How to Compare Internet at Two Addresses
When two properties are otherwise close, connectivity is one of the few differences you can quantify before you commit. SignalScout has a compare mode for exactly this: two US addresses, side by side, with the measured numbers rather than the marketing ones.
What gets compared
Each address is geocoded through the US Census Geocoder, which returns the county it falls in, and that county is looked up in Measurement Lab's published statistics for 2024. For each side you get:
- Median measured download, upload and latency for the county
- The county each address resolved to, so you can see what is actually being compared
- A Connectivity Score from 0 to 100, which ranks that county against every other US county
- A dash rather than a score where a county has fewer than 300 measurements, because a guess would be worse than an admission
Weight the metrics to your actual life
A comparison is only useful if you know which column you care about. Most people default to download, which is usually the least discriminating number.
| What you do | Metric that decides it | What a meaningful gap looks like |
|---|---|---|
| Video calls all day | Upload, then latency | 12 Mbps vs 90 Mbps up is a technology difference, not noise |
| Competitive gaming | Latency | 20 ms vs 60 ms is felt; 20 ms vs 24 ms is not |
| 4K streaming, several TVs | Download | Below roughly 50 Mbps starts to matter; well above that rarely does |
| Uploading video, photos, backups | Upload | Every doubling halves your wait, so this scales linearly with your day |
| Remote desktop or VPN to an office | Latency, then upload | Consistency matters more than peak speed |
| Moving somewhere rural | Upload, then the measurement count | A county with fewer than 300 measurements is not reported at all |
Which differences are real
County medians are estimates built from a finite number of measurements, so small gaps mean nothing. Three rules of thumb:
- Check the measurement counts first. If one county rests on 40,000 measurements and the other on 400, you are not comparing like with like. The thin side could move a long way on a few hundred new results.
- Ignore differences under roughly 20% on download when both counties are already fast. 95 Mbps versus 110 Mbps tells you nothing useful about which house is better connected.
- Take upload gaps seriously. Download varies with the plans people bought. Upload varies with the technology in the ground. A large upload gap between two counties almost always means fiber is common in one and rare in the other, which is the most durable difference visible at this resolution.
Use the percentile scale for context. Across US counties, the measured median download sits at 26 Mbps at the 5th percentile, 60 at the 25th, 85 at the median, 110 at the 90th and 138 at the 99th. Moving from the 25th percentile to the 50th changes daily life. Moving from the 90th to the 95th does not.
When both addresses share a county
Two addresses in the same county return the same numbers, because county is the finest geography M-Lab publishes openly. That is not a limitation to work around so much as the answer itself: at this resolution there is no measured difference between the two houses, and anything you have heard about one street being better than the other has to be confirmed another way.
The comparison therefore earns its keep across county lines — a move to the next state, or a choice between a city county and a rural one. Within a single county, go straight to provider address checkers, the FCC map and the neighbours, which are the only sources that see one street at a time.
What the comparison cannot settle
Be clear about the boundaries before you make a decision on this alone.
- No provider names. SignalScout does not list which ISPs serve an address. The FCC's bulk availability dataset needs a registered account and its per-address API blocks automated requests, so provider lists are planned rather than present. Get names from the FCC National Broadband Map and each provider's own checker.
- No prices or plan tiers. An address with a slower median might have a cheaper gigabit option nobody on the street has bought.
- Data vintage. 2024, M-Lab's newest published county statistics. A build completed since then will not show.
- Medians reflect purchases, not possibilities. This is the single most important caveat, and it applies to both sides equally — which is why comparisons are more reliable than absolute readings.
A workable process
- Run both addresses in compare mode and note the county each address resolves to, along with download, upload, latency and score.
- Discard any metric where the gap is small or the sample is thin.
- Rank the remaining gaps by what your household actually does.
- Look up providers for both addresses on the FCC map, and read it with the scepticism the filed-versus-measured guide recommends.
- Share the comparison link with whoever you are deciding with — results are shareable, so the person who is not in the room can see the same numbers.
- Before signing anything, run through the pre-move checklist for the address you pick.
The goal is not to find the address with the biggest number. It is to avoid the one where the number that matters to you is quietly ten times worse.
Common questions
How do I compare two addresses in SignalScout?
Use the compare mode, enter both US addresses, and the tool geocodes each one and shows the measured county data side by side. You get the median download, upload and latency for each address's county, plus a Connectivity Score ranking that county against every other US county. If both addresses fall in the same county, both sides show the same figures.
What size difference in measured speed actually matters?
It depends on the metric. Download differences under about 20% between two already-fast counties are effectively noise, whereas a large upload gap usually reflects a real difference in the technology serving each street, and latency differences of 30 ms or more are noticeable on calls and in games.
Can I share a comparison with someone else?
Yes, results are shareable via a link, which is useful when a partner, agent or family member is not looking at the same screen. They see the same measured numbers rather than a summary you have paraphrased.
Does the comparison tell me which providers each address can get?
No. SignalScout does not currently list providers, because the FCC's provider-availability data requires a registered account for bulk download and its per-address lookup blocks automated queries. Pair the comparison with the FCC National Broadband Map to get provider names for each address.
Check a real address
See the measured download, upload and latency for any US address, plus a Connectivity Score ranked against every county in the country.