FCC Issues New Broadband Stats

The FCC has issued a report (pdf) that claims that 78 percent of adults in the U.S. access the Internet and that 67 percent have broadband at home. Before you go jumping up-and-down in excitement about these adoption rates, there are serious caveats and questions.

The biggest caveat is how the FCC defined broadband:

For the purposes of this report, home broadband users are those who said they used any one of the following technologies to access the internet from home: cable modem, a DSL-enabled phone line, fixed wireless, satellite, a mobile broadband wireless connection for your computer or cell phone, fiber optic, T-1 (p 3).

Whoa, Nellie! That’s a definition big enough to drive a semi-truck through. It’s the “if it’s not dial-up, then it must be broadband” definition. The FCC definition of broadband is so, well, broad because of teleco lobbying. In the context of this report, do not confuse broadband with “really fast.”

There’s also an issue with how well  folks know what type of ‘broadband’ service they are using. More on that later, first let’s focus on why broadband doesn’t necessarily mean “fast.”

In September 2009, Reuters reported (emphasis added):

A 2008 study by the Organization for Economic Co-operation and Development showed that the United States ranked 19th with an advertised rate of 9.6 megabits per second (mbps). The top three countries were Japan with 92.8 mbps, Korea with 80.8 mbps and France with 51 mbps.

The U.S. has no broadband system for residential adoption today that comes close to the speeds that Japan, Korea or France had in 2008.

Earlier, in July, John Shepler of T1Rex.com had warned:

If you’ve been eagerly anticipating FTTF (Fiber to the Farmstead), you may find the definition of broadband a bit disheartening. The bar has been set at a breathtaking speed of 768 Kbps download and 200 Kbps upload to meet the criteria of “broadband” Internet service.

Let’s do do the math.

• One kilobit per second (Kbps or kbps) is 1,000 bits per second (bps)
• One megabit per second (Mbps) is 1,000 Kbps or 1,o00,000 bps
• One gigabit per second (Gbps) is 1,000 Mbps or 1,000,000 Kbps or 1,000,000,000 bps

The 2008 data showed that the average U.S. broadband connection was about an order of magnitude  slower than Japan (1/10th as fast). The 2009 FTTF broadband definition is yet another order of magnitude slower: 1/100th as fast as Japan’s average advertised broadband in 2008.

In the public comment period, Free Press advocated defining broadband as being a minimum upstream and downstream speed of at least 5 mbps, half the average US advertised speed in 2008. However, the telecos argued that this bar was too high because, in part, of our rural (low-density) swath between the Rockies and the Appalachians.

This FCC report begs the question: what are ‘average’ speeds for the plethora of technologies the report counts as “broadband”?

Technology	Speed
Cable	Basic: 4 Mbps to 6 Mbps High End: 12 Mbps to 16 Mbps and faster
DSL	Basic: 768 Kbps to 1.5 Mbps High End: 3 Mbps to 7 Mbps
Fiber Optic Cable	15Mbps – 25 Mbps
Mobile – EDGE	Up to 58Kbps, average 22Kbps
Mobile – 3G	AT&T: Download, 700-1.7 Mbps; Upload, 500 Kbps – 1.2 Mbp Sprint: Download, 600Kbps – 1.4 Mbps Verizon: 600 Kbps to 1.4Mbps
Mobile – 4G	Download: 3-6 Mbps
Satellite	10 – 20kbps
WiMax (like Clear)	Download: 3-6 Mbps
South Korea	1 Gbps (2012)
Japan	Average advertised: 93.6 Mbps (2007)
France	Average advertised: 44.1 Mbps (2007)
Data: ArsTechnica: FiOS speeds, AT&T Cell Phone Coverage, CLEAR WiMax 4gG, GigaOm: Korea, High Speed Internet Access Guide: DSL v Cable, ModMyi.com Forum (EDGE), Sprint 4G, Verizon 3G Basics, Website Optimization: France, Japan, Wikipedia: Satellite Internet Access

These data ignore issues of latency (satellite, mobile), privacy (cable) and line-sharing (cable). The table also ignores price, both the difference in costs for the various technologies in the U.S. and the vast difference in pricing plans when the U.S. is compared with Asia and Europe.

The second caveat is this: read the data on  how people access from home with a large dose of skepticism.

Why would the survey be designed to accept more than one answer to the question: “At home, what do you now use to connect to the Internet?” I do not believe that most people have multiple “broadband” access plans in order to provide for redundancy. However, that’s what this chart (p 14) leads you to believe. The total should add to 100 percent; instead, it adds to 212 percent.

Thus, the question appears very poorly worded. The question should have been what is your “primary” connection followed up by “secondary methods.” When we had DSL, for example, we also had a dial-up number that we could use if we were (a) traveling or (b) the DSL died. [The firm elaborates on its followup questioning technique.]

Another cause for pause.

Forrester estimated in November (p 21) that 17 percent of the U.S. population has a smart phone. But 28 percent of the respondents in the FCC survey claim to be accessing web pages from their cellphones and 26 claim to access email (p 22). The report authors then claim that 30 percent of American adults have smart phones, according to their survey. I find it hard to believe that Forrester — an analytical firm with a pro-business bent — would underestimate smart phone adoption by almost 50 percent.

The fourth caveat is actually a question.

I cannot figure out how the analysts came up with the “78 percent  of American adults access the internet” soundbite.

The 22 percent data point is on page 24 as a partial explanation of why one-third of Americans do not use broadband at home. However, the methodology section does not provide a list of the questions asked of those surveyed. For those critical of data tweaking I refer you to Table 2 on page 48, which shows how the demographics of those responding had to be “modified” to reflect the national demographic on age, gender, education, race.

Not surprising: the correlation between age, education and income and ‘broadband’ adoption (p 13). I say “not surprising” because these factors have been correlated with Internet use from the git-go.

Also not surprising: the data suggest that people who own a mobile phone are more likely to have some form of non-dial-up access to the Internet (p 47).