Separate capacity from responsiveness
Download capacity matters when receiving streams, games and large files. Upload capacity matters for sending backups, video and work files. Latency describes response time to a destination. Jitter describes variation between sampled delays. A high download rate can coexist with poor calls if uploads saturate the connection or delay varies.
Start by listing activities that happen at the same time. Use each service’s published requirements rather than treating one universal plan size as a rule. For example, Netflix lists 3 Mbps for 720p, 5 Mbps for 1080p and 15 Mbps for 4K in its connection speed guidance. Those are service-specific recommendations, not a guarantee for every stream or household.
Build a demand estimate
Illustrative calculation, not observed customer data: two concurrent 4K streams at 15 Mbps each imply 30 Mbps of download demand before other traffic. Add a file download or another service and the demand changes. Leave headroom for variation instead of selecting a plan whose advertised maximum exactly equals the sum.
Do not add upload and download requirements as though they are the same direction. A large cloud backup may affect a call through upload contention even if the download direction is mostly idle. A plan with a large download number and a small upload allowance can still be a poor match for repeated large uploads.
Translate speed into time
At a constant 20 Mbps, a decimal 1 GB file takes at least 400 seconds: 1 × 8,000 ÷ 20. That is 6 minutes 40 seconds before added overhead or slowdown. Mbps means megabits per second; MB/s means megabytes per second. Divide Mbps by eight for the simple payload-rate conversion.
A server limit, Wi-Fi variation or pauses can make actual transfers slower. The transfer-time calculator estimates a constant-rate time and explains its assumptions. It does not run a new speed test.
Measure where the activity happens
- Run three sequential tests on the device and connection you normally use. Record time, room, wired or wireless, VPN and other traffic.
- Compare median and range, keeping failed attempts separately. One unusually high number is not a stable capacity estimate.
- Repeat near the router; compare Ethernet when available. Change only one condition at a time.
- Check the affected service’s own statistics. Its route and server may differ from this website.
Decide what to investigate before paying more
If wired and near-router tests are consistently better than tests in the problem room, investigate the wireless path before upgrading the plan. If multiple wired devices degrade at the same busy time, collect repeated observations and ask support to investigate; the pattern does not prove where congestion occurs.
If all household activities work comfortably, a larger advertised speed may provide little practical benefit. If a particular download is slow while other services work, the destination may be the constraint. Avoid buying hardware or a plan solely because one test displays a low number.
Work through the controlled comparison · Understand this test’s limits