Absent New FCC Broadband Maps, Local Govs Plot Coverage

In the national broadband conversation, people disagree about many things, including the type of technology that should connect users, the organization that should provide the service, the way a project should be paid for and the very definition of high-speed Internet. But one point seems to unite all stakeholders: The Federal Communications Commission’s broadband coverage maps are inadequate.

The FCC maps use Form 477 data, which is broadband coverage information reported by Internet service providers. For several years, the maps have determined which geographic areas will receive federal broadband subsidies. Under this system, unserved areas are prioritized. The problem comes with the definition of “served.” If an ISP says it provides high-speed Internet, defined as 25 Mbps/3 Mbps, to one location within a census block, the census block is considered served. But why should one assume that a single served location is representative of an entire geographic unit?

After years of criticism directed at the FCC maps, former President Donald Trump signed the Broadband Deployment Accuracy and Technological Availability Act in March 2020, which requires the FCC to produce more detailed and accurate maps. Although two years have passed since that legislation’s signing, the FCC has not yet released a date for completion of the updated maps. And so, in a world where broadband is more critical than ever, states and local areas find themselves trying to answer an old question: Where does the digital divide exist? The only way to answer this question is to put together a more granular broadband map — a tall order given challenges related to data, methodology, coordination and cooperation.

SAN ANTONIO

Interim Digital Inclusion Administrator Candelaria Mendoza said broadband mapping in San Antonio is part of a larger digital inclusion initiative. During late 2019 and early 2020, the city surveyed its residents on whether they have high-speed Internet access at home and, if so, how they are getting the connection (through a router, cellular plan or both). The University of Texas at San Antonio helped the city create a GIS map using the survey responses, which totaled over 6,000. In addition, San Antonio has been able to map out where some broadband infrastructure exists, as the city has its own fiber.

The mapping has given the city some “middle ground” on the truth about broadband availability in its communities. Without this effort, San Antonio would only have Form 477 data to base decisions on. In regard to relying on available ISP-reported data, Mendoza said it can be frustrating hearing that San Antonio is almost completely connected when the city knows that many community members face affordability barriers and have to rely on outdated copper infrastructure. In her view, there should be more attention paid to the “quality of Internet” that people are receiving.

Moreover, more densely populated urban areas tend to get the short end of the federal funding stick when it comes to broadband subsidies. Mendoza expects rural areas to be favored again when the funding formulas are created for the $42.5 billion broadband deployment program created by the Infrastructure Investment and Jobs Act. And yet through its own mapping program, San Antonio can show that its digital divide disproportionately affects people of color and households with low income and low education levels.

“We want to make sure we don’t forget them and that they’re part of our storytelling,” Mendoza stated.

In terms of the future, Mendoza said San Antonio will continue having a dialog with its provider community. The city has found that it’s best to hold “one-on-one conversations” with providers. Mendoza pointed out that while companies aren’t sharing data that they consider proprietary, provider input can help shape and inform local strategies. She also noted that Texas recently established its own broadband office and has issued an RFP for a state broadband mapping project.

“We’ve already volunteered to help [the state],” Mendoza said. “If there’s anything we can share from our lessons learned, we’re definitely an open book.”

LOUISVILLE

Louisville, Ky., has had multiple broadband-related mapping projects. Grace Simrall, chief of civic innovation and technology, said the first effort, named Louisville Fiber, involved identifying which ZIP codes had a demand for fiber to the home. The second effort, Speed Up Louisville, initially used crowdsourced data to map out Internet speeds at people’s homes. The city would later integrate this data with speed test information obtained from Measurement Lab (M-Lab), an open source global measurement platform.

“We had decent participation [with crowdsourced speed data],” Simrall said. “But it’s nothing like people going to their browser and taking the M-Lab Internet speed test, which is embedded in multiple search engines.”

Louisville also maps Form 477 data as it becomes available and supplements that with American Community Survey (ACS) information as well as Emergency Broadband Benefit (EBB) data, which allows the city to identify EBB enrollment numbers by ZIP code. The data sets have revealed that, like San Antonio, broadband gaps tend to show up in areas of concentrated poverty.

“There’s a lot of value in mapping that FCC Form 477 data,” Simrall explained. “Bear in mind, this is the ISPs self-reporting, so it is their opportunity to portray themselves in the best light. At least in the case of Jefferson County, I can tell you that even in the best light, there are areas where it’s very obvious that there are gaps. And again, those gaps directly mirror our social divide gaps. There’s value in being able to map that and demonstrate that.”

Louisville faces two main challenges with its broadband mapping. First, like San Antonio, Louisville hasn’t convinced its providers to publicly share any broadband availability data. However, Simrall said the city has decent relationships with its two main providers, which are willing to share maps with Louisville in private. Second, different data sets use different geographic units, which can make analysis tricky.

“If you want to get down to anything lower [than a county], whether that’s a ZIP code or census block, the challenge you might run into is that the reference data that you’re dealing with is broken down in a different way,” Simrall said. “We’re receiving this EBB data by ZIP code, but when I look at my American Community Survey data, it’s actually broken down by census tract. Marrying the two things, you might have a census tract that splits across two or more ZIP codes. … You can’t assume that all your census tracts are going to fit neatly into ZIP code boundaries and vice versa.”

One major advantage of mapping EBB enrollments is that it can help a local area reach underserved populations. Simrall said her data inclusion manager will leverage the map data to figure out where Louisville should have targeted campaigns to ensure that residents know a federal broadband stipend exists and whether they are eligible for it. As confirmed by a report from BroadbandNow, only a fraction of eligible households took advantage of EBB, which was replaced by the Affordable Connectivity Program (ACP) at the start of the year.

