Broadband review desk
Reading your speed test results
You ran the test. Now the screen shows four or five numbers, and at least one of them does not match what you were sold. Before you call anyone, it helps to know which figures describe your connection, which describe the trip your data took, and which describe the server on the other end. This page works through that comparison.
The number you compare against the plan
When a provider sells a 500 Mbps plan, it is describing the capacity available at your access point under good conditions, not a figure you should expect on every device at every hour. Wi-Fi overhead, in-home wiring, and the number of active devices all reduce what reaches a laptop. A wired test is the fairest comparison against the advertised number.
That is why the same line can report a comfortable figure on a desktop with an ethernet cable and a disappointing one on a phone across the house. Both readings are real. Only one of them describes what you are actually paying for.
Four readings, four different jobs
- Download
- How fast data arrives. It governs streaming and large file transfers, and it is the figure providers quote most.
- Upload
- How fast data leaves your home. Video calls, cloud backups, and file uploads all depend on it.
- Latency
- The round trip in milliseconds. It sets how quickly a page or a voice replies, regardless of bandwidth.
- Jitter
- How much that round trip varies from moment to moment. High jitter breaks calls even when the speed looks fine.
Advertised tiers are ceilings, not averages
The plan is a promise about capacity at the access point. Everything past the access point — the router, the walls, the number of screens competing for the same signal — is yours to manage. When you compare a measured result to a tier, compare the wired number first. If it lands near the plan, the service is doing what was sold and the rest of the gap is inside your home.
Wireless results are still worth recording. They tell you what people in the house actually experience, and they shift the moment you move the router or change its channel.
"A connection with plenty of bandwidth and unstable latency still produces a miserable call."
Latency and the video call problem
Latency is dominated by distance and by the number of networks a request crosses — not by how wide the pipe is.
Latency and the video call problem
Latency is the round trip in milliseconds, and it is dominated by distance and by the number of networks crossed. If your calls break up while downloads feel fine, latency and jitter are the numbers to look at, not download speed.
Picture a request leaving a home in Seattle and reaching a game server on the other side of the country. The distance adds time, and every additional network the request passes through adds a little more. A speed test result that shows a healthy download figure and a sluggish call is not a contradiction — it is two different measurements of two different things.
Jitter matters here too. A steady 30 ms round trip feels better on a call than a round trip that swings between 20 ms and 150 ms. When a provider changes routing or a path gets congested, jitter is often the first thing to move.
Why evening speeds drop
Residential networks are shared. Capacity is provisioned for a neighborhood, and when most of that neighborhood streams at once, each connection gets a smaller slice. This is normal engineering, not a malfunction.
The useful question is not whether the evening figure falls, but how far it falls and whether it stays above what you actually need for the things you do.
A simple way to see the pattern for yourself
-
1
Run the same wired speed test at a quiet hour, then again at the busiest hour you use the connection. Keep the device, the cable, and the server the same each time.
-
2
Note the drop, if any. A modest dip is normal; a collapse that persists night after night is worth a call.
-
3
Compare the evening figure against what your household actually uses — calls, streaming, backups, gaming — rather than against the plan number. If the evening figure still covers those uses, the connection is fine.
Upload matters more than it used to
Home connections were built around downloading, and many plans still reflect that. Remote work changed the balance. Video calls, cloud backups, and file uploads all depend on the upload figure. If your plan offers a large download and a small upload, you may be paying for capacity you do not use and missing the one you do.
Symmetric plans
Fibre plans often sell the same figure up and down. For a household that both consumes and publishes, that shape tends to fit the work.
Cable and DSL
These older technologies frequently offer far less upload than download. That gap is the reason a big-upload house feels slow on a fast plan.
What to watch
If the upload figure in your test sits well below your household's daily habits, ask the provider what is available in your area before assuming the plan is the problem.
Wi-Fi versus the wire
A wireless result includes everything between your device and the router: distance, walls, interference from neighboring networks, and the age of the radio. None of that is the provider's responsibility.
If wired results are close to plan and wireless results are not, the fix is in your home, and it may be as simple as moving the router or changing a channel.
