Why Most Speed Tests Are Wrong

Run a speed test three times in one afternoon and you can get three different answers. That is not a broken tool — it is almost always a broken test setup. A laptop on Wi-Fi with a phone streaming video in the next room will never show your real connection speed, no matter how good your plan is. Before you call your ISP or upgrade your plan, make sure the number you are looking at actually reflects reality.

The 8 Most Common Speed Test Mistakes

These mistakes show up in almost every "my internet is slow" complaint. Fix them before you trust any result.

Testing over Wi-Fi instead of Ethernet. Wi-Fi has its own speed ceiling, interference, and signal loss on top of your actual internet connection. You end up measuring your router, not your ISP. Fix: plug your laptop or PC directly into the modem/router with an Ethernet cable and test again.
Other devices are active on the network. A console downloading an update, a smart TV streaming 4K, or a phone backing up photos in the background all eat into your available bandwidth. Fix: pause downloads and streams on every device, or test at a time when nothing else is connected.
A VPN is still turned on. VPNs route your traffic through an extra server and add encryption overhead, which lowers both speed and increases ping. Fix: disconnect the VPN completely before testing, then reconnect it afterward if you need it.
Too many open browser tabs and background apps. Cloud backup tools, auto-updaters, and even dozens of idle tabs can quietly use bandwidth and CPU. Fix: close unnecessary tabs and check your system's network usage panel for anything running in the background.
Testing only once. A single run can be skewed by a momentary network hiccup, a busy test server, or a brief Wi-Fi drop. Fix: run the test 3 or more times and compare the median result, not the highest or lowest.
Using the wrong test server. Some tools default to a distant server while a closer one is available, which understates your real speed and inflates ping. Fix: check which server the test used and, if possible, choose one physically closer to you.
Testing only during off-peak hours. A 3 AM test can look excellent even on a connection that slows to a crawl every evening when the neighborhood gets busy. Fix: also test during your actual usage window — typically 7 PM to 11 PM — since that is when congestion is most likely to show up.
Ignoring upload speed and jitter. Most people only look at the download number, but video calls, game hosting, and cloud backups depend on upload, and real-time apps depend on jitter more than raw speed. Fix: always check all four numbers — download, upload, ping, and jitter — not just one.

Bonus Basics: Mbps vs MBps

ISPs advertise speed in megabits per second (Mbps). File download managers and browsers usually show megabytes per second (MB/s). There are 8 bits in a byte, so a 100 Mbps plan downloads at roughly 12.5 MB/s. If your file transfer looks "8 times slower" than your advertised plan, this is almost always why — not a fault with your connection.

The Right Way to Test

A reliable test takes five extra minutes and saves you from chasing a problem that does not exist. Follow this sequence:

Go wired. Connect the test device directly to your router or modem with an Ethernet cable.
Get the network idle. Pause downloads, close streaming apps, and ask anyone else on the network to hold off for two minutes.
Run the test at least 3 times. Space the runs a minute or two apart and record each result rather than trusting the first one.
Test at different times of day. Run it once during the day and once during your typical peak usage hours in the evening.
Record everything. Note download, upload, ping, jitter, the test server used, and the time of each run. A simple table or notes app entry is enough — this record is what turns a vague complaint into evidence.

You can run this exact routine with a speed and ping test that shows the server location and all four core metrics on one screen.

What Each Metric Actually Means

Bandwidth is only part of the picture. Depending on what you use the internet for, a different metric matters most.

Metric	What it measures	Matters most for
Download speed	How fast data arrives from the internet to your device	Streaming, browsing, large downloads, game updates
Upload speed	How fast data leaves your device to the internet	Video calls, live streaming, cloud backup, hosting a game
Ping (latency)	Round-trip time for a single packet, in milliseconds	Online gaming, video calls, remote desktop
Jitter	Variation in ping between consecutive packets	Real-time voice/video quality, competitive gaming
Packet loss	Percentage of data packets that never arrive	Any real-time application — causes stutter and disconnects

A connection can have excellent download speed and still feel terrible for gaming or calls if ping, jitter, or packet loss are bad. Always read all five numbers together, not just the big download figure.

Why Different Sites Give Different Numbers

It is normal for two speed test tools to disagree, sometimes by a wide margin. The main causes are the physical location of the test server, how your ISP routes traffic to that server, and how many parallel connections the test opens to measure your line. A test that picks a server in your own city will almost always show a lower ping and often a higher speed than one that routes to a data center hundreds of kilometers away. For the full breakdown of why this happens and how to compare results fairly, see our guide on why speed test results differ between websites.

Is It Your Wi-Fi or Your ISP?

Before blaming your provider, rule out your own network. This takes two tests:

Test wired, at the router. Connect a laptop directly to the modem or router with a cable and run the test. This shows the actual speed your ISP is delivering, with your home network removed from the equation.
Test on Wi-Fi, in the room you actually use. Run the same test on the device and location where you normally notice problems — for example, a bedroom two floors from the router.

If the wired result is close to your plan but the Wi-Fi result is much lower, the problem is local: distance from the router, wall interference, an overloaded 2.4 GHz band, or outdated router hardware. See our Wi-Fi vs Ethernet comparison for what typically causes the gap and how to close it. If both the wired and wireless results are low and consistent, the bottleneck is upstream of your router, and it is time to look at your ISP.

What to Do If Speeds Are Consistently Below Your Plan

A single slow test proves nothing to an ISP. A pattern does. If your wired, idle-network results are repeatedly below what you pay for, build a case before you call:

Collect at least 5-10 test results spread across different days and times, all taken wired with the network idle.
Note the pattern. Is it consistently low, or only slow during specific hours? Evening-only slowdowns usually point to peak-time congestion; consistently low results any time of day point to a provisioning or line issue.
Include the test server and timestamp for each run — this heads off the "your test tool is wrong" response.
Call support with the data ready. State the plan speed you pay for, the average measured speed, and the number of tests, and ask them to check line signal levels and provisioning on their end rather than just restarting your modem.
Ask for a technician visit if the issue persists after a modem/router reset and the evidence still shows a shortfall — most providers are contractually required to deliver close to advertised speeds and will escalate faster with concrete numbers in hand.

Keeping a simple log turns "my internet feels slow" into a specific, provable claim — and that is what gets a real fix instead of a generic troubleshooting script.