You ran a speed test on your phone and got 45 Mbps. Your provider says you're getting 80 Mbps. One of them is right. The answer depends on how you ran the test and what was happening on your network at the time.
Speed tests are useful, but they're not infallible. A single result on a Wi-Fi-connected phone tells you almost nothing about the connection speed at your router. Getting a meaningful reading takes about two minutes more effort than most people put in.
Why broadband speed test results vary
Five things affect your speed test result more than your actual broadband connection.
Wi-Fi versus ethernet
This is the biggest one. Wi-Fi is shared, contended, and affected by walls, interference, and distance from your router. A phone in the kitchen might be getting 45 Mbps from a router delivering 200 Mbps to an ethernet-connected device in the same room.
If you want to know your actual broadband speed, plug a laptop or desktop directly into your router with an ethernet cable and test from there. No Wi-Fi in the chain. The result you get then is your line speed.
If you don't have an ethernet cable, the result you get on Wi-Fi is still useful. It tells you the speed your device can access. But it's not the same as your line speed.
Time of day
Broadband networks are shared between all the users in your area. Peak hours, typically 7pm to 10pm on weekdays, mean more people competing for the same capacity. FTTC connections slow down during peak hours more than FTTP, because the copper portion of the network is shared.
Ofcom's Connected Nations research tracks this evening slowdown. Some FTTC users see speeds drop by 20 to 30% between 8pm and 9pm compared to a midday test. If your provider promised "up to 80 Mbps" and you're testing at 8pm, a lower result is expected.
Test at different times of day to get a proper picture. A midday test and a 9pm test together will show you whether you have a general speed problem or a peak-time congestion issue.
Speed test server distance
Speed test tools like Ookla Speedtest and Fast.com connect to the nearest available test server. "Nearest" sometimes means London for a test run from Edinburgh. The further the server, the higher the latency and the greater the chance of a result that doesn't reflect your actual connection.
When running a test, check which server it's connecting to and try changing it if the auto-selected one seems far away. Ookla lets you pick your server manually before running.
Your device's capabilities
An older laptop running on a 5-year-old Wi-Fi card might not be capable of receiving more than 100 Mbps over Wi-Fi, even if your connection can deliver 300 Mbps. The device becomes the bottleneck.
Similarly, a phone processing other apps in the background during a speed test will show lower results. Close other apps, put the phone down flat (antenna position matters on mobiles), and run the test again.
Background activity
Automatic updates, cloud backups, streaming devices, and other computers on your network all compete for bandwidth during a test. Before running a test, disconnect other devices from Wi-Fi, pause any downloads or cloud sync, and make sure no one else in the house is streaming.
This sounds excessive. It takes 30 seconds. And the difference in results can be dramatic. We've seen households lose 40% of available bandwidth because someone's laptop was quietly downloading a Windows update in another room.
How to get an accurate speed test result
Follow this order for a result you can actually trust:
- Connect your laptop or desktop directly to the router via ethernet cable.
- Disconnect all other devices from Wi-Fi temporarily.
- Pause any background downloads, updates, or cloud backups.
- Open a browser and go to a speed test. Our speed test tool runs directly from your browser.
- Run the test twice and take the average.
- Note the time. Repeat the test at 8pm to compare.
The download result from an ethernet-connected device during off-peak hours is the closest you'll get to your actual line speed.
What the numbers actually mean
Download speed is how fast data comes to your device. Streaming, browsing, downloading files. Most people focus on this number.
Upload speed is how fast data leaves your device. Video calls, file uploads, cloud backups. Usually much lower than download on FTTC. Important if you work from home. See our guide on broadband for working from home for more on why upload matters.
Ping (latency) is how long it takes a signal to travel to a server and back. Measured in milliseconds. Matters most for gaming and video calls. Under 20ms is excellent. Over 60ms starts to cause problems in real-time applications.
Jitter is the variation in ping over time. A consistent 30ms ping is better than a ping that varies between 10ms and 80ms. High jitter causes choppy video calls and lag spikes in online games.
What to do if your speed is genuinely low
If your ethernet-connected speed test comes back well below what your provider promised, you have grounds to complain.
Ofcom's voluntary code of practice for broadband speeds means most major providers must offer an exit from the contract if they can't deliver speeds within the range they guaranteed at sign-up. Keep records of your test results with timestamps.
Before contacting your provider, try restarting your router (unplug for 30 seconds, plug back in). A significant number of support calls are resolved by a simple router restart. If that doesn't help, ask your provider to run a line check. They can often spot faults remotely that don't show up on your side.
If speeds are consistently low on a FTTC connection, it may be time to switch to a full fibre provider. Check what's available at your postcode with our broadband comparison tool. Our guide to UK broadband speeds covers what you should expect from each technology type.