Man drives 6,000 miles to prove Uncle Sam's cellphone coverage maps are wrong – and, boy, did he manage it
Amazing how a big cash payout focuses the mind
A Vermont state employee drove 6,000 miles in six weeks to prove that the cellular coverage maps from the US government suck – and was wildly successful.
In fact not only did he prove conclusively that reports delivered to the Federal Communications Commission (FCC) by mobile operators aren't worth the paper they're printed on but also swung a spotlight on just how bad bureaucracy can get when it comes to Washington DC.
Corey Chase, a telecommunications infrastructure specialist who works for the Vermont Department of Public Service (PSD), undertook the monster road trip with some specialized equipment: six phones, each connected to a different mobile nework, and a custom piece of software, G-NetTrack, that carried out constant measurements of download speeds.
Everyone knows that the reports sent to the FCC on both broadband and cellphone internet speeds are awful, and are often carefully manipulated by operators to make it seem that Americans are getting excellent service when they aren't. But what sparked Vermont and Chase to go to so much trouble was, of course, money.
The FCC has set aside $4.5bn to ensure that 4G LTE speeds are available across the country and plans to divvy it up according to their coverage maps. But when Vermont took a look at the claimed coverage of its state, it was amazed to find that as far as the FCC was concerned 95 per cent of the state was receiving speeds of 5Mbps or higher. Just two swathes of territory accounting for 1,300 square km out of the state's total 25,000 square km were deemed eligible.
Officials knew that the reality on the ground was quite different and so decided to take advantage of the FCC's appeal process to try to get a bigger slice of the federal pie.
Vermont decided to undertake a statewide road trip to get a bigger picture. And what it built is the largest-scale accurate survey that has been published. And the results are not encouraging.
Rules, rules, rules
The FCC's challenge process was suitably bureaucratic: results had to be measured in a specific way and states were divided up into thousands of one-kilometer square blocks. To successfully challenge coverage in a block, a state is required to submit download speed tests that cover 75 per cent of that block, with a maximum radius of 400m.
In theory this means that each block could be split up into roughly nine (three by three) areas and a state would need to send results from six or more of these areas to be taken a valid challenge.
Of course in reality many of those areas are not easily accessible, especially in a state with large urban areas, hills and mountains etc. Vermont's Chase decided the best solution given limited resources was to basically drive every major road in the state at 40 miles per hour.
Why 40mph? Because the cellphone testing system in the back of his Prius ran a test sequence lasting 20 seconds, consisting of a 10-second download test, five-second ping test, and a five-second pause. At 40Mbph, that meant each test was conducted within 360 meters of one another – within the FCC rules.
The result of six weeks of driving and covering every corner of the state was 187,506 speed tests. And rather than finding that just 5 per cent of the state wasn't covered by a 5Mbps signal, Vermont discovered that an incredible 72 per cent of its results showed that its citizens were received slower than 5Mbps speeds.
Those results doesn't correlate exactly to geographic space but the state has published the results in the form of an interactive map that show much, much large gaps in mobile coverage than claimed in the official maps.
Because of the 75 per cent requirement however, Vermont notes that it is only able to directly challenge 789 of the 25,438 blocks that its state was sub-divided into – just 3 per cent. The reality will be much higher and Chase notes in his report [PDF] that the state wants to challenge a further 13 percent of the blocks. He also notes that he was unable to test in a remarkable 76 per cent of the blocks because they weren't accessible from the main roads he took.
As the report notes: "Meeting the 75 per cent threshold would require several tests at least 400 meters apart within each block. This could be met if the route went directly through the middle of a block. In most cases however, the main roads transect the blocks obliquely, that is, along a side or a corner. In these blocks, a drive test would not meet the 75 per cent requirement."
It goes on: "The PSD lacked time or budget to conduct a test thoroughly enough to meet the 75 per cent threshold on a wide basis. The PSD considered testing only small targeted areas with a goal of testing on side roads to meet the 75 per cent threshold in those areas."
America's broadband speed map is back! And it doesn't totally suck!READ MORE
"Ultimately, the PSD decided that the results of a drive test throughout the state could provide insight beyond the challenge process. Therefore, the PSD determined that it would conduct a drive test of all major roads (roads that receive federal aid) even though only a portion of the transected blocks would reach the 75 per cent threshold."
Or, in other words, the state decided that having good statewide data was more importantly than going along with the FCC's terrible process for challenging terrible map data. Or, put another way, the FCC knows its maps are pretty worthless and devised a system to improve them that is just as flawed as the maps in the first place. Your tax dollars going to good use.
Either way, Vermont's exercise should allow the state to challenge the FCC's figures and argue that rather than five per cent of its state not receiving decent mobile speeds, it should be funded for 15 per cent, while in reality the figure is much higher.
What it also indicates is that the most likely solution to figuring out the true situation of internet and mobile speeds across the United States is to incentivize local governments to do the work themselves.
No doubt the FCC will immediately start work on ensuring it has accurate data to work from. FCC chair Ajit Pai's former employer Verizon has made its claim of 98 per cent LTE coverage of the United States' population its main marketing message for years; we have no doubt it would love to update that figure to somewhere between 68 and 88 per cent. ®