Fraud Detection

Porsche Westlake’s Services Director, Sam Abregel, addresses auto warranty fraud

By October 10, 2019 No Comments

10 categories of warranty fraud to watch out for.

Sam says, all the warranty repairs are normally categorized into about 10 different categories. The final price charged varies between dealerships as the prices of the service and parts are marked up by different factors by each dealership. However, each of the service categories has a standard labor time allocation, and this categorization of warranty repairs is standard across all the automobile retailers. The major service categories, and some of the most common types of fraud seen in each area are:

  • Missing Steps In The Service – The 10 standard steps of warranty repair analysis and rectification are devised to mitigate fraudulent activity. If the technician bypasses some of the steps while performing the task, it usually indicates some fraudulent intent.
  • Less Mileage Vehicle – Sometimes the technicians find a car with less mileage and uses that car for making false warranty claims.
  • Email Changes – Usually if there is a frequent change of email ids that is an indicator of unusual, and potentially fraudulent activity. It is a problem that needs to be looked at to ensure that it’s not a part of some scam.
  • Frequent Repairs – Every store and region has a profile that they maintain. If there are frequent repairs in a vehicle it may indicate that there is a problem. The standard profiles can help in identifying how frequent and necessary were the repairs that were ordered.
  • Repetitive repairs – If a repair is often repeated and very frequently the same repair is done, it is most likely a part of a fraud scheme. These may be identified each system has a part failure rate (E.g. – Engine, brakes, etc.), which can help in identifying anomalies.
  • Quick Fixes – Indexes that show the necessary time for certain repairs help to identify outliers. Some technicians find a short cut to finish a repair work, so that they can pocket the difference. Thus if a repair falls outside the repair time index (positive or negative), it raises a red flag that may indicate fraud.
  • High Staff Turnover – Staff turnover is another indicator of fraudulent activity. There is a need for reliability for the repair and warranty work. If the staff keeps on moving around, finding the person accountable for fraud is difficult.
  • Unnecessary Repairs – There are repairs ordered and done which are not needed as a part of the current service.

Huge opportunity for Machine Learning

Given the vast amount of transaction data and the presence of anomalous patterns that indicate fraudulent activity as outlined above, there is a huge opportunity to harness predictive analytics and machine learning models to identify and address potential fraud and anomaly issues before they have the chance to harm your business.  Machine learning models are ideally suited to:

  • IDENTIFY suspicious patterns across multiple, disparate groups
  • Provide clear INSIGHTS about the nature of the anomalies
  • Score and prioritize potential SUSPICIOUS behaviors to go after highest-yield cases first
  • Support case investigation to take ACTION & STOP fraud schemes
  • LEARN from case resolution patterns and continuously improve detection.

From its inception 120 years ago the Retail Automotive industry has faced many types of risks. But a constant challenge is a small group of unethical dealers, or more so specifically, the technicians committing service and maintenance fraud.  According to NADA Data (2018), dealerships wrote more than 310 million repair orders, with service and parts sales totaling more than $116 billion across the 16,753 franchised new car dealerships in America. With this humongous amount of service transactions, there is a commensurate challenge to identify fraudulent activities, and therefore any automated solution to overcome this challenge would be very welcome.

X