Prpoosed formatting standards for machine-readable labels and archiving of prior labels.
Please review and make suggestions, either by opening an Issue or creating a Pull Request.
The FCC deadline is 10 October 2024. https://docs.fcc.gov/public/attachments/FCC-22-86A1.pdf
Excerpted from the FCC document:
-
Machine-Readable Format. We require providers to make the information included in the label available to the public in machine-readable format.142 By “machine readable,” we mean providing “data in a format that can be easily processed by a computer without human intervention while ensuring no semantic meaning is lost.”143 Providers should make each label’s information available by providing the information separately in a spreadsheet file format such as .csv. These files should be made available on a provider’s website via a dedicated URL that contains all of a provider’s given labels. We require providers to publicize the URL with the label data in the transparency disclosures required under 47 CFR § 8.1(a). These machine-readable files must provide the same categories of information as those presented in each label, including the unique identifier described below. We direct CGB, in consultation with other relevant bureaus, to make available on the Commission’s website resources that may help providers satisfy the machine-readability requirement, such as sample machine-readable spreadsheet files.144 Further, given the importance of this requirement, we will monitor providers’ implementation of machine readability to ensure providers’ implementation of this requirement is useful to third parties and the Commission in its data collection efforts.
-
Although section 60504 of the Infrastructure Act does not expressly address the format requirements for broadband labels, implementing broadband labels with a machine-readability requirement advances the statutory objective of providing consumers with sufficient key information needed to evaluate broadband Internet access service plans in a manner that is available when they need it and most effective for them.145 We agree with commenters that making the label information machine readable will yield a number of benefits to consumers. For example, machine readability will enable third parties to more easily collect and aggregate data for the purpose of creating comparison-shopping tools for consumers.146 These tools may include browser add-ons or websites that compare plans offered by different providers.147 Making the information machine readable also helps ensure that the data third parties use is both accurate and up to date. Because providers often “adjust . . . [their] business offerings,” we believe it may be simpler for them to “re-enter the new information and re-upload [their] labels” in a machine-readable format.148
-
Machine readability also promotes both competition as well as transparency and accountability. Consumers may use the data collected in this manner to compare typical speeds reported by subscribers versus those reported on a broadband label.149 And, as AARP explains, the generation of shopping tools like these helps promote “digital equity” for groups lacking the necessary expertise to parse what is often complicated language contained in service agreements.150 These tools can assist such groups, including older Americans, to more easily obtain the information they need to select the service plan that best meets their needs.151
-
Further, requiring ISPs to post machine-readable label information will allow the Commission to more easily collect data about broadband markets.152 Information collected via machinereadable labels may also make monitoring for compliance with Commission rules and enforcement more efficient as well.153 A machine-readable label could, for instance, help determine if “a provider has published [a] properly formatted label . . . online.”154
-
While each of the foregoing benefits would be sufficient to persuade us to adopt this requirement, we further observe that a machine-readability requirement will make data more easily available for research as well.155 As New America’s Open Institute of Technology explains, broadband affordability research that is reliant on manual review of existing provider advertising can be a “timeconsuming and laborious process that many organizations are unable to undertake.”156 The Institute for Local Self-Reliance, which itself has “been forced to abandon research projects because of the industry’s information gaps,” observes that the broadband consumer label provides “an excellent opportunity to facilitate research efforts” by “allow[ing] researchers to aggregate data at a large scale and analyze this data.”157 Such research can serve industry, policymakers, consumers, and advocacy groups by providing a clearer picture of the marketplace.158
-
The record shows that these benefits can be achieved at a low cost to providers, with no commenters providing cost data to suggest otherwise.159 We agree with AARP that making the broadband consumer label data machine readable does not impose a high burden or require special technical expertise.160 We find ACA Connect’s argument that such a requirement would “tax the resources of small providers with limited in-house technical resources” unpersuasive, as they fail to elaborate why or substantiate their claim with any evidence.161 Further, we do not believe that publishing the label information in a spreadsheet file would impose a high technical burden. And as noted above, the Commission will offer resources to ease compliance with this requirement.
