Newly released transparency reports from the city of Menlo Park, California show that license plate reader data might not be as reliable as device manufacturers claim. The report casts doubt on the viability of a technology police and federal agencies spend millions of dollars each year procuring and maintaining.

In May 2014, the city council in Menlo Park passed an ordinance requiring regular reporting on the city’s use of license plate readers and the data they generate. The first transparency report is now public. It shows that between July 1 and October 1 2014, the license plate readers captured 263,430 images of ordinary people’s license plates. Out of those, only 141 provoked a “hit” in the system, notifying police that the car was on a hot list. But according to the report, “The vast majority of the hits were subsequently deemed to be a “false read” after further review by the ALPR operator.” In other words, police collected over a quarter million records of the movements of ordinary California residents, but the data was only relevant to a crime less than one one thousandth of a percent of the time. The transparency report only highlights one case in which the data was used to solve a crime, the recovery of a stolen car.

1 to 263,430 is not such a great ratio.

Across the country, government agencies and private companies are storing detailed records about our movements by collecting, sharing, and retaining location records captured by automatic plate readers attached to cars and affixed to traffic posts. In most states and cities, this technology is completely unregulated: Cops and companies can collect as much license plate location information as they want on anyone and everyone, and store it forever—doing god knows what with the sensitive records. But some cities and states, like Menlo Park, have implemented laws and regulations to limit data retention and require automated reporting on the use of the location tracking technology. Menlo Park’s first transparency report shows why these regulations are critical.

Absent public disclosure of surveillance technology purchases and the details of how police are using it, taxpayers have no way of judging whether or not it’s worth the money (and threats to liberty) to operate these costly technologies. The first report out of Menlo Park suggests it’s not. And that California city isn't alone. Over an 18 month period in Vermont, police captured 7.9 million license plate reader records. How many crimes did they solve with this location tracking dragnet? Fewer than five.