The wire takes the wireless question out of the picture. If the wired number matches the plan and the wireless one does not, the wireless side is where the work is.
Comparing providers without fooling yourself
Comparisons go wrong when someone runs a speed test with one provider at nine in the morning and another at nine at night. Test the same way, at the same hours, with the same device and cable. Otherwise you are comparing time of day rather than service.
What to keep equal
Why it changes the answer
What to do instead
Time of day
A quiet-hour result and a busy-hour result describe two different neighborhoods of load.
Run both providers in the same window of the day, morning and evening.
Device and connection
A phone on Wi-Fi and a laptop on ethernet will not produce comparable figures.
Use one device and one cable for every test in the comparison.
Test server
Different servers sit at different distances and on different paths.
Pick the nearest server and keep it the same, or note which server each result used.
Number of runs
A single reading can land on a bad moment and mislead both sides.
Repeat across several days and compare the pattern, not one reading.
There is no shortcut here, and no provider ranking on this site pretends otherwise.
When the numbers do not add up
Two problems can look similar and have different owners. Sorting out which one you have saves a wasted call.
If the line is the suspect
If a wired speed test against multiple servers, repeated across several days, sits consistently below your tier, gather the log and contact your provider. Note the times, the servers, and the device. A complaint with a pattern behind it is easier to act on than a single disappointing reading.
If only one service is slow
If the numbers are fine but a specific service is slow, the problem is likely that service's servers or the path to them, not your line. Both are real problems; they have different owners. Your test tells you which door to knock on.
State programs work from many individual tests. One result is a personal answer; thousands together show where construction is needed.
What a public tool adds
State-level tools collect results across providers and turn them into an aggregated coverage map. Vermont's VT Speed Check, introduced October 6, 2026, is designed for exactly that: officials said the data will help identify areas that need stronger internet service support.
Your individual test helps you; thousands of them together help a state decide where to build. That is the practical reason to use a public tool even when your own connection looks healthy — the reading you leave behind becomes part of the picture regulators and providers both work from.
The desk wrote through the tool's rollout when it was announced, and keeps tracking how states use aggregated speed data when they plan construction.
Read the Vermont rollout storyCommon questions after a test
Why does my result change between runs on the same day?
A speed test result is a snapshot of one moment, one device, and one path. Demand on your neighborhood, the load on the test server, and the state of your own home network all move from minute to minute. That variation is the reason a single reading rarely settles a question — a pattern across several days does.
Which server should I choose in the test?
Pick the nearest server to your location and keep it for every run in a comparison. A distant server adds travel time to the request, which shows up as latency rather than as a fault on your line. If a provider's own server is available, it can still be useful, but note that you are measuring a shorter and friendlier path than a general one.
Does a good speed test result mean my connection is fine for everything?
Mostly, but not automatically. Bandwidth, upload, latency, and jitter describe different qualities, and a service can be limited by one of them while the others look healthy. If one kind of use is consistently poor while the test looks fine, look at the latency and jitter figures next, then at the service itself.
Should I test on Wi-Fi or on a cable?
Both, for different purposes. The wired test tells you what the provider is delivering, and it is the fair comparison against the plan. The wireless test tells you what people in the house actually get. If the two differ sharply, the gap is inside the home and worth working on.
Working through your own reading
Start with the wired number and compare it to the plan. Then look at upload, latency, and jitter against the way your household actually uses the connection. If the picture is still unclear, the reliable test guide and the policy coverage below are the next steps.
Re-run the test properly
A repeatable method for clean readings
Server choice, cable versus Wi-Fi, and the order of runs — the setup steps that make a result worth comparing.
Open the reliable test guideRead the FCC policy beat
The rules behind the measurements
Spectrum decisions, funding programs, and the policy moves that shape what providers can build and where.
Go to the policy sectionSubscribe for updates
New guides and policy coverage
Practical measurement writing and broadband policy reporting, sent when the desk publishes something new.
Sign up for the desk updatesQuestions about a reading you cannot explain?
The desk answers reader questions about speed test results, broadband comparison, and what a provider is responsible for. If your numbers still do not make sense after working through this page, the reliable test guide walks through the method step by step, and the newsroom is reachable on weekdays.