Summarise this article with:
What the Numbers Mean
A speed test result is a snapshot of your connection at one moment in time. Here is what each line is actually measuring.
Download: 285 Mbps
Upload: 11 Mbps
Ping: 14 ms
Jitter: 2 msDownload Speed
Download speed is the number most people care about, and it is measured in Megabits per second (Mbps). It tells you how fast data can flow from the internet to your device. Streaming a 4K Netflix show, loading a webpage, or pulling a software update all depend on download.
The test works by opening several parallel TCP connections to a nearby server and requesting large blocks of data. It measures the total bytes received across all streams, divided by elapsed time, and expresses that as Mbps.
| Download Speed | What it handles |
|---|---|
| Under 25 Mbps | Basic browsing, SD video, email |
| 25-100 Mbps | HD streaming, casual gaming, video calls |
| 100-300 Mbps | 4K streaming, multiple users, faster downloads |
| 300 Mbps+ | Heavy multi-device households, large file transfers |
The FCC raised its fixed broadband benchmark to 100 Mbps downstream in 2024, up from the old 25 Mbps floor set in 2015. That 100 Mbps figure is now the FCC's definition of "broadband."
Upload Speed
Upload speed governs anything you send outward: video call quality, live streaming on Twitch or YouTube, backing up photos to the cloud, or sending large email attachments. The test reverses the download process and sends data from your device to the server.
Most home internet plans are asymmetric. Cable connections often give you a fraction of your download speed on upload because the network is engineered around consumption, not creation. Fiber plans are more often symmetric, delivering similar upload and download figures.
| Upload Speed | What it handles |
|---|---|
| Under 5 Mbps | Basic video calls, email |
| 5-20 Mbps | HD video conferencing, routine cloud backups |
| 20-50 Mbps | Live streaming, fast large-file uploads |
| 50 Mbps+ | Professional content creation, symmetric work-from-home |
If you video call frequently and your calls look fine but others on the call say your image is choppy, upload is usually the culprit.
Ping (Latency)
Ping is the round-trip time in milliseconds for a tiny data packet to reach the test server and return. Lower is always better. Latency does not tell you how much data moves, it tells you how fast the network responds.
This matters most for real-time activities. Online gaming, video calls, and remote desktop sessions are all sensitive to latency. Streaming video is not, because players buffer ahead.
| Ping | Experience |
|---|---|
| Under 20 ms | Excellent: competitive gaming, tight video calls |
| 20-50 ms | Good: most gaming and calls work well |
| 50-100 ms | Acceptable: casual use, some gaming delay |
| 100-200 ms | Noticeable lag in games and calls |
| 200 ms+ | Problematic for any real-time use |
You can probe latency in more detail with the ping tool, which sends repeated packets and shows you the distribution.
Jitter
Jitter is the variation in ping from packet to packet. Where latency measures the average trip time, jitter measures the inconsistency. If one packet arrives in 14 ms and the next in 55 ms and the one after in 12 ms, that swing is jitter.
A low average ping with high jitter still produces a poor experience. Voice and video calls depend on a steady, predictable stream. High jitter causes choppy audio and frozen video frames even when average latency looks fine.
In my testing, jitter under 5 ms on a wired connection is typical. Wi-Fi can push that into the 20-40 ms range depending on interference.
| Jitter | What to expect |
|---|---|
| Under 5 ms | Stable: calls and gaming are smooth |
| 5-20 ms | Acceptable for most uses |
| 20 ms+ | Expect choppy calls and gaming hiccups |
How the Test Actually Works
When you start a speed test, the sequence is:
- Server selection. The test locates a server near you, usually by pinging several candidates and choosing the lowest latency one. Distance introduces delay; testing against a server in another country will show lower speeds than your actual connection supports.
- Latency and jitter measurement. Several small packets travel to the server and back. The test records the round-trip times and calculates average latency (ping) and variation (jitter).
- Download test. Multiple parallel TCP connections open and request large data payloads. The test runs for a fixed duration (typically around 10-15 seconds) and calculates throughput across all streams.
- Upload test. The same parallel-connection approach runs in reverse: your device sends data to the server.
Using multiple connections is intentional. A single TCP stream can be limited by the protocol's congestion window behavior, so parallel streams reveal the connection's true capacity.
Why Results Vary
Run the same test twice and you may see different numbers. Several things cause this.
Server Location
Testing against a far-away server always shows lower performance than testing against a nearby one. This is physics: more network hops, more potential for congestion, more time in transit. Always test against the server with the lowest ping first. Most test tools auto-select it, but you can manually pick servers to compare.
Wi-Fi vs. Ethernet
A wired Ethernet connection removes several variables. Wi-Fi performance depends on distance from the router, walls and physical obstructions, interference from neighboring networks and household devices, and the number of devices connected simultaneously. For a fair baseline, plug in an Ethernet cable when testing.
Time of Day
Network congestion follows human schedules. Evening hours and weekends typically see more shared traffic, especially on cable connections where nearby homes share the same network segment.
Background Activity
Software updates, cloud sync, and other devices streaming video all consume bandwidth during a test. Pause them to get a clean reading.
Are You Getting What You Pay For?
ISPs advertise "up to" speeds, not guaranteed speeds. FCC data from its Measuring Broadband America reports indicates that most major cable and fiber providers deliver measured speeds that meet or come close to advertised tiers. DSL is the notable exception: performance depends heavily on your physical distance from the provider's distribution point, and speeds degrade significantly over longer copper runs.
If your results consistently fall well below your plan's speed, try this sequence before calling your ISP:
- Test with a wired Ethernet connection directly into your modem, bypassing the router.
- If the wired result is also low, the problem is likely upstream of your router.
- If wired speeds are fine but Wi-Fi is slow, the router is the issue.
- Test at different times of day to rule out peak-hour congestion.
The Honest Story on ISP Throttling
Throttling is real and distinct from congestion. ISPs can use deep packet inspection to identify specific traffic types and slow them selectively, for example, capping video streaming or peer-to-peer traffic without touching other connections. This means a standard speed test can look fine while Netflix performs poorly, because the test traffic does not look like streaming video to the ISP's network equipment.
A practical detection method: run a speed test normally, then run the same test while connected to a VPN. A VPN encrypts your traffic so the ISP's inspection cannot classify it by type. If speeds increase substantially with the VPN active, selective throttling is a likely cause. If speeds drop with a VPN (which is common due to encryption overhead), that does not confirm throttling.
Net neutrality rules at the federal level were effectively vacated following a 2025 court ruling, so ISPs currently have wide legal latitude in the US. State-level rules vary.
Common Problems and What They Indicate
Good download, slow upload
This is often normal on cable plans, which are engineered asymmetrically. If upload matters to you (remote work, streaming), check whether your ISP offers a symmetric fiber option.
Slow on all devices, fast on wired test
The router or Wi-Fi environment is the bottleneck. Try rebooting the router, checking for Wi-Fi interference, or moving devices closer to the access point.
Good speeds, poor call or gaming quality
Look at ping and jitter, not the raw speed numbers. High jitter on an otherwise fast connection explains choppy calls and lag spikes. Ethernet typically reduces jitter substantially compared to Wi-Fi.
Speeds vary wildly across tests
Run the test at different hours and note which times are slow. Consistent evening slowdowns on a cable connection point to neighborhood congestion, which is an ISP infrastructure issue.
FAQ
Why is my upload speed so much lower than my download speed?
Most cable and DSL plans are deliberately asymmetric. The network infrastructure for these technologies was designed when households consumed far more data than they produced. Cable's DOCSIS standard allocates more frequency channels to downstream traffic. Fiber-optic plans are more commonly symmetric. If your plan is listed as, for example, "500/20," the first number is download and the second is upload, so the gap is expected.
Does using a VPN affect my speed test results?
Yes. A VPN routes your traffic through an additional server and adds encryption overhead, both of which reduce throughput. Expect VPN speeds to be lower than your raw connection. However, if a VPN makes your test faster than without it, that is a reliable signal that your ISP is throttling certain traffic types, since the VPN prevents the ISP from classifying your traffic.
Should I use Wi-Fi or Ethernet for a speed test?
Use Ethernet for the most accurate result. Wi-Fi introduces variables (signal strength, interference, channel congestion) that can make your results lower than your actual line speed. A Wi-Fi test tells you about your wireless setup, not your ISP's connection. Test both and compare to see how much your Wi-Fi environment costs you.
Why does my speed test show 500 Mbps but videos still buffer?
Speed and latency are different dimensions of performance. A high-bandwidth connection with high latency or high jitter can still produce buffering, especially on a Wi-Fi connection with variable signal quality. Also check whether the buffering happens on a specific service: a platform's servers or your path to them may be the constraint, not your home connection. The ping tool can help you check latency to specific destinations.
What is a good speed for a household with multiple people working from home?
Video calls require roughly 3-5 Mbps per person for HD quality, more if using 1080p or 4K webcams. Add background cloud sync, software updates, and streaming on other devices. A household with two people on video calls simultaneously, plus streaming, should have at least 100 Mbps download and 20-30 Mbps upload as a practical floor. Fiber plans at 300-500 Mbps symmetric remove headroom concerns for most households. The FCC's current minimum broadband benchmark is 100/20 Mbps.