-
We disagree with commenters that argue that requiring the label to be machine readable creates difficulties for providers because of “information on the label [that] cannot be boiled down to a binary response.”162 First, commenters opposed to machine readability fail to describe what kind of information is lost and how that may impact consumer choice.163 NCTA only cites descriptions of onetime fees as an example where oversimplification may be required.164 However, NCTA does not explain how “semantic meaning is lost” or what inaccuracies might be introduced.165 To the extent that providers request “flexibility” to provide additional information in the label not required by the Commission,166 information that may not be easily reducible to binary responses, we note that this is not the label’s purpose. Indeed, to the extent that machine readability promotes “apples-to-apples” comparisons that do not reflect every nuance that differentiates plans,167 we agree with AARP that this does not necessarily represent a flaw.168 One of the goals of the broadband consumer label is to simplify the process of comparison shopping and make the most critical information readily available to consumers.169 Thus, we agree with AARP that conveying the type of information opponents argue may not be picked up by a program “is secondary to label data needed to make apples-to-apples comparisons.”170 We also agree with commenters that the benefits outlined above outweigh these concerns over flexibility.171
-
NTCA and WISPA’s invocation of the nutrition label model, which they argue “is not designed to serve as [an] on-ramp to electronic comparison shopping,”172 to oppose a machine-readability requirement also proves unconvincing. Nothing about a machine-readability requirement undermines the broadband consumer label’s ability to provide “rapid and comprehensible comparison among products.”173 Simultaneously, shopping for broadband is a more involved process than purchasing a food product. It involves selection of a service that normally requires ongoing, periodic payments, that may involve a contract, and that impacts various facets of an individual’s life.174 Such a choice reasonably takes more time and research than that spent in a food aisle, making NTCA and WISPA’s comparison in this regard inapt.
-
We also disagree with AT&T’s assertion that machine readability is not “designed to help the consumer at the point of sale but rather to facilitate third parties’ desire to conduct various forms of research or analysis,” which AT&T claims is “not the purpose of the labels.”175 As described above, machine readability enhances the point-of-sale experience in a variety of ways, including in the form of third-party shopping comparison tools. While AT&T claims that machine readability “could fatally compromise broadband providers’ ability to . . . convey accurate information on the labels,”176 AT&T does not elaborate as to how. To the extent that machine readability fails to capture all the benefits of a given plan, we agree with Consumer Reports that the Commission can expect “the creativity of ISPs” will lead to solutions for “further explain[ing] the details of their service offerings to appeal to a wide range of audiences.”177
-
We recognize, however, that the Commission did not include a machine-readability requirement in 2016 and that this will take some additional effort. We therefore delay compliance with this requirement until one year after the Office of Management and Budget completes its review of this new information collection.
-
Unique Plan Identifiers. We require ISPs to develop unique identifiers for each of their plans and attach them to the broadband label. The unique identifier should consist of a unique ID for fixed plan or mobile plan (“F” for fixed plans and “M” for mobile plans), followed by the broadband provider’s FCC Registration Number,178 and ending with a provider-chosen string of precisely 15 alphanumeric characters uniquely identifying the specific plan within the broadband provider’s offerings. The Unique Plan Identifier shall not include special characters such as, &, *, and %. For example, AT&T could specify a fixed broadband offering as F + 0005937974 + 123ABC456DEF789. This would appear on the label as F0005937974123ABC456DEF789. Unique identifiers should be sufficiently distinctive so that third parties and the Commission can identify the specific plan identified by the unique identifier.179 Additionally, reuse of identifiers must not occur; even if a given plan is no longer offered, its string should not be repurposed for a new or different plan.
-
Unique identifiers are useful for a variety of purposes. For example, use of a unique identifier would enable ISPs, which often change their plan offerings, to reuse a given plan’s name without creating confusion. While NCTA argues that unique identifiers are unnecessary for this purpose, they do not describe the “significant burdens” they claim would be imposed.180 USTelecom notes that requiring provider-created unique identifiers would not “creat[e] undue burden on providers or increas[e] administrative costs.”181
-
Additionally, unique identifiers may be helpful in reducing ambiguity in other contexts as well. Third-party shopping tools might benefit from ISPs’ use of unique identifiers. And researchers may find it helpful having a shared, consistent means of identifying ISPs’ plans as opposed to use of descriptive language that could result in confusion about which plan is being discussed.
Footnotes
143 See 44 U.S.C. § 3502(18).
144 AARP Reply at 9-11 (stating that machine-readable data requires a consistent structured data format, which could be enabled by the Commission providing an API for inputting data, or by providing a template in Excel format, to allow the creation of information in both .csv and .xml formats and the machine-readable format requirement should impose no significant burden on the ISP).
145 Infrastructure Act § 60504. We thus find Lumen’s statutory assertions unavailing. See Lumen Comments at 13 (arguing that making label information machine readable “is clearly not something the Infrastructure Act mandates”). Although we impose a machine-readability requirement and require that providers associate their plans with a unique identifier, we do not address issues related to any information collection required under sec. 60502(c) of the Infrastructure Act. Such issues remain pending in WC Docket No. 21-450. See Affordable Connectivity Program, WC Docket No. 21-450, Notice of Proposed Rulemaking, FCC 22-44 (rel. June 8, 2022).
