As a stolen Bentley is loaded onto a tow truck in Georgia to be returned to its rightful owner, a months’ long investigation is coming to an end. But this story actually begins several months earlier near Chicago, at the desk of an analyst. While combing through vast amounts of vehicle related data, the analyst spots an outlier. A new high-end Bentley pops up in the vehicle title history information with no previous service or registration history, strange for a high-end vehicle not to be serviced regularly or have other records. After extensive research, a lead is sent to an NICB Special Agent which initiates the investigation that ultimately leads to the recovery of the $300,000 vehicle.
But it doesn’t stop there because the analyst also found a Range Rover in Illinois, a Mercedes-Benz in Florida, and another in New Jersey…all connected to the same organized theft ring. The National Insurance Crime Bureau (NICB) has been seeing an increase in Vehicle Identification Number (VIN) switching where a stolen vehicle’s VIN is changed to hide the vehicle’s true identity. Along with this increase has been a noticeable rise in organized vehicle theft ring activity. To combat this specific problem, NICB’s analysts have employed a number of different analytical techniques and tools but really, these techniques can be used in almost any analytical scenario to increase your odds of success.
1. Combine data to see a bigger picture
It’s rare, or at least more difficult, to uncover fraud in one stand-alone dataset. But when you start leveraging different datasets against each other, you can paint a bigger picture of what is going on. In the world of vehicle theft analysis, we are often combining data from vehicle manufacturers with title history information and public records data. Together, these datasets can reveal the full story of a vehicle and all its owners. This allows us to pinpoint when the theft or fraud occurred and even more importantly, it allows us to tie multiple stolen vehicles together. Combining these datasets has been instrumental in identifying organized rings ranging from small ones involving only a few vehicles in a single state, to large nationwide rings involving hundreds of vehicles. Taking it a step further, we also add in vehicle theft data whenever possible. This involves looking for active thefts that match the possible VIN switched vehicles we are looking at. By doing this, analysts have uncovered how the thieves are manipulating VINs to hide stolen vehicles and identified trends in where the vehicles are being stolen from.
It could be incorporating an individual’s previous address history from public records, claims data from other companies, or even adding a social network analysis into your data that leads you to a key connection not seen elsewhere. Whatever it is, try to think of new ways to combine the data you have available and look for new types of data you can add to the mix.
2. Master your software
Every organization utilizes different software tools from the way they store their data to the programs they use to perform their analysis. There is no “one size fits all” and no perfect software tool that will do all the work for you. What every analyst should do is become aware of all the tools they have at their disposal and become a master at them. If the only tool you have to manipulate data is Microsoft Excel, then learn everything there is to know; from the basic pivot tables to the more advanced features of Get & Transform, Power Pivot, and creating data connections. If you have more advanced analytical tools, make sure you understand what the different features are doing and how to make the best use of them. NICB’s analysts are continually exploring new methods with our software programs to look for quicker and easier ways to accomplish the tasks at hand and to test out new ideas to look at our data in different way. The more familiar you are with how your software programs work, the more creative you can be in using them to manipulate your data and highlight new findings. Even the best analytical minds won’t get very far if they can’t fully operate the software tools their data is in.
3. Use advanced analytical concepts and automation to save time
While the future of advanced analytics involves complex computer algorithms, predictive data models, and machine learning; not every fraud analyst is going to have access to the software (or a data scientist in their back pocket) in order to produce these. However, there are still some steps you can take to improve your analysis using the fundamentals of these concepts. In its most basic form, an algorithm is just a set of steps a computer follows. You give the computer your data as an input, the algorithm runs and gives you an output based on those steps. At NICB, we use algorithms to quickly pinpoint the most obvious fraud in our data. This ranges from filters that are automatically applied to every new dataset so we can instantly see the latest records with the most fraud indicators to more complex versions where we are combining numerous datasets and running calculations on the data to identify potential fraud. For vehicle theft and VIN switching analysis, this means we can automatically run these algorithms every time we get new data and get a targeted list of the VINs that are most likely VIN switched, so we can hit the ground running. The goal is to quickly find the most obvious fraud. To do that you need to figure out what indicators are the most important and then create a way to efficiently and automatically find that in your data. The best way to do this will vary greatly depending on your software but the idea is to identify the steps you use to find fraud and then automate those steps so you can quickly generate a targeted output to review for fraud.
4. Play time isn’t just for kids
Once you have employed automation to speed up your analysis of the highest priority fraud, you have freed up time to examine the more complex or less obvious fraud schemes. Giving analysts time to play around with the data, try new techniques, and look for new patterns is critical in staying on top of new fraud schemes. As insurance companies and law enforcement identify and crack down on different fraud schemes, the fraudsters find new ways to work around the system. NICB has seen this consistently in the vehicle theft and vehicle fraud arena over the years. They change the vehicles they target, the theft methods, the type of VIN switching, the states where they re-title and sell the vehicles. While we can use algorithms and automation to quickly identify known patterns, we can’t rely on them to catch everything. They are based on the past and we want to be working towards the future which is why it is vital that our analysts have time to look for new patterns and schemes. It is during this “play” time that our analysts have identified states that are seeing a sudden increase in VIN switching and found new patterns in how vehicles are stolen and re-titled. As the fraudsters pivot to new types of theft and fraud schemes, we must continually innovate our approach to stay at the forefront of fraud analysis. This is one of the most important elements of successful fraud analysis. The absolute best method of finding fraud today may be totally useless a few years from now as trends change and criminals adapt. Fraud analysts and their methodologies need to evolve in response to what is happening their data. Without the free time needed to explore and innovate, this is near impossible.
Whether you are new fraud analyst or a seasoned vet, the world of fraud schemes and fraud analysis is continually changing and we must continue to adapt and improve in response. Vehicle theft has seen an increase in the last few years, after steadily falling for years and VIN switching, vehicle fraud, and organized groups have made an even bigger comeback. But at the desks of NICB’s vehicle analysts in Chicago, we are fighting back using all the analytical tools at our disposal.
Jessica Rust is a Supervisory Analyst at the National Insurance Crime Bureau (NICB) in Chicago, IL where she oversees a team of analysts focused on vehicle-related insurance fraud. She has 15 years of experience in insurance fraud and vehicle theft analysis. Her primary area of expertise is VIN switching and organized vehicle theft and she has helped pioneer NICB’s efforts to proactively identify stolen vehicles and organized rings. Jessica’s work has led to over 20 million dollars in recovered vehicles and multiple theft rings being disbanded. In addition, Jessica helps instruct the NICB Analyst Academies which teach new analysts the techniques to identify and analyze fraud.