Exclusive: Check Out the Terrible State of US ISP Competition

We wouldn't need net neutrality if we had a free market in home internet. But exclusive data from Ookla Speedtest Intelligence shows that most Americans don't have anywhere near the choices they want.

Losing net neutrality means internet service providers will be able to charge content providers and favor their own content over those of competitors. For instance, Verizon could speed up its own Go90 video service while demanding payment from Vimeo, Vevo, and YouTube for the same quality of service.

The loss of net neutrality isn't likely to reduce your access to internet giants like Google, Facebook, or Amazon, which can afford to pony up to keep the bits flowing. But it could very well choke down smaller, newer content services, or fringe or independent information sources that can't pay the ISPs for access.

In a competitive ISP market, this wouldn't matter so much. If one ISP has a problem with the Daily Caller, for instance, fans of that site would be able to switch to an alternative ISP with a better deal.

But that's not the reality of competition in the US. Especially at 25Mbps, the FCC's definition of broadband, most Americans have two or fewer choices, Speedtest Intelligence data shows.

FCC chairman Ajit Pai and his allies would argue that slower broadband options, satellite, and LTE mobile carriers, should be considered as home competition as well. We disagree. The most intense competitive concerns right now come around video services, and to serve multiple high-def video streams to an American home right now, you need a median connection speed over 10Mbps.

Driven by that video consumption, the average American home also consumes more than 190GB of data per month. As 4G "unlimited" and satellite data plans de-prioritize your speeds at between 10GB-50GB depending on carrier, they can't replace a fast home internet connection.

Anything less than real competition at 25Mbps is sentencing Americans to second-class internet services.

Here's How Bad Competition Is

The Speedtest Intelligence data set shows results from 21,511 of the nation's 33,092 ZIP codes. Because it's a crowd-sourced data set, that reflects where people actually live in the US, as opposed to ZIPs that have no, or almost no residents.

Because the data was built from real speed tests, the data set also shows actual, rather than promised speeds. Data was gathered from connections to home ISPs from July 1 through December 13, 2017.

Taking speed out of the equation, only 30 percent of those ZIP codes have a choice of three or more ISPs at any speed. The situation gets even worse when you look at 25Mbps connections. Seventy percent of our ZIP codes showed either zero or one option for 25Mbps internet. (That's true at my house, where my only choice right now is Spectrum. Verizon Fios has been promised for years, but it hasn't yet arrived.)

Now, we thought some of that might be biased by the wide open spaces of rural America, which have a well-known broadband drought. So we looked again, at results from only the 5,000 most populous ZIP codes—the nation's major cities and suburbs.

Of those ZIPs, about 58 percent had a choice of three or more ISPs of any speed. But if you want a 25Mbps connection, only about 24 percent of ZIP codes had three or more options. The map below shows how few ZIPs in the country have three or more options at 25Mbps or greater.

I'm harping on having three options because that's the bare minimum to have any sort of competitive market. In wireless, for instance, the FCC and DOJ have repeatedly shot down mergers that would result in fewer than four nationwide carriers, for competitive reasons.

Broadband competition is also very unevenly distributed, even within cities. In Chicago, the city's North Side has three 25Mbps providers—RCN, Comcast Xfinity (the dominant cable company), and Webpass. Down on the South Side in ZIP code 60619, though, RCN and Webpass go away, and all you have is big cable.

On the west side of Houston in ZIP code 77084, you can jump off your cable company and join GigaMonster, but folks southeast of downtown in 77061 have no such competitive option—all they have is their cable company. Often, these lines follow economic divisions, with wealthier neighborhoods having more options. The map below shows which ZIPs in the Houston area have three or more ISP choices at 25Mbps.

The Speedtest numbers even slightly overestimate competition, because they were analyzed at the ZIP code level. Some ZIP codes are split between dominant cable providers, and those two providers would show up as "competing" in that ZIP, while they don't compete on a house-by-house basis. Similarly, if a competitive overbuilder such as RCN or Verizon Fios is available at one home running a test in a ZIP, it's listed as a competitor for the whole ZIP.

Why So Bad?

There's plenty of debate over the lack of US high-speed competition. In my judgment, it comes from two sources: the fact that it's very hard to build new infrastructure in the US, and the fact that regulators won't force the incumbent providers to share.

Building new "last mile" connections to homes is extremely expensive here. Some of that is just because running cable is expensive, but some of it is because America's system of powerful towns and villages with complex development regulations (often, anti-development regulations) makes building anything more expensive than it is in more centralized countries.

The Telecommunications Act of 1996 spurred competition by forcing existing copper-wire phone companies to wholesale their home connections to competitors, something called "local loop unbundling." During an era when dial-up, ISDN, and DSL over copper phone lines were the dominant means of internet access, that meant we had a lot of different ISPs to choose from.

But a 2004 court ruling and a series of FCC rulings severely cut down on local-loop unbundling, and prevented it from being extended to new technologies like cable and fiber internet. The FCC and the courts decided that fiber-to-the-home had so much profit potential that competitors would run multiple fiber lines to each home.

By and large, this hasn't happened. In fact, as competitive fiber providers have popped up, they've by and large dialed back their buildout plans after a few years, seeing buildout costs as too high to make decent profits from; just look at Google Fiber and Verizon Fios.

Without any help from regulators, the next chance for home ISP competition will come from 5G wireless, which may not have the capacity limits we've seen with 4G wireless. Verizon has said that its new 5G home service will be competitive with other home offerings, but it's coming slowly; only three to five cities next year, starting with Sacramento.

A free-market approach to internet access doesn't work if there's no free market. Without net neutrality, we're at the mercy of our monopoly ISPs, and it doesn't look like those monopolies are going to change in the next few years.

Editor's Note: Ookla is owned by PCMag parent company Ziff Davis.

This article originally appeared on PCMag.com.