The Problem: In late January, California's Employment Development Department, a state agency that provides job-placement services and disability insurance, mailed 63,000 tax statements that included Social Security numbers and income data—to the wrong addresses.
The misdirected statements, which represented 4% of the 1.5 million the agency sent out, were supposed to go to individuals who had received state unemployment, disability or family leave benefits but who had moved in the previous 18 months. The agency re-mailed statements when it discovered the error, and informed all the affected individuals in March.
The cause of the glitch: The software that synchronized the agency's mailing database with the U.S. Postal Service's address-update database incorrectly combined individuals' old street addresses with their new cities and ZIP codes, according to Kevin Callori, spokesman for the Employment Development Department.
As a result, most of the letters were either delivered to the wrong addresses or returned as undeliverable, although Callori estimates that up to 20,000 in the original mailing were correctly forwarded by the U.S. Postal Service.
The department is "implementing more extensive testing procedures" to ensure a similar error doesn't happen again, Callori says.
Key Lesson: Put in controls at strategic points to ensure data accuracy.
"When adding or appending new data to existing records, the fact that you are changing data demands that you have check routines in place," says Frank Dravis, vice president of product marketing at business intelligence software vendor Business Objects, which sells database-cleansing software. A check routine, in this context, is a piece of code that would automatically verify that a database merge has produced the expected results.
Do you have a tip about an information-systems snafu? Send it to todd_spangler@ziffdavis.com.
