Your speed test results, explained
Four measurements, one clearer picture. What download, upload, latency, and jitter actually tell you.
Start with what you want to do
A speed test becomes useful when you connect its numbers to an everyday question. Can you make a clear video call while another person streams? Does a game feel slow because of capacity or response time? One headline speed does not answer every question.
Download is capacity coming in
Download speed describes how quickly data reaches your device from the test endpoint. Higher capacity can help with large files and several people streaming at once. It does not automatically make every website open instantly: small requests can be more sensitive to latency and server processing.
A reading in Mbps means megabits per second. File sizes are commonly shown in megabytes. Eight bits make one byte, so a 100 Mbps connection has a theoretical transfer ceiling of about 12.5 megabytes per second before protocol overhead and other constraints.
Upload matters more than it seems
Video calls, cloud backups, sending attachments, and streaming your own video all use upload. A connection with a fast download and a modest upload can work well for watching films but struggle when several devices are sending data at once.
Latency is the wait for a response
Our HTTP latency is the time a small request takes to reach the endpoint and return. It includes network travel, browser scheduling, and server work. A distant endpoint will often show higher latency even on a healthy connection. Compare like with like: the same device, endpoint, and approximate conditions.
Jitter is variation
Two connections can have the same average latency but feel different. If one alternates between quick and slow responses, calls and interactive applications may notice the inconsistency. Our jitter value averages the absolute change between successive idle round trips. It is an indicator from a short sample, not a complete analysis of every packet.
Watch what happens under load
Loaded latency samples the endpoint while data transfers are running. If it rises far above the idle figure, data may be waiting in queues. Pausing other uploads or comparing Ethernet can help you narrow down the issue. A busy server or route can also contribute.
Keep a small comparison diary
Run a test near the router, another in the room you use most, and one on a wired connection if possible. Repeat at a different time of day. The History tab stores recent results on this browser so that you can compare without creating an account.
For the exact formulas and activity thresholds, read our testing methodology.