Who Has Fast Broadband?

The Department of Commerce issued the following graph recently as part of a report titled Competition Among US Broadband Service Providers. The graph is a little hard to read, so I include it here for easier viewing. The purpose of the graph is to show that there is not very much competition for data speeds of 25 Mbps and higher. The graph shows the number of households that are able to buy various speeds of broadband from a low of 3 Mbps to a high of 1 Gbps. It does a good job of demonstrating the story that the Commerce Department wants to tell – which is that for most markets there is only one provider of fast Internet.

But there is another story told by this graph that nobody seems to want to talk about, which is how the government is doing a really lousy job of counting the number of households that still can’t get the slower speeds of Internet access. This graph shows, for example that only two million US households can’t get 3 Mbps download speeds and that 6 million can’t get 10 Mbps. And those numbers are total bosh.

These numbers are based upon the database that supports the National Broadband Map, and frankly, for rural America, this data is a farce. The problem with the map is that the data speeds are self-reported by the telcos and cable companies and supposedly represents the speeds that the carriers advertise in various markets, not the speeds that people can actually get. Further, the speeds are counted by census block, and if only a few homes in a block can get a certain speed then it is assumed that most can. This blog is too short of a forum to go into a detailed discussion of why the sampling method behind the map is a misuse of statistics, which would require a math-based whitepaper, and maybe one of these days I’ll take the time to write it.

Consider the actual speeds that can be achieved with DSL, where the speed drops rapidly with distance from the DSL hub, called a DSLAM. On very good copper DSL can deliver 10 Mbps of speed for about 7,000 feet. That’s not 7,000 feet as the crow flies, but 7,000 feet of pole lines. In most places that is not a whole lot more than a mile. Good copper can deliver 3 Mbps up to about 13,000 feet, which is almost twice as far as 10 Mbps but which only still equates to maybe 2.5 miles. And the sad reality is that most rural copper is not very good and so actual distances that can be achieved are going to be less than these theoretical distances. It’s well known that the largest telcos like AT&T, Verizon and Qwest neglected rural copper for decades and their copper in many places is a disaster.

Further, a lot of rural towns are still equipped with older DSL technology that tops off at a capacity of between 1 Mbps and 6 Mbps. In those communities even the people in town can’t get 10 Mbps as is suggested by the National Broadband Map. The reality is that the large telcos generally advertise the same products everywhere in a region. So small towns will see advertisements that say something like “speeds up to 20 Mbps, with fine print at the bottom of the ad that will say “where available”. Thus the large companies are not lying when they say they advertise fast speeds in rural areas, but the whole point of the Map is supposed to be to measure the broadband that people can buy.

What is even worse is that the large telcos will sell you DSL that barely works. I’ve heard of rural customers who buy DSL with speeds as low as 144 kbps, just a little faster than dial-up. At peak times some of these people revert to dial-up which is faster. And they pay the same price as a customer in the nearby town who might have 6 Mbps. Households with these slow speeds are undoubtedly shown on the Broadband Map as able to buy broadband.

Cable companies have a more definable situation. Cable ends where the coaxial cable ends and if you live one home beyond the last connection you can’t get cable modem. Almost all of the faster speeds on this chart are from urban and suburban cable modems. Many of those systems have been upgraded to deliver over 100 Mbps, and in competitive places like Austin TX they can deliver over 300 Mbps. But rural cable systems share a problem with rural telcos in that many of the systems are old. Some older cable systems don’t yet offer cable modems. And many other older systems can’t deliver speeds greater than 6 Mbps, making them very similar to rural copper systems.

So this graph works for the story it wants to tell, which is that there is often only one provider in a market that will offer speeds of 25 Mbps and higher. But the government really needs to stop publishing statistics  that make it look like most of the country already can get speeds of 10 Mbps, because it simply is not true in much of rural America. I get a little angry every time I see the government make pronouncements based upon this bad data. It feels like a cover-up with the government and the FCC together to deny how poor our broadband is. I wish it was true that  only two million people can’t get 3 Mbps and only 6 million can’t get 10 Mbps, because we wouldn’t have as far to go to improve.