Top 25 Key Creative Datamining Queries for SIU Analysts

June 7, 2021 Frequent flier fraudsters spend a lot of time thinking of ways to get around your Claims Teams.  One of the most effective and efficient ways for SIU's to stay ahead of them is analytics - creating innovative datamining queries to bring the fraud to you for investigation. 

What is your favorite query? Which ones have the highest fraud impact?  Highest hit ratio?  Here are some common, interesting, and creative policy and claim queries to experiment with to bring in the meaty questionable cases for your investigators:

• Car to Driver Ratio - 4 cars and 1 driver on a single policy?  Your insured might be hiding youthful drivers or a commercial business. Be sure to check out any unusual car to driver combinations.
• Popular IP's - Multiple policies and/or claims from the same IP address often have suspicious origins.
• Midnight Munchies - Are you reviewing internet sales with limited data that occur between midnight and say, 4am?  Do that many people really buy insurance in the middle of the night while going for a snack? 
• Drop Box Dreams - Check your policy and claim addresses against drop boxes.  Often drop box / non-post office box addresses look like a street address but are really paid drop boxes for Underwriting or Claim fraud.
• Rate Jumping Ruckus - Watch out for varying addresses.  Is your policy rated for one state, but your loss and driver’s license are in another state?  May be some rate jumping, or PIP fraud here.
• New Vehicle Aficionado - Rotating new vehicles might be hiding a commercial business, such as a small car lot.
• Salvaged Vehicles - Previously salvaged vehicles often end up being used again in staged losses
• Tried and True New - Loss reported within x days of inception or reissue.  For certain within the first week to catch those 'fortune teller losses', but often up to and beyond 90 days.  Frequently flier fraudsters are getting smarter and pushing their loss dates out.
• Vicious Viaducts - Accidents in which a known or unknown vehicle pushes an alleged victim into a concrete barrier, such as a bridge viaduct, are a favorite for many large city rings.
• Frequent Fliers - Check out any combination of multiple claims on a policy, especially if they are similar in nature. A nice payout on the first one often entices follow-up claims, staged, caused, or otherwise. 
• Jinxed Neighborhoods - Rings will often target certain areas of the city for staging if they are free from cameras and witnesses.  You can get to this through Underwriting area market frequency spikes, and by running streets and addresses with large numbers of associated losses.
• Hail Sale - New car added to the policy followed closely by hail- be sure it was not a hail sale purchase from a car lot.
• Twin Vehicles - are you insuring the same VIN number on multiple policies or paying a total loss on the same VIN on multiple claims - positive indicator for fraud!
• Parked Unoccupied - This is an 'easy' way to cover for a negligent loss someone does not want to report. Look for an EDR here!
• Phantom Hit and Run - Another 'easy' way to either stage a loss, avoid reporting a negligent one, or create an injury claim.  EDR possibilities!
• UM Reborn as BI / a Phantom Twist - Watch out for losses that start out with a mysterious hit and run negligent party that can easily be defeated.  Frequent fliers are often more than ready to point the finger at your insured if the mysterious party story does not work out.  Remember the EDR here.
• Party Line Calls - Watch out for shared phone numbers between unrelated parties either current or tied together in past claims.  This is often hard to see with multiple interested parties and where i2ANB shines!  Insured and witness?  Claimant passenger and body shop?  A solid indicator for a potential ring, so be sure to add these 'party line' numbers to your future datamining queries.
• Ghost Provider - Check those new provider addresses against Google Maps and a field visit to be sure they exist.
• CPT Blues - Watch for rapidly rising CPT codes in your billing data as an indicator for abuse. Then get ahead of those by pulling them in advance for review.
• Party Bus - Most bus lines, commercial, government, and private, have internal cameras.  You can get ahead of jump-ins by proactively requesting camera footage.
• Overflowing Vehicles - Vehicles loaded down with BI or UM injured passengers are often provide a solid lead for potential fraud.
• Downtown Shoppers - Many areas have camera footage downtown, be it government or local businesses, that capture pedestrian losses. Many businesses will allow you to view the camera footage, and FOIA requests usually work with governments.  This can not only circumvent fraud, but also give you a clearer picture of liability and injury quickly on legitimate claims.
• Humdrum Minimal Impact Crash - This query is a common one but be sure to include it in your arsenal.  Injury cases with minimal impact often contain abuse of some form, either via the customer or the provider.
• Instant Replays - Be sure to stay ahead of 2nd, 3rd, and 4th showings from proven fraudsters in your claim and policy books by developing a POI list.  Identity resolution alerts are critical here as frequent fliers often try to mask their identities.  
• Postmortem Portrait - Don't forget to postmortem your organized fraud cases as a part of your wrap-up.  Looking backwards via i2 Analyst Notebook will uncover the trending indicators that can be turned into new queries.

These are just some of the combinations you can use to get ahead of the fraud through analytics.  To keep ahead of the fraud, we must continue to 'outthink' the competition > fraudsters.  Be creative.  And keep in mind there is nothing wrong with trying a query that does not work - learn from it, tweak it, put it on the shelf to retest down the road, or retire it and try something else.  Fraud Analytics is all about learning, trending, and following the data to get ahead of the fraud.


And now for my mini-commercial to close - At Globlue, fraud analytics is our passion.  If you are interested in putting enhanced datamining queries into play, including advanced identity resolution capabilities, through a streamlined easy-to-manage process, reach out so we can show you what we can do for your team!
June 7, 2021