“You don’t need the data to be perfect to make decisions,” Simrall said. “What you’ll probably discover, especially when you’re looking at this EBB data or the new ACP data, is it will give you very clear areas for you to focus your efforts and make sure that as a community you’re not leaving money on the table.”

WISCONSIN

Despite the fact that Wisconsin has mapped broadband availability for a grant program to some extent since 2013, the state was hardly getting any sub-census-block data from its provider community at the start of the COVID-19 pandemic. Thanks to consistent dialog with the state’s estimated 95 providers, the situation has significantly shifted in two years.

“We have about 45 providers who are giving us data in some form or another at a finer-grained level than the census block, usually using GIS data of some sort,” said Colter Sikora, broadband mapping coordinator with the Public Service Commission of Wisconsin. “Knowing that we’re reaching that 50 percent point of our Internet service provider community that’s giving us that data, I think that’s a real victory for improving the accuracy of the broadband maps. Of course, it’s yet to be seen what the quality of that data is.”

According to Sikora, Wisconsin poses a simple inquiry to providers: What can you serve at this point? Although the question is straightforward, the providers are still trying to figure out how to map effectively for their own purposes. With 45 different providers submitting data, the state could be seeing 45 different mapping procedures. The state has no uniform procedure to date, though if the feds come up with an official process, that might change.

“We don’t have our data collection in statute or anything insanely formal … but it is along the lines of if someone calls you up and says, ‘I need service, and I’m in your blob,’ then there better be, at least in my eyes, a better than 95 percent chance that that person’s going to have service within the standard 10 business days,” Colter said. “Not ‘We can build this, and next year you can have it.’ That’s not OK. There needs to be infrastructure that’s reachable.”

Interestingly, Wisconsin isn’t asking for address-level data. From the state’s perspective, addresses are limited as a unit of analysis when it comes to having true foresight about where high-speed Internet can exist. The goal is to identify where service can be achieved, not merely where service is available.

“Especially in agrarian Wisconsin or rural anywhere USA, you have plenty of turf where, sure, you could put a house there, but right now it’s a cornfield, and there are no address points other than maybe a driveway to access that cornfield,” Sikora pointed out. “To limit the scope of broadband mapping to, say, coverages available at this address and that address and the other address doesn’t necessarily answer the whole question of, ‘Where can a provider provide service right now and where can’t it?’ Because if there are houses on one street corner and houses on another and fields in between, what’s happening in between still matters as far as development decisions are concerned.”

Key to the state’s success is, regardless of political administration, its broadband office’s ability to maintain healthy relationships with providers. Sikora said states can also take advantage of private-sector entities that serve as a middleman between government and providers to make processes like data collection and mapping easier. There also needs to be actual vision behind what a state will do with the data it collects, because there’s little point in doing something if it’s not going to be helpful down the road.

Another point involves being considerate of providers. The state believes waiting for procedural guidance from the FCC is a good strategy, as one doesn’t want to lay down strict guidelines that might lead to redundant work for providers once a future FCC granular mapping process is established. Providers could also be reluctant to participate if a state’s expectations for data are too high.

“Asking too much certainly can be and, probably for some providers, is a deterrent,” Sikora said. “That, I don’t think is a question. Where that boundary is varies from provider to provider.”

GEORGIA

In contrast to how Wisconsin handles mapping, Georgia specifically utilizes address-level data and aggregates it up to the census-block level to create a comparison with Form 477 data. All in all, the state has mapped about 5.3 million locations in 159 counties. And starting in 2021, Georgia added unserved locations during its annual June 30 map refresh, said Eric McRae, associate director at the University of Georgia’s Carl Vinson Institute of Government, which does the technical work for the state map on behalf of the Georgia Department of Community Affairs, the lead agency.

Like most complex and unprecedented endeavors, the path to Georgia’s statewide broadband map had its share of bumps. Initially when the project started in 2019, McRae’s team collected the address data and then gave that information to the providers so that the providers could map it. This approach ended up being too heavy of a lift for providers, who then suggested that the state collect and map the provider data.

“We worked with the 44 providers to get all of their data in various forms and then standardize it against a master address file that we actually commercially purchased in the state of Georgia, because we didn’t have a master address file for the state at the time,” McRae recounted.

As is the case in Wisconsin, communicating with providers has been a major element of Georgia’s progress. Georgia has a provider advisory board, which informed the state that companies didn’t want served locations to be collected and shown based on competition concerns. But McRae said providers may one day become comfortable with going in this direction. It’s all about respect and gradual negotiation, so coming to providers with lofty expectations isn’t the way to go.

“Start talking with them immediately and find out what they’re comfortable with and what they’re not comfortable with,” McRae advised. “I’ve talked with several states that are starting a mapping process, and they’re basically going for everything. They want to collect every single speed, every single this. They want to have all of this information. I told them, I said, ‘That’s a heavy lift for a provider, especially since you’re not paying them to give you this information. You’re asking.’”

McRae also warned against attempting to use legislation to tell providers what types of information they must fork over.

“They’re not a regulated community,” McRae said. “You don’t have the teeth that you think you do.”

McRae shared another piece of advice that recalls what Simrall communicated: Don’t worry about having everything perfect. There are going to be data inconsistencies. There are going to be conundrums that you may not decipher. Being almost correct is going to get you further than relying only on Form 477 data to make decisions.

A principle that goes hand in hand with this logic is not being overzealous about map updates. McRae flatly stated monthly map updates are too much to handle for everyone involved and can get in the way of planning. Data will change after you publish the map, but that information can be saved for a bigger planned update, and the state can still keep the information in mind so that overbuilds don’t occur in particular areas. McRae wouldn’t suggest updating a broadband map more than twice a year.

“Trying to publish the map every time someone does an update, you’re going to basically be chasing your tail all the time,” McRae said.