Five years ago, Jamie Barnett took his boss on a tour of a 911 call center in Fairfax, Virginia, a 20-mile drive from his office at the Federal Communications Commission in Washington, D.C. At the time, Barnett was the chief of the FCC’s Public Safety and Homeland Security Bureau, the division that oversees, among other things, the operation of the national emergency number. The purpose of the trip was to give the FCC’s then-chairman, Julius Genachowski, a firsthand look at a persistent problem with 911: The system couldn’t accurately trace the locations of emergency calls made from cellphones.
At the call center, Barnett recalls, his boss put on a headset and observed a few calls, then pulled out his cellphone to find out what would happen if he called 911 from within the call center itself. An operator answered and requested permission to locate Genachowski using the GPS on his phone.
When his location popped up, it showed that Genachowski was at a Costco a half-mile away.
On police TV dramas, cellphones allow cops to do seemingly miraculous things. GPS can track kidnapping victims from a cellphone stashed in the trunk of a moving car, or pinpoint a call to a single creepy basement. But in reality, figuring out the exact location of a cellphone—and accurately transmitting that location to an operator—is nearly impossible.
In most cases, that doesn’t make a difference. Most people who call 911 know where they are, and can communicate it clearly to the operator. But every so often callers don’t know where they’re located, or they’re in a situation where they can’t communicate their location out loud, forcing operators to spend precious seconds or minutes figuring out where they are.
“It’s costing you time,” says Barnett, now the director of Find Me 911, an organization set up to lobby the FCC to adapt better location technology. “You go there and you say: Where is this person? I can’t see them; I can’t find them.”
During the first half of the 20th century, if you wanted to make an emergency call, it was routed through a switchboard operator, waiting its turn to be placed alongside all the ordinary calls to relatives or the neighborhood pharmacy. Once the emergency call went through, the operator would connect the caller to the police or fire station nearest to their home—a cumbersome job that could take considerable time. (New York City alone currently has 77 different police precincts.)
It wasn’t until 1957 that the National Association of Fire Chiefs first suggested a separate number to speed up emergency-response times. Around a decade later, the Federal Communications Commission finally created our current system. Throughout the 1970s, specialized call centers cropped up across the county to dispatch first responders to emergencies.
The system worked so well because it was so simple—with just two numbers and three digits, “911” was easy to remember and dial—and because it could precisely locate callers based on their phone numbers.
Tracing a landline call to an exact location doesn’t require complex technology. The 911 call centers have databases of every landline phone number in their area, and the street address associated with each one. In the earlier days of 911, locating a caller was as easy as drawing up a document and running a basic search.
Cellphones made the process much more complicated. Like landlines, each cellphone has an owner and a billing address—but that address tells you virtually nothing about where the call is coming from. In the early ‘90s, with more and more emergency calls coming in from cellphones, it became clear than 911 call centers were going to need advanced, location-based technology to find people.
In 1992, the FCC created a committee to figure out a solution. But the committee members failed to predict just how ubiquitous cellphones would become. Brian Fontes, the CEO of the National Emergency Number Association (NENA), says that most of the early cellphone-tracking methods—most of which are still in use today—focused on locating people outside, with the assumption that indoor calls would come from a landline.
According to a recent report from the National Center for Health Statistics, 45 percent of U.S. households now have cellphones but not landlines, a number that’s grown steadily for more than a decade. But emergency-call centers haven’t kept pace. “There’s been a lot of advances in technology, but it hasn't advanced to the point where they can actually determine where someone is,” says Fontes.
Calls placed on cellphones bounce off of the nearest cellphone tower, which directs them to the closest emergency-call center. But calls can be intercepted in unpredictable ways. When a lot of calls are being made at once, there’s a greater chance that the towers become overwhelmed, meaning some calls will be picked up by a tower farther away.
In an emergency situation, this can be disastrous. Last December, for example, a woman named Shanell Anderson lost control of her car, which plunged into a pond in a town north of Atlanta. When Anderson dialed 911 from her cellphone, the call was directed to the wrong dispatch center. When she described her location to the operator, it was out of the call center’s district, and the operator couldn’t immediately locate it on a map. It took 20 minutes for an ambulance to arrive; by that point, Anderson was unconscious. She died 12 days later.
Even assuming a call reaches the correct phone center, it doesn’t mean that an operator can quickly trace it to the right location. That’s because many cellphone carriers rely on triangulation, which locates callers by measuring the distance between the towers that pick up the call.
Triangulation is accurate to a point, but it’s not perfectly precise. The process can generally pinpoint someone’s location within 100 to 300 square meters. This can be enough to help a dispatcher locate someone in, say, a small town or rural area. “If there’s nothing but two trees and a coyote in sight, you’re doing pretty good,” says Roger Hixson, the technical director of NENA. But if you’re calling from a dense city block—or from an apartment building, with hundreds of residents and potential emergency situations—the 300 square meters start looking dauntingly large. That circle can get police officers to the right neighborhood. But finding the right house, let alone right floor, is much more difficult.
GPS technology can help, but it’s at its most accurate when a cellphone has the location-tracking system turned on for hours on end, allowing satellites more time to calibrate where the phone is. And besides, 911 has to operate accurately not just for the next generation iPhone users of the world, but for the millions of Americans who don’t own a GPS-enabled smartphone.
Hixson believes that GPS isn’t quite accurate enough to gage location in an emergency. “You wanna get it right,” he says. “If you don’t quite get the right information for where the pizza shop is, no one dies. For 911, people die.”
Reformers like Hixson are putting pressure on the FCC to update the federal regulations for emergency calls—first issued in 1996, and last revised in 2010—to account for the realities of modern phone use. For decades, there were few mechanisms in place to make sure cellphone carriers adhered to FCC emergency guidelines, says Barnett. “The providers just said, ‘Yes, we are compliant.’”
After more than a year of debate, the agency passed new rules in January requiring that cellphone carriers correctly identify 40 percent of 911 callers’ locations. But the rules gives providers several years to phase in the change, and a longer timeline to implement the “vertical location” requirement for finding callers on a specific floor in a multi-story building.
While the emergency-dispatch community waits for the federal laws to catch up, some states and companies are working on ways to find cellphone users who can’t transmit their exact location. For callers who can’t speak to a 911 dispatcher, some operators deploy a series of yes-or-no questions that can be answered by taping the phone. (Certain states, like Massachusetts, have this built into their emergency-call systems, allowing a number that’s pressed on a phone to pop up on operator’s dashboard.) Wi-fi access points, which locate a phone based on all of the wireless networks in the near vicinity, also seems promising. Some companies are looking at software that can allow 911 operators to automatically access a phone’s GPS. Another seemingly simple fix—like sending location-based text messages—has potential, too, but texts become complicated when routed through the 911 system.
Other potential solutions come from the private sector, and the cadre of tech startups like SafeTrek, Guardly, and Panic Guard that allow users to call for help directly from an app. But many of these new applications also aren’t compatible with the technology already in place at call centers, and each has its pitfalls.
“Some of those applications are a bit scary,” says James Lake, who manages emergency call centers in Charleston, South Carolina. “They may alert five or six people, who may in turn dial 911 and say, ‘I just got an alert from my best friend’s phone.’ You ask, ‘OK, where’s your best friend?’ ‘Well, I don’t know.’” In those cases, no one’s any more knowledgeable than the dispatcher, who’s still in the dark.