to their customers resulting in potential savings of $3.1B. $.4B $.7B $1.1B $1.4B $1.8B $2.2B $2.5B $2.9B $3.1B $.0B $.5B $1.0B $1.5B $2.0B $2.5B $3.0B $3.5B 10% 20% 30% 40% 50% 60% 70% 80% 87% MODEL PREDICTION ACCURACY INVENTORY LOSS REDUCTION Inventory Loss Reduction Savings