Big Data on Fraud
Whether it be security fraud, insurance fraud, billing fraud, or even unemployment fraud, big data and our ability to crunch big data is an important tool.
As so much is not automated anymore, we must begin to capture the data and see how it can be used.
Here are some examples of data slipping though our fingers:
IP login addresses
A recent case involved over 70 people applying for unemployment insurance from the same IP address and that was not even in the state from which the befits were sought
Address capture
Over 70 people had applied for car insurance and lived in the same 800 sq foot apartment.
Regional Analysis
Clients living in 22 Zip codes of the 41,758 zip code locations are where 14% of the insured losses occurred.
With big data you can generate smarter analytics:
- Spot emerging trends and hotspots to both jump-start antifraud efforts
- Prevent fraud at the time of engagement
- Predict and detect fraud at the intake of claims
- Identify fraud during the adjustment process
- Discover fraud patterns in all types of information
- Investigate fraud more efficiently by reducing false positives
- Concentrate investigations where the time will prove the most fruitful.
The insurance industry has accepted some level of fraud loss, about 10%, as a cost of doing business. What has happened is the fraud losses are no longer the lone individual opportunist. The ones that were paid off with $5,000 to go away. But like any real damage, the damage is not done by the larger storms such as hurricanes and tornados it is done by small persistent damage – like termites. The same is happening to the insurance industry time and time again.
Organized criminal gangs all outfitted with lawyers, doctors, experts and bad Samaritan witness at accidents are gnawing away at the industry. New scams such as low velocity impact injury, the phantom victim, and staged accidents are all on the rise. The claims are never very large by they are persistent. Big data can help stem the drain.
By collecting the names of witnesses, driver’s license information, telephone numbers, lawyers, doctors, insurance agents, and medical expert’s one can assemble a picture of fraud patterns. What is the likelihood that one witness saw 11 accidents? It is trillions to one. Why is it always the same lawyers, doctors and experts on the same types of cases – specialization or organized accidents?
Gather data on the insured and the victim as well as all of the witnesses is of the utmost importance if you are going to be able to mine data.
One defense attorney now demands access to all victims and witness’ social media and cell phone and computer records. As she said, it looks bad when the victim, the lawyer, a witness, and the medical doctors are already corresponding before the accident.
Most who implement big data mining and big data solutions see saving right away. Many in the neighborhood of a 40% to 70% reduction in fraud claims over a given industry average.
The process of the use of big data is: Identify, Prevent, Detect, Investigate then update the new “tells” of fraud to aid in the prevention and identification of fraud.
A Factoid:
80% of people do not see an accident – they here the crunch and only then see what happened after the impact.
Use big data to sport the outliers of fraud to protect your core business, its profitability and its competitive advantage.
