About DownloadSpeed

Why FCC Broadband Data Collection deserves a public-facing home

The Federal Communications Commission Broadband Data Collection (BDC) is the federal record of internet availability across the United States. Every twice-yearly filing cycle, every Internet Service Provider with a presence in the U.S. submits a list of locations they serve and the technologies and speeds they offer at each. The data is the most comprehensive public source for who has broadband and what they actually have.

The data quality has been controversial. The BDC’s predecessor (Form 477) was widely criticized for overstating coverage because it asked providers to report by Census block — a single served address in a block became ’served’ for the whole block, which produced famous over-coverage maps. The BDC requires more granular location-by-location reporting and includes a challenge mechanism that lets residents and local governments dispute incorrect entries. BroadbandGrade reads the BDC data and presents it per-ZIP, per-county, per-state, with explicit notes on where challenge-driven revisions are in flight.

How the pipeline pulls FCC data

The pipeline pulls from the FCC BDC public dataset on the semi-annual release cadence — typically May and November releases covering the prior December and prior June filing periods. Each pull touches every served location filed by every covered provider, aggregates to ZIP and county levels, and produces the per-provider, per-geography pages. The site stamps the BDC release date on every value.

A practical detail: the BDC challenge process can produce revisions to a provider’s reported coverage months after the original filing. The site does its best to capture both the originally-filed and revised values, but in some cases the source data overwrites the original without preserving the challenge history. For specific addresses where the BDC report disagrees with the resident’s actual experience, the FCC challenge portal is the right corrective channel.

Where broadband data has caveats

Three caveats. First, provider-reported availability is widely acknowledged to overstate real coverage at the block level. The BDC’s per-location reporting is more accurate than the older Form 477 per-block reporting, but the underlying incentive — providers want to look comprehensive — still produces over-counting at the margin. The pages note where coverage looks unusually high for the actual housing stock as a hint that the reported figure may merit a challenge.

Second, the ’served’ threshold is 25/3 Mbps for the FCC’s traditional definition or 100/20 Mbps under the more recent BEAD program definition. The two thresholds produce very different coverage maps; the site uses the BEAD threshold by default because it is the operational standard for current federal investment decisions, but the 25/3 figure is also available on the per-geography pages.

Third, actual speeds at any specific address depend on local infrastructure, time of day, and oversubscription on the provider’s local node. The BDC reports the maximum advertised speed the provider says they can deliver to the address; actual experienced speeds can run substantially lower, particularly during peak evening hours. For ground-truth speed data, the M-Lab speed test dataset is a useful supplement and the site links to it from the methodology page.