Lightbridge Inc. recently activated two new weapons for fighting subscription and true identity cellular fraud. Fraud Sentinel is the complete package including Postalpro, Fraud Detect and the company’s hallmark product, ProFile.
Carriers “were clamoring for analysis on the front-end,” said Michelle Wheeler, product manager for the Waltham, Mass., company. Postalpro and Fraud Detect are “data scrubbers” said Wheeler, that weed out fraudulent or high-risk customers before they are signed up. ProFile is an inter-carrier system that detects if an applicant has ever been shut-off-when the carrier disables an account past due-or written-off. Carriers can chose any combination of the products and within each can customize around the needs of its markets.
Wheeler said using the Sentinel package reduces the number of credit checks needed, which is important as customer acquisition costs have increased and per subscriber revenue has dropped.
Postalpro interprets U.S. Postal Service data to see if an address is non-existent or incorrect, on the basis of street, city, state and zip code.
In billing and direct mail efforts, Postalpro also can save carriers money by preventing mailings to incorrect or false addresses, a huge problem for carriers today. Wheeler said of 8,500 bills mailed in a week, about 1,000 are returned. On average that costs the carrier $5 per bill in mailing costs.
Fraud Detect is the second layer of identity verification and has a hit rate-the number of applications found questionable-of 10 percent to 15 percent. Developed by Trans Union credit bureau, Fraud Detect is licensed to Lightbridge for resale, an agreement which is exclusive for the next six months. Using U.S. Census data and telephone company data on each tower’s radius of coverage, Fraud Detect compares components of applicant information. Algorithms unscramble the data.
To verify an applicant’s social security number, Fraud Detect checks it against the Social Security Administration’s death master file, matches the applicant’s said date of birth against the date of birth on file for that number, alerts to recently issued numbers and flags on fake numbers printed by manufacturers on wallet cards.
Area codes are a substantial trigger for subscription fraud, said Wheeler. The applicant may have all other information about a legitimate user, except the correct phone number. Fraud Detect matches an area code against the state, the exchange numbers against the city, determines a pager or cellular phone number and flags when a number is not in service. The program cannot, however, flag unpublished numbers. Fraud Detect flags on prison and military base zip codes, the latter where more detailed address information is needed for billing. Comparisons of distance between zip code and phone, employer phone to home zip code and employer phone to employer zip code are completed almost instantly by Fraud Detect. The system does crosschecks on driver’s license information as well.
In a Fraud Detect case study, using a base of 500 applications, Lightbridge’s ProFile found 261 were reported as shut-offs or write-offs by other carriers. Those 261 applications were run through Fraud Detect, which found fraud matches with 39 percent. This is reason, said Wheeler, for carriers to deploy Fraud Detect on the front end.
ProFile detects in real-time if an applicant is a high-risk subscriber or a thief. Each carrier can specify the dollar amount at which to reject an applicant reported as a shut-off or write-off. If a previous write-off is only $25, the applicant may be accepted.
If a write-off customer applies for service with a new carrier, Lightbridge notifies the previous carrier of the applicant’s current address and other information. Lightbridge also notifies a carrier that a customer recently signed on, has since been determined a write-off somewhere else. In this case carriers can limit usage and roaming. Wheeler said more savvy criminals know their window of opportunity-the time between a balance being overdue and when he or she is reported as a shut-off or write-off. ProFile detects if an applicant’s name and address information is similar to stored data on write-offs and fraudulent users, even if a few characters vary or are transposed.