146 See, e.g., AARP Comments at 4, 18; Cloudflare Comments at 11; Consumer Reports Comments at 6; NYC Comments at 3; OTI Comments at 11-12; NCC 4/7/22 ex parte at 2; Cranor 8/18/22 ex parte at slide 16.
147 OTI Comments at 11; see also Consumer Reports Comments at 6 (observing that machine readability would simplify the task of third parties to “compare service plans and their cost[s]”).
148 ILSR Comments at 7; Consumer Reports Comments at 6; see also NCTA Comments at 16-17
149 ILSR Comments at 7; see also Cloudflare Comments at 11 (arguing that machine readability would “help ensure transparency of . . . measurement” by enabling third parties to collect such data).
150 See AARP Comments at 21.
151 AARP Comments at 18.
152 See Consumer Reports Comments at 6; ILSR Comments at 7.
153 ILSR Comments at 7; Boston Joint Commenters Reply at 8-9; see also Consumer Reports Comments at 6 (observing that doing so would facilitate the Commission’s collection of broadband pricing data).
154 ILSR Comments at 6.
155 See AARP Comments at 3, 18; Cloudflare Comments at 11; Consumer Reports Comments at 6; OTI Comments at 11-12.
156 OTI Comments at 11-12.
157 ILSR Comments at 8.
158 AARP Comments at 21; Cloudflare Comments at 11; ILSR Comments at 8.
159 See AARP Reply at 10 (observing that use of a template in a “ubiquitous Excel format,” such as .csv and .xml, “should impose no significant burden on the ISP”).
160 AARP Reply at 10 (arguing that the creation of information in a machine-readable format does not “require[] any exceptional effort”).
161 See ACA Connects Comments at 11-12.
162 NCTA Comments at 17; see also CCA Comments at 6 (“machine[] readability across labels from different providers risks creating the misimpression that consumers can simply make an apples-to-apples comparison based solely on the information on the labels”); AT&T Reply at 22-23 (arguing that machine readability “could fatally compromise providers’ ability to maintain the flexibility they need to convey accurate information on the labels”).
163 While we acknowledge that information obtained by consumers, such as from third-party price aggregators, may not fully reflect all the costs and benefits of a given plan, this wider criticism applies to the use of a broadband consumer label generally and is not specific to a machine-readability requirement.
164 NCTA Comments at 17.
165 Id.
166 See AT&T Comments at 22; NCTA Comments at 17.
167 CCA Comments at 6.
168 AARP Reply at 10.
169 See NPRM, para. 1 (stating the need for “access to accurate, simple-to-understand information about Internet access services helps consumers make informed choices and is central to a well-functioning marketplace” and “enabl[es] consumers to comparison shop when choosing broadband services and providers that best meet their needs and match their budgets”).
170 AARP Reply at 10.
171 See, e.g., Boston Joint Commenters Reply at 8-9.
172 NTCA and WISPA Comments at 15-16.
173 Id. at 16.
174 The Commission, for example, has recognized how, since the start of the COVID-19 pandemic, individuals have increasingly relied on broadband Internet for “telework, remote learning, telehealth, and other online applications to meet our personal and professional needs.” Improving Competitive Broadband Access to Multiple Tenant Environments, GN Docket No. 17-142, Report and Order and Declaratory Ruling, FCC 22-12, para. 1 (rel. Feb. 15, 2022).
175 AT&T Reply at 22-23; see also NCTA Comments at 17 (arguing that machine readability is unnecessary for the purpose of comparison shopping); Letter from Brian Hurley, Vice President of Regulatory Affairs, ACA Connects, to Marlene H. Dortch, Secretary, FCC, CG Docket 22-2, at 3 (July 1, 2022) (ACA Connects 7/1/22 ex parte) (arguing that machine readability is not “directly related to the core purpose of the labels—providing timely information to consumers to inform shopping decisions”).
176 AT&T Reply at 22.
177 Consumer Reports Comments at 6.
178 Providers must use the FCC Registration Number that is used when submitting data to the Broadband Data Collection.
179 ISPs might consider use of other indicators, such as a ZIP Code of where the plan is offered, to set their identifiers apart. See Fields and Miller Comments at 1-2. 180 NCTA Comments at 17. 181 USTelecom Comments at 2