- A speed test measures download speed, upload speed and latency at the moment it runs.
- A wired connection to the router gives the most accurate result by removing WiFi as a variable.
- Results vary with the time of day, so testing at peak and off-peak reveals the effect of contention.
- Closing other apps and pausing downloads during a test avoids skewing the result.
- Ofcom provides an official speed test, and other reputable tools are widely available.
A speed test measures download, upload and latency at that moment. For accuracy, test wired, close other apps, and try different times. Ofcom offers an official test, and results show how the line is performing.
Last reviewed: June 2026
What a speed test measures
A broadband speed test measures how a connection is performing at the moment it runs, reporting three main figures. Download speed is how fast data arrives, upload speed is how fast data is sent, and latency, often shown as ping, is the delay before data starts to move. Some tests also show jitter, the variation in latency. Together these give a snapshot of the connection's performance. Knowing what each figure means helps interpret the result and judge whether a connection is delivering what it should.
A single test is a snapshot, not a complete picture. Because performance varies with conditions, running several tests under good conditions, and at different times, gives a more reliable view than one reading taken in passing.
Why testing method matters
How a test is run has a large effect on the result, which is why method matters as much as the tool. A test taken over WiFi, with other devices active and apps running, can report a speed far below the line's actual capability, because the WiFi and the competing activity, rather than the line, are the limit. To measure the line itself, the test needs to remove these variables as far as possible. This is the difference between testing the broadband and testing the home network, and confusing the two leads to misleading conclusions.
| Factor | Effect on result | How to control it |
|---|---|---|
| WiFi versus wired | WiFi can understate the line | Test by ethernet |
| Other active devices | Compete for capacity | Pause other use |
| Time of day | Peak times reduce speed | Test off-peak and peak |
| Test server distance | Distant servers add latency | Use a nearby server |
| Device capability | Old devices cap results | Use a capable device |
Use a wired connection
The single most important step for an accurate test is to connect the testing device to the router with an ethernet cable rather than using WiFi. This removes the wireless link, which is often the real limit, and measures the line as delivered to the router. Where a device cannot use ethernet, testing close to the router on a strong 5 GHz signal is the next best option, but a wired test is the benchmark. Comparing a wired result with a WiFi result also reveals how much the home network, rather than the line, is contributing to any shortfall.
Control the conditions
Beyond using a cable, controlling the conditions improves accuracy. Closing other applications and pausing downloads, streaming and backups on all devices ensures the test measures the line rather than competing with household activity. Testing on one device at a time avoids other devices skewing the result. Restarting the router before testing can clear temporary issues. These steps create a clean measurement, so the figure reflects the connection's capability rather than the noise of a busy network.
Test at different times
Because consumer broadband is shared, speeds can vary with the time of day, dipping at peak times in the evening when many users are online. Testing at both a quiet time, such as late at night, and a busy time, such as a weekday evening, reveals how much contention affects the connection. A large gap between off-peak and peak results points to congestion rather than a fault. Recording results over several days at different times builds a fuller picture than a single reading.
Which tools to use
Several reputable speed test tools are available. Ofcom provides an official speed test as part of its consumer information, which is a trustworthy choice. Other widely used tools are also reliable, and many providers offer their own tests. Using a recognised tool, and ideally testing with more than one to compare, gives confidence in the result. Testing to a nearby server gives the lowest latency and the truest measure of the line, while testing to a distant server shows how far-away services add delay.
How to read the results
Interpreting the results is the final step. The download and upload figures can be compared with the package's advertised speed and, importantly, with the personalised minimum guaranteed speed the provider gave at sign-up. Latency indicates responsiveness, which matters for gaming and calls. If a wired test consistently falls well below the minimum guaranteed speed under good conditions, that suggests a problem worth raising with the provider. If the line speed is good but WiFi tests are poor, the home network is the issue rather than the broadband.
Acting on the results
A well-run speed test is the basis for action. Where a wired test under good conditions sits below the minimum guaranteed speed and the issue persists, reporting it to the provider allows the line to be investigated, and where the minimum is not met it may open the right to exit the contract. Where the line is fine but WiFi is the problem, the focus shifts to router placement, interference and devices. Keeping a record of test results, with the date, time and method, provides evidence that supports any complaint and helps track whether a problem is consistent.
Testing as an ongoing habit
Running occasional speed tests is a useful habit rather than a one-off. It confirms a new connection is performing as expected, helps diagnose slowdowns when they occur, and provides a record over time. Testing consistently, using a wired connection and controlled conditions, makes the results comparable and meaningful. With a clear method and a reliable tool, a household can monitor its connection confidently and distinguish genuine line problems from in-home factors, which is the foundation for getting issues resolved.
Frequently Asked Questions
Which broadband speed test is most accurate?
Accuracy depends as much on method as on the tool. Ofcom provides an official speed test, and other recognised tools are reliable too. The most accurate approach is a wired test to a nearby server, with other apps closed and downloads paused. Testing with more than one tool and comparing results gives added confidence.
Should I use WiFi or ethernet to run a speed test?
Ethernet gives the most accurate result, because it removes WiFi as a variable and measures the line as delivered to the router. WiFi can report a speed far below the line's capability. Comparing a wired result with a WiFi result also shows how much the home network, rather than the line, is contributing to any shortfall.
What time of day should I run a speed test?
Testing at both a quiet time, such as late at night, and a busy time, such as a weekday evening, reveals how much contention affects the connection. A large gap between off-peak and peak results points to congestion rather than a fault. Recording results at different times builds a fuller picture than a single reading.
Why does my speed test result vary?
Results vary because of the time of day and contention, the WiFi conditions, other devices and apps using the connection, the device used, and the test server chosen. Controlling these factors, by testing wired with other activity paused, gives a cleaner and more consistent measurement of the line's capability.
What do the numbers on a speed test mean?
Download speed is how fast data arrives, upload speed is how fast data is sent, and latency, or ping, is the delay before data moves. Some tests also show jitter, the variation in latency. Download and upload can be compared with the package and the minimum guaranteed speed, while latency indicates responsiveness for gaming and calls.
What should I do if my speed test is below the guaranteed speed?
If a wired test under good conditions consistently falls below the minimum guaranteed speed the provider gave at sign-up, and the issue persists, report it to the provider so the line can be investigated. Where the minimum is not met and cannot be fixed within a set period, it may open the right to exit the contract.
