Why is it so incredibly difficult to figure out how much of the nation has access to affordable broadband? For the past 15 years, the Federal Communications Commission has been required by law to collect data on high-speed Internet access. For the most part that information has been fairly useless for the public or even for policy types. Up until recently, for example, a telecom company only needed to serve one customer in a ZIP code to get credit for serving everyone. Even so, good luck figuring out which company it was; the FCC's report scrubbed the names of the actual providers. In essence, the FCC was going out of its way to prevent useful information from being publicly released. Attempts by public interest and consumer advocates to get the FCC to release this information repeatedly failed.
So in 2009, the stimulus act allotted an additional $350 million to give the FCC another chance to gather this information and present it in an understandable way. That project, the National Broadband Map, was released last February, and it's a least an improvement over the old effort. But it didn't require all that money to make this happen. All the FCC really had to do to produce virtually the same map was ask better questions in its existing survey of telecom companies and release more of the data. Instead, taxpayers had to fork out another few hundred million in grants to collect much of the same data a second time.
Fortunately, there are ways to collect even better information that are much cheaper and don't require bureaucratic soul-searching. At the New America Foundation, where we both work, we've partnered with the Planet Lab Consortium and Google to offer a tool that lets anyone measure the performance of his or her broadband connection, called Measurement Lab. To date, people have run this test more than a half-billion times, giving us 300 Terabytes of data to work with. (That's more than 150,000 iPod shuffles.) We make all this data public, allowing people to see actual broadband speeds and compare them among countries, U.S. states, and cities.
There are a few reasons why this approach is superior to the FCC's efforts. While the National Broadband Map does offer a snapshot in time of where broadband is and where it isn't, you have to take what you see with a big grain of salt. In much same way that the coverage maps provided by cell phone companies are notoriously inaccurate, the map's information often fails to reflect on-the-ground realities. For example, the FCC reported in its National Broadband Plan that only 4 percent of the country has access to more than two landline-based broadband providers. But users who search the map for broadband service providers at their home address may be surprised find that they potentially have dozens of options. You get these bogus results for two reasons. First, the map relies heavily on self-reported data by the providers, who often paint their coverage areas with a broad brush. Second, it merges residential and business class services, so even if no residential options are available, your area may be considered covered. Together, these shortcomings give the impression of a market filled with lots of choices for consumers.
What's worse is what the map excludes. The information you get provides the name of the provider and a range of their fastest advertised speed offerings. But anyone who's ever checked their connection speed knows that real-life speeds tend to be substantially lower than what you're paying for. In the U.K., for example, Ofcom—the equivalent of the FCC—found that actual speeds were systematically half the advertised speeds. Even more puzzlingly, the map offers no information about price. According to the FCC's own research, cost is cited among the top issues for not adopting broadband at home. Thus, one would expect the National Broadband Map to provide this critically important piece of information. It doesn't.
To its credit, the FCC incorporated M-Lab into its consumer broadband test, but it has yet to release the data it has collected to the public. Meanwhile, telecommunications companies continue lobbying the FCC against releasing accurate information to consumers about where their services are available, how much they cost, and what are their actual speeds.
The FCC and the National Telecommunications and Information Administration could fairly easily fix all the problems with the map listed here. M-Lab has offered to donate tens of millions of actual speed measurements from throughout the country to help get this process started and is working to create an overlay for the National Broadband Map to do exactly this.
There was once considerable optimism surrounding the National Broadband Map. We think that with a few vital improvements, the map could easily become an exemplar of government data transparency as well as an incredibly useful tool for U.S. residents and policymakers. But without these improvements, the National Broadband Map runs the risk of becoming a $350 million boondoggle—a map to nowhere filled with inaccurate and useless information.