The FCC is in the process of reworking its broadband maps. The task of doing so is complicated and the new maps are likely going to be a big mess at first. In a recent article in Slate, Mike Conlow discusses two of the issues that the FCC will have trouble getting right.
One issue is identifying rural homes and businesses. We know from recent auctions that the FCC assumption on the number of homes in a Census block is often wrong. It’s hard to count homes without broadband if we don’t know how to count homes in general. The mapping firm CostQuest suggests counting homes using satellite data. But the article shows how hard that can be. For instance, it shows a typical farm complex that has multiple buildings. How does an automated mapping program count the homes in this situation? Mixed among the many farm buildings could be zero homes, one home, or several homes.
If you have ever looked at satellite maps in West Virginia, you see the opposite problem. There are homes under total tree cover that aren’t seen by a satellite. To really complicate matters, there are several million rural vacation homes in the country, many not more than a shack or small cabin, many without power. How is satellite mapping going to distinguish a cabin without power from a home with full-time residents? It’s unlikely that a national attempt to count homes using satellite data is going to get this even close to right – but it means many millions to CostQuest to try.
The second mapping issue comes from ISPs that will have to draw polygons around service areas that have broadband or can get broadband within 10 days of a service request. The article shows a real example where it’s easy to draw a polygon along roads that will leave out homes that are back long lanes or driveways.
When ISPs convert to this new mapping with the polygons, especially if housing data comes from satellite imagery, the resulting maps are going to have a lot of problems. The first iterations of the new maps will differ significantly from today’s mapping and it’s going to be nearly impossible to understand the difference between old and new.
As complicated as these two issues are, they are not the biggest problem with the mapping. The big issue that nobody in Congress or the FCC wants to talk about is that it’s nearly impossible to know the broadband speed delivered to a home. For most broadband technologies, the speed being delivered changes from second to second, and from minute to minute. If you don’t think that’s true, then run a speed test at home a few dozen times today, every few hours. Unless your broadband comes from a stable fiber network, the chances are that you’ll get a wide range of speed test readings. After taking these multiple tests, tell me the broadband speed at your house. If it’s hard to define the speed for a single home, how are we supposed to tackle this in mass?
But let’s just suppose that in some magical way that the FCC could figure out the average speed at a home over time. That still doesn’t help with the FCC mapping because ISPs will be allowed to report marketing speeds and not actual speeds to the FCC. The Slate article suggests that the biggest problem in today’s maps comes from counting broadband by Census blocks – where if one home has fast broadband, the entire Census block is counted as fast. That is a much smaller issue than people assume. The majority of misstated rural speeds today come instead from ISPs that claim they sell a speed that is much faster than what is delivered. Big telcos today report rural areas as having 25/3 capability for no reason other than the ISP says so – when in reality there might not be even one customer in that area that has even 10/1 Mbps DSL. The big telcos have successfully been lying about speed capability for years as a way to shield areas against being overbuilt by grants. Recall that Frontier tried to sneak in over 16,000 speed changes for Census blocks just before the deadline of the RDOF grant. The new mapping is not going to be a whit better as long as ISPs can continue to lie about speeds with impunity.
There are a few simple ways to fix some of the worst problems with the maps. First, the FCC could declare that all DSL is no longer broadband and stop bothering to measure DSL speeds. They could do the same with high-orbit satellites that have huge latency issues. But even doing this solves only a portion of the problem. There are still numerous WISPs that report marketing speeds that are far faster than actual speeds. The FCC maps are also about to get inundated by the cellular companies making the same overstated speed claims for fixed rural cellular broadband.
What is so dreadful about all of this is that a rural home may have no real option for broadband but might have FCC maps that show they can buy fast broadband from DSL, one or more WISPs, and one or more fixed cellular providers. The FCC is going to count such a home as a success because it has competition between multiple ISPs – when in reality the home might not have even one real broadband option.
I hate to be one of the few people that keep saying this – but I’m sure that the new FCC maps won’t be any better than the current ones. Unfortunately, by the time that becomes apparent, Congress will have assumed the mapping is good and will have moved on to other issues.
Great commentary as usual Doug! In South Carolina due to our mapping efforts, we can verify much of what you say is true.
Indeed, the speed test results from the WA State Broadband Office show too many conflicts with the FCC maps. The FCC shows northeastern WA as 100% served at 100/100 Mbps but look at all the black (no service) and red (less than 10 Mbps) speeds tested. https://tinyurl.com/WARealitySpeeds
Completely agree. We’ve done some work over the last year in that part of the state and the FCC maps are way off base. Unfortunately, I think this is true to some degree in 90% of the counties in the country.
I have seen suggestions to use Ookla results, but they don’t tell you what is available. I just ran seven speed tests using 2 wireless computers and one hardwired on my DSL and the Ping ranged from 6-89ms. The Download was 22.99 to 64.90Mbps and Upload was 2.47 to 8.13Mbps. I subscribe to the highest DSL speed available. What if I only subscribed to 10/1? Ookla can only tell what speeds were tested and cannot tell if they were using a wireless router or direct connect. They cannot tell what wireless card/chip I have in the computer. They cannot even tell what speed I am paying for. They also cannot tell whether there are a bunch of other computers, TVs, cell phones, etc. connected simultaneously that would reduce bandwidth available for speed test, or how much overlap there is between wireless networks in the vicinity.
Once we start creating the polygons, it will be very difficult to determine if a location qualifies. For instance if the house is a mile from the road, but the parcel is within 500′ of the fiber cable, it qualifies. However, if the location is 600′ from the cable, but the parcel isn’t within 500′ of the cable, it doesn’t qualify. Of course, if you already serve the location with a fiber, then it qualifies. This is just for fiber. Wireless depends on propagation maps, etc.