Fast Net News, DSL Prime, DOCSIS Report

In 2007, 96% of homes in California could get broadband according to a careful study by the state Broadband Task Force.  That's 4% unserved. In 2010, according to the NTIA broadband map over 8% were unserved.

One of the studies is seriously in error.  The official FCC data indicates a slight increase in coverage. Reviewing the public data, I became almost certain the 2007 data were on target and the NTIA/Broad Map data incorrect. 

At right: The graph from the 2007 California Broadband Task Force. They worked from the best available data and concluded all but 6% of California could get 5 megabits and all but 4 % could get broadband in 2007. There's been additional deployment since then and improved speed in the slower areas. 

At left: The data for California according to the National Broadband Map, which claims that only 83.2% of Californians can get cable modems. The FCC reports 99% of Californians who can get cable TV can get modems as well, so the NTIA/National Map data implies fewer than 85% of California can get cable TV. I marked this as a likely error. Nationally, over 95% of homes can get cable service. It's prohibitively unlikely that California has three times as many homes unserved by cable TV as the national average.  California is more urban and wealthier than the national average, both attributes making it likely California has at least as much cable coverage as most states.

    When two presumably reliable sources differ, the best thing to do is to go to primary sources. That's what NTIA presumably is doing now that reporters around the country are pointing out the data errors. I compared the California NTIA data to the State Broadband Task Force, FCC data, the National Broadband Plan estimates, the Columbia CITI report to the broadband plan, and to the financial reporting of the companies involved.  I'm touching base with the cable associations as well. The California Cable Association, with representatives of all the major cablecos in the state, reports 96% of homes can get cable modems. 

    Troy Wolverton in the San Jose Mercury, one of the nation's best tech reporters, writes "some of the data is simply wrong." Reporters in other states are finding similar, and the private buzz in D.C. is NTIA totally blew it. So far, WSJ and WP, the papers Washington reads, haven't picked up the story but it's almost inevitable they will. When they do, they find even more that went wrong. The rumor is that the big carriers sabotaged the work by refusing to provide the prequal data for each address, which is the only truly accurate data. NTIA and the states then filled in the information using some rules of thumb that proved wildly inaccurate and some tract and area data. In some cases they simply left entire regions out because they didn't get answers from the carriers. 

    A small part of the discrepancy may be because the NTIA set a cut off of 3 megabits for their 8.3% figure. The 2007 data showed that all but 6% could get 5 megabits while another 2% had speeds slower than 5 megabits. Since that time, DSL speeds in particular have improved because of DSM and other technical improvements as well as U-Verse. Both FCC data and news reports indicate that some areas not covered in 2007 have since been reached. If the 2007 data were updated to 2010 and adjusted to the 3 megabit threshold, the "unserved" at 3 megabits would almost certainly be less than 6% and probably 3-4%.

    Here's some of the original California Broadband Task Force report. I'm told that Larry Smarr and Michael Byrne deserve a great deal of credit. Anne Neville was the administrator in California when they discovered and solved most of the problems. Neville also led the NTIA project, where apparently political decisions from above created the issues. 

"Through analysis of CBTF’s broadband mapping project and independent research, the Task Force determined that California is better positioned than most states on broadband availability and adoption, yet the state lags behind key foreign competitors. Specifically, the CBTF found:

• 96% of California residences have access to broadband.

• 1.4 million mostly rural Californians lack broadband access at any speed.

Working on behalf of the CBTF, the California Institute for Telecommunications and Information Technology (Calit2) partnered with one of the

nation’s largest providers of broadband measurement, Speedtest. Net, to acquire 2006 California data. This set included 350,000 users in California who conducted 1,243,278 tests of their effective bandwidth.

As policymakers determine whether and where to direct scarce resources, it is critical for them to have access to current broadband deployment information. The data underlying these maps is the result of a project undertaken by the CBTF to map broadband availability throughout California. Through the California Emerging Technology Fund, broadband providers voluntarily submitted availability data to a neutral third party. The Office of State Health Planning and Development, acting as staff to the CBTF, created these maps using an aggregated dataset gathered by the third party. This project would not have moved forward without the investment of significant time and effort by the individual

broadband providers in California and the support of the industry trade groups. All of the largest wireline broadband providers participated in this effort, and a high percentage of the small providers also submitted the requested data.

    Through this effort, over 15 million supplied addresses and 7 to 8 million address equivalents were processed. In total, 11 of 12 Incumbent Local Exchange Carriers (ILECs) and 6 of 10 cable companies participated in this project. "

From NTIA, data provided for perspective. I'm printing verbatim.

1.       National Broadband Map Data is at the census-block level.

NOFA clarification: http://www.ntia.doc.gov/frnotices/2009/FR_BroadbandMappingNOFA_090708.pdf

Background:

 All data is provided at the census block level, blocks over 2 square miles we collected at the road segment level. The census block is the smallest unit of geography for which the Census Department makes records available. Most of the country lives in census blocks that are less than two square miles.  In many urban areas, the blocks are much smaller. In aggregate, providers generally don’t have data at the address level and most states do not possess a complete address file. Requiring address-level data would have lead to inaccurate or incomplete information. NTIA will continue to strive to develop the most granular data possible in the future and is even helping to fund several states to develop address files.

2.       Explanation of  our 5-10% number from the press release

 Statement attributable to NTIA from our press release:

 “The map shows that between 5 - 10 percent of Americans lack access to broadband at speeds that support a basic set of applications, including downloading Web pages, photos and video, and using simple video conferencing. The FCC last July set a benchmark of 4 Mbps actual speed downstream and 1 Mbps upstream to support these applications. NTIA collected data in ranges between 3 – 6 Mbps and 6 – 10 Mbps maximum advertised download speeds, which are the closest measurements to the speed benchmark for broadband that the FCC set.”

Background:

See the FCC’s press release on the 706 item described broadband: http://hraunfoss.fcc.gov/edocs_public/attachmatch/DOC-299989A1.pdf

While the FCC refers to broadband as actual speeds of 4 Mbps down and 1 Mbps, that is not a metric used in the NBM.  NBM data reflects maximum advertised speed tiers, not actual speeds. The closest are  maximum advertised speeds of:

a) equal to or greater than 3 Mbps and less than 6 Mbps

b) equal to or greater than 6 Mbps and less than 10 Mbps

These speed tiers are not available to approximately 5 and 10 percent of Americans, respectively. Note that we adopted our tiers BEFORE the FCC set the new 4 Mbps threshold

3.       Background point regarding the CA map:

-          One study was done in 2007 with less than 30 providers, the other was done more than three years later with data from over than 80 providers. The studies used different variables such as speed tiers and data collection methodologies.

Additional Background:

-We encourage all users to use our crowdsourcing features to continue to improve the map.

-All of our data is downloadable.

-More than 1650 unique providers from across the country provided data for the map. We encourage small providers who did not submit data to participate in the future.

See also Anne Neville’s recent blog: http://www.broadbandmap.gov/blog/2486/25-million-records/

Thanks,

Moira

25 Million Records

Posted on February 25, 2011 by admin

Last week we released the first version of the National Broadband Map (NBM). The data it contains represents the hard work of 50 states, 3 territories and the District of Columbia over the last year and a half. With over 25 million records, this dataset is the first of its kind and provides an invaluable resource of information on broadband availability in the Nation. As with any first, however, there is information that needs to be corrected that does not display correctly.

We identified some of this information at launch and listed it on our FAQ section.

We are currently updating this section to address other issues we have identified. We want to call particular attention to how Arkansas is displayed on one of our maps. Arkansas provided its data to us, but due to some processing issues, that data is not currently displaying on the Broadband Availability across Demographic Characteristics . As we work to fix this gap, we recommend you look at the Analyze section of the NBM to see this type of information about Arkansas.

You may also want to check out Connect Arkansas which has done a great job of gathering this data and organizing communities to expand broadband opportunities. They’ve also completed a survey about broadband adoption in their state that yields some fascinating information.

We’re very excited about the response to the National Broadband Map, appreciate all your feedback and encourage you to keep using the crowdsourcing tools on the website.