Internet Speed Test Results Explained: What Is Actually Good?
⚡ Quick Answer
A speed test reports three numbers: ping (delay, in milliseconds), download (data coming in, in Mbps) and upload (data going out). For everyday use, a ping under 30 ms is excellent, 25 Mbps download comfortably covers a 4K stream, and 5 Mbps upload is the sensible floor for one HD video call. Expect your result to sit below the advertised package speed — WiFi and network overhead always eat some — but if you're consistently getting less than half of what you pay for on Ethernet, that's a genuine ISP conversation.
The Three Numbers, Properly Defined
Run the TestMyPC internet speed test and you'll get three results. Most people glance at the biggest number and ignore the rest — which is exactly backwards for many of the problems that send people to a speed test in the first place.
Ping is the round-trip time for a tiny message to travel from your browser to the test server and back, measured in milliseconds. It's not a measure of how much data can move — it's a measure of how quickly the connection responds. Ping is what makes a video call feel natural or awkward, a game feel crisp or laggy, and a remote desktop feel local or treacly. A gigabit connection with 200 ms ping still feels sluggish for anything interactive.
Download is how fast data arrives, in megabits per second (Mbps). The test requests timed chunks of data for roughly ten seconds and calculates how much moved. Download governs streaming quality, page loads, game downloads and file transfers — it's the number ISPs advertise, because it's the biggest one.
Upload is how fast data leaves your machine. On most UK home broadband it's a fraction of the download figure, because providers deliberately allocate the line asymmetrically. Upload is what carries your face and voice in a meeting, your files to cloud backup, and your attachments out of your inbox. When calls freeze while everything else feels fine, upload is the first suspect — we cover that scenario in depth in our upload speed fix guide.
One unit trap worth knowing: speeds are measured in megabits (Mb), file sizes in megabytes (MB). There are eight bits in a byte, so a 100 Mbps connection moves about 12.5 MB of file per second. That's why a "fast" connection still takes over a minute to fetch a 1 GB download — nothing is wrong.
What Counts as "Good"? Thresholds by Activity
There's no single good speed — only enough speed for what you're doing. These figures are per activity, happening at once; two 4K streams need double the 4K row.
| Activity | Download | Upload | Ping |
|---|---|---|---|
| Browsing & email | 5–10 Mbps | 1 Mbps | Under 100 ms |
| HD (1080p) streaming | 5–10 Mbps per stream | Minimal | Not critical |
| 4K streaming | 25 Mbps per stream | Minimal | Not critical |
| Online gaming | 10–25 Mbps | 3 Mbps | Under 50 ms |
| One HD video call | 3–5 Mbps | 5 Mbps floor | Under 80 ms |
| Several simultaneous calls | 10+ Mbps | 10+ Mbps | Under 80 ms |
| Cloud backup of large files | Minimal | 20+ Mbps to be painless | Not critical |
Notice the pattern: streaming and downloading live entirely in the download column, while calls and backups live in the upload column, and gaming barely touches either but is brutally sensitive to ping. Diagnose against the column your problem actually lives in. A household that streams 4K on two TVs while someone works on video calls upstairs needs roughly 60 Mbps down and 10 Mbps up before anything else on the network gets a look-in — well within a mid-tier fibre package, but beyond many entry-level ones.
Ping bands are worth memorising: under 30 ms is excellent for everything including competitive gaming; 30–80 ms is fine for everyday work and calls; 80–120 ms is where video calls start to feel slightly out of step; over 120 ms feels noticeably laggy for anything interactive.
Free Tool
📶 Test Your Connection Now
Ping, download and upload in about thirty seconds — no downloads, no account.
Run Speed Test →Why Browser Tests and ISP Tests Disagree
Run our test and your ISP's test back to back and the ISP's number will usually be higher. Neither is lying — they're measuring different journeys.
An ISP's speed checker typically talks to a server inside the provider's own network, a short, uncongested hop away. It answers the question "what is the line into this property capable of?" A browser test crosses the open internet the way your real traffic does — through your WiFi, your browser's networking stack, any VPN, and the public routes to the server. It answers a different and often more useful question: "what do web apps, video calls and cloud services actually get on this device, right now?"
That's why a browser result reading a little lower is expected and not a fault. The interesting case is when the two disagree wildly — an ISP test showing 70 Mbps while the browser sees 8. That gap lives somewhere inside your home or your device: WiFi conditions, a VPN, an overloaded router, or a busy machine. If it's one specific laptop underperforming, work through our one-slow-laptop triage guide.
Why You Never Get the Advertised Number
UK providers advertise an "average" speed measured at peak time, but even that figure describes the line at the router — not what reaches your screen. Several layers each take a slice:
WiFi overhead. Wireless is a shared, half-duplex medium with retransmissions, encryption overhead and protocol chatter. Even a healthy 5 GHz connection typically delivers 50–70% of what the same router manages over Ethernet, and a 2.4 GHz connection through two walls can deliver far less. This is the single biggest gap for most homes.
Contention. Your line shares upstream capacity with neighbours, which is why evening results dip on some connections. Inside the home, every phone syncing photos and every console updating a game is contending with your test.
Device and browser limits. An older laptop's WiFi adapter, a heavy antivirus inspecting traffic, or a VPN can each cap throughput regardless of how fast the line is.
The practical implication: always judge your ISP on an Ethernet test next to the router, and judge your WiFi on the difference between that result and the one from where you normally sit. Run the test two or three times in each position — a single run can be skewed by a background update.
👤 Example scenario
A remote worker on a "150 Mbps" package measures 22 Mbps at their desk and assumes the ISP is short-changing them. Testing again next to the router shows 141 Mbps — the line is fine, and the real problem is a 2.4 GHz WiFi connection travelling through a chimney breast. Switching the laptop to the 5 GHz band brings the desk up to 96 Mbps, and no ISP call was ever needed.
When the Gap Justifies an ISP Complaint
Complain when the problem survives your best-case test. The evidence that carries weight is: wired to the router by Ethernet, WiFi ruled out, VPN off, other devices quiet — and the result still sits well below the advertised speed, consistently, across several days and times.
As a rule of thumb, under half the advertised speed on Ethernet at off-peak times is worth a formal complaint; most major UK providers are signed up to Ofcom's automatic compensation scheme and minimum-speed guarantees, which typically let you exit the contract penalty-free if they can't fix a confirmed shortfall within 30 days. Keep a simple log — date, time, wired result — because a pattern of ten dated readings is far harder to wave away than "it feels slow". Wildly unstable readings between runs, or ping that regularly spikes above 120 ms on a wired connection, are also legitimate faults worth reporting even when the headline download number looks respectable.
What doesn't justify a complaint: a WiFi result lower than the package speed (that gap is yours to fix), or a browser test reading modestly below a wired ISP test (that's measurement context, not a fault).