Unserved and Underserved is a distinction without a difference

and a follow up on states that might fall short on reaching the Unserved

Senator Thune recently asked broadband stakeholders what Congress could do to improve the Infrastructure Bill with respect to broadband. My suggestion: remove the meaningless distinction between Unserved and Underserved (capital “U”s), allocate money to states based on the total number of locations that are on the wrong side of the Digital Divide, and give states the tools and flexibility to blanket the underserved (lower case “u”) with good Internet.

Notwithstanding the [i think] strong case I’m about to make that there is no distinction between Unserved and Underseved, the word “Unserved” appears in the broadband section of the Infrastructure Bill 32 times, and it is currently the law. So I want to take a moment to follow up on my previous post about how far the BEAD money might be able to go in each state. Specifically, there are 11 states that might not be able to reach all of their Unserved, with Utah, Alaska, and Idaho potentially only reaching 45%, 52%, and 69% respectively. This estimate takes into account that RDOF locations are “free” to the state, but does not take into account ARPA-funded locations (more on that another time). In addition to RDOF, the big factor at play here is how expensive it might be to bring fiber to a location. Expensive states with low RDOF funding, like Montana, are bad off. The BEAD allocation formula just doesn’t move enough money into expensive states.

Now back to the Unserved and Underserved and how there’s no difference. A lot of us have this expectation that very rural areas are completely Unserved, then there’s some kind of less rural segment (rural towns, for example) which are Underserved. And people that are Unserved have no access to the Internet and people who are Underserved have some kind of passable but not ideal access. None of it works like that in the real world.

The chart above ranks every Census block in the country from least dense (on the left of the chart) to most dense (on the right). Then it adds the number of Served, Underserved, and Unserved Broadband Serviceable Locations in those Census blocks. The most rural parts of the country are where we find most of the Unserved. That isn’t surprising. What does surprise is that we’re finding Underserved as well, at a ratio of about 2:1 Unserved:Underserved.

By the time we reach the 30th percentile — the top 30% most rural Census blocks — we’ve reached 72% of all the Unserved and Underserved. At this point, there’s still a healthy mix of Unserved and Underserved locations — 1.4 Unserved locations for every Underserved location. The point: all the way up the curve as neighborhoods become more dense and more Served, there’s still a mix of Unserved and Underserved.

Here’s another way to think about it, especially for those of us who worked with the Form 477 block-level data for a long time. Previously a block was entirely Served, entirely Underserved, or entirely Unserved. Now with the new maps, a block can still be completely Served or Underserved, but it can also be a mix of two, or even all three categories. At the 20th percentile, which is still quite rural, 23% of locations are in blocks that have Served, Underserved and Unserved locations in them. Additionally, 15% of locations in that same type of block are both Unserved and Underserved (but not at all Served).

Let’s look at some random examples that show what’s happening. First up is some farms outside Prescott, Arkansas. Here, what’s happening is that CenturyLink is “offering” 60/5 DSL to some locations (green dots, which are Underserved), but 10/1 DSL to other nearby locations (red dots, Unserved). (One of those red dots is a small pond in a forest - see my post about overcounts).

Here’s the same general area, where 40/3 DSL (which is Underserved) are mixed with 10/1 DSL just hundreds of feet apart.

Here’s another example in Iowa. In this case, US Cellular “offers” 25/3 fixed wireless to this location, which is almost certainly a guest house or barn (again, overcount). But they do not offer service to the main house, which is Unserved.

Here’s another example nearby. (I could come up with lots more examples because I have a list of hundreds of thousands of blocks where this happens).

Cable and fiber offerings are almost all Served — 98% or more of cable and fiber’s footprint have access to 100/20 throughput. But DSL and fixed wireless both straddle the all-important thresholds at 25/3 and 100/20. Fifty-nine percent (59%) of DSL offerings are Unserved while 35% are Underserved. In licensed fixed wireless, 43% of the offerings make the location Unserved, while 39% of offerings are Underserved.

As most readers know, ISPs can put almost anything into the “maximum advertised speed” field in their filings. As I discussed previously, challenging the “maximum advertised speed” numbers is essentially impossible with the current FCC rules. An approach that’s more likely to succeed is showing that a provider doesn’t offer any service to the location. That’s the approach Nevada is trying for 20,000 locations, with the support of their two senators. It might work, but it might not.

These DSL and fixed wireless filings wouldn’t be a big issue if the 25/3 threshold didn’t make such a big difference for funding levels. After all, who cares if US Cellular claims 25/3 for a service that no one uses? It’s unusable for home Internet if they claim 25/3 or 10/1 or no coverage at all. But when 2,043,676 locations hang in the balance — the number of locations that are Unserved but for the presence of a 25/3 FW or DSL offering on the number — it makes a big difference. The way it’s currently set up is costing North Carolina about $700 million because they have 193,000 of these locations.

There are big things we could do to fix this problem like rewriting how ISPs report the throughput they offer, or trashing the whole idea of reporting throughputs and just inferring the throughput from the technology. Those are hard — or at least improbable. The easy approach is to remove the distinction-without-a-difference between Unserved and Underserved, do the allocation based on the combined number, and build broadband to as many of those people as we can. That’s my suggestion to Senator Thune.