A broadband speed test measures three key things about your connection: download speed, how fast data comes in; upload speed, how fast it goes out; and latency, the short delay before data moves. Running one is simple, but how you run it affects the result, so a few precautions make the difference between a reading that reflects your connection and one skewed by your Wi-Fi or other devices. The most accurate test is done on a device connected directly to your router with a cable, with little else using the connection.
The short answer is to test under clean conditions and interpret the numbers in context. A single test is a snapshot, so running a few at different times gives a fuller picture. Comparing the results with your deal, including any average download speed, helps you tell whether your connection is performing as expected or whether something is holding it back.
This guide explains how to run an accurate speed test, why a wired test matters, and how to interpret download, upload and latency. Because conditions vary, treat results as a guide and confirm with your provider if something looks consistently wrong.
How do you run an accurate speed test?
Accuracy comes from removing the variables that distort a reading. The most important step is to test on a device connected directly to your router with a cable, since Wi-Fi introduces variability that can understate your connection's true speed. If you must test over Wi-Fi, stay close to the router with a clear line of sight.
It also helps to pause other activity during the test. Streaming, downloads and other devices share the connection, so a test run while the network is busy will show lower speeds. Closing other applications and asking others in the home to pause heavy use gives a cleaner result. Running the test a few times, and at different times of day, accounts for normal variation.
These precautions matter because a careless test can mislead you. A reading taken over weak Wi-Fi during a busy evening might look alarming when the connection itself is fine. A clean, wired test shows what your connection is actually delivering.
What do the three measurements mean?
Each number a speed test reports tells you something different. Understanding them helps you interpret the result rather than fixating on a single figure. The table below summarises the three measurements.
| Measurement | What it shows | Why it matters |
|---|---|---|
| Download speed | How fast data comes into your home | Affects streaming and loading |
| Upload speed | How fast data goes out | Affects calls, backups and sharing |
| Latency | The delay before data moves | Affects responsiveness and gaming |
The table shows why all three matter. A high download speed alone does not guarantee a good experience if the upload is weak or latency is high. For working from home, the upload figure is important; for gaming, latency is. Reading the three together gives a fuller sense of how your connection performs for what you do.
How do you interpret the results?
Once you have a clean reading, compare it with your deal and your expectations. On a wired test with little else active, your download and upload should be reasonably close to the deal's figures, though a modest shortfall is normal, and comparing against any average download speed is realistic for busy hours. If the numbers are far below your deal on a wired test, that suggests an issue worth investigating.
Latency is interpreted differently, since lower is better and there is no single target; what matters is whether it is low and stable enough for your activities. For everyday use, ordinary latency is fine, while gaming and video calls benefit from low, consistent figures. A test that shows good download and upload but high latency can still feel sluggish for real-time tasks.
Because a single test is a snapshot, look at several results before drawing conclusions. Consistent readings that match your deal suggest a healthy connection, while consistent shortfalls, especially wired, point to something to address with your provider or your equipment.
What can throw off a speed test?
Several things can distort a result, and knowing them helps you avoid false alarms. Testing over Wi-Fi, especially far from the router, often understates your connection because of signal loss. Other devices using the connection during the test divide the bandwidth and lower the reading. An older device or router may not be capable of your deal's full speed, capping the result.
The time of day matters too, since peak-hour congestion can reduce speeds temporarily, which is why an average download speed is lower than a maximum. A single low reading during a busy evening is not necessarily a problem. Running the test under clean conditions, wired, with little else active, removes most of these distortions and gives a result you can trust.
If results remain low under clean conditions across several tests, that is the signal to look further, either at your equipment or by confirming with your provider. Otherwise, a clean test usually reflects your connection accurately.
Frequently asked questions
Why should I test with a wired connection?
A wired connection removes the variability that Wi-Fi introduces, such as signal loss from distance and obstacles. This shows what your connection itself is delivering, rather than what your Wi-Fi can carry to a particular spot, giving a more accurate reading.
How many times should I run a speed test?
Run it a few times, and at different times of day, since a single test is only a snapshot. Multiple readings account for normal variation and peak-hour congestion, so consistent results give a more reliable picture than one test alone.
What is a good latency figure?
Lower is better, and what counts as good depends on the activity. Everyday browsing tolerates more latency, while gaming and video calls benefit from low, stable figures. There is no single target, so focus on whether it is low and consistent enough for what you do.
My speed test is lower than my deal. Is that a problem?
A modest shortfall is normal, especially over Wi-Fi or during busy times, and comparing against any average download speed is realistic. A large, consistent shortfall on a wired test with few devices active is more likely a problem worth investigating with your equipment or provider.
Conclusion
A broadband speed test measures download speed, upload speed and latency, and running it on a wired device with little else active gives the most accurate reading. Interpreting the results means comparing download and upload with your deal, including any average download speed, judging latency by whether it is low and stable, and looking at several tests rather than one. Wi-Fi, busy networks and older equipment can all distort a result, so clean conditions matter. If readings stay low under clean, wired conditions, investigate further with your equipment or provider.