Why the FCC Broadband Map Says You Have Service When You Don't

Thousands of American households can point at the FCC National Broadband Map, see their own home marked as served by 100/20 Mbps or better, and then look at the DSL modem that has never cracked 8 Mbps. Both things are real. The map is not lying so much as answering a question nobody asked.

What the FCC map actually measures

The National Broadband Map is built from the Broadband Data Collection (BDC), the system that replaced the old Form 477 regime in 2022. Twice a year, every facilities-based provider files a list of locations where it offers service, along with the technology and the maximum advertised speed.

Those locations come from the Location Fabric, a national dataset of broadband-serviceable structures. This was a real improvement: under Form 477, a provider that served a single home in a census block could cause the entire block to be shown as covered, and rural blocks can be enormous. BDC fixed that specific flaw by moving to individual locations.

What BDC did not change is the underlying definition, and that definition is where most of the remaining overstatement lives.

The "could serve" problem

A provider may report a location as served if it currently provides service or could provide service within ten business days of a request, without an extraordinary commitment of resources or construction charges beyond the provider's standard installation charge.

Read that again, because every word is doing work. "Could provide" is a judgement call made by the provider about its own network. "Extraordinary" is not a number. "Standard installation charge" varies by company. A provider with a fiber terminal 800 feet from your driveway can reasonably conclude it could serve you in ten days, file you as served, and then quote you $14,000 in construction costs when you actually call.

Five more ways the picture drifts from reality

• Speeds are maximum advertised, not typical. A cable operator that sells a 1 Gbps tier files 1000 Mbps for every location on that node, regardless of what evening congestion does to it.
• Mobile coverage is modelled, not measured. Wireless carriers submit propagation maps generated from tower parameters and terrain. The models assume outdoor use and specified signal thresholds; they do not know about your metal roof, the hill behind the house, or a sector that is already full at 7pm.
• Capacity is invisible. A network can be technically present and commercially closed. Fixed wireless and 5G home internet providers routinely stop accepting new subscribers in a sector while remaining "available" on paper.
• The correction mechanism relies on you. The FCC runs an availability challenge process, and it works, but it puts the burden of proof on residents who must document a failed order. Most people never file one.
• Filings age. Data is submitted twice a year and published on a lag, so both new builds and quiet retirements take time to show up.

Filed availability versus measured performance

Notice that the biases run in opposite directions. That is exactly why the two sources are useful together. Filed data over-promises; measured data under-promises. The truth for a given address usually sits between them, and when the gap is enormous — "gigabit available" against an 18 Mbps measured median — that gap is itself the most informative thing on the page.

How to check an address honestly

1. Start with measured data. Put the address into SignalScout and read the median download, upload, latency, and test count from the Ookla Open Data tile covering that spot, plus the tiles around it.
2. Then get provider names from the FCC map. SignalScout does not list providers — the FCC's bulk availability data requires a registered FCC account and the map's per-address API blocks automated lookups, so provider lists remain a planned feature here rather than a current one.
3. Cross-examine the gap. If the map claims a technology the measured uploads contradict — fiber claimed, 12 Mbps upload measured — assume the claim needs proof.
4. Ask for a real order. The only definitive test is placing an order and getting an install date and a written construction quote. Availability that evaporates at that step is exactly what the FCC challenge process exists for.
5. Ask a neighbour. One honest answer from the house next door beats any dataset.

The map is still worth using

None of this makes the National Broadband Map useless. It is the only public source that tells you who might sell you service at a specific location, it is far better than the block-level data it replaced, and the challenge process gives it a self-correcting mechanism that no marketing page has. Just read it as a directory of claims rather than a record of performance.

If you want to go further on the mobile and ISP-map side, see how accurate ISP coverage maps are, and if you are trying to work out whether fiber is genuinely in your street, the fiber detection guide shows what the measured numbers give away.

Common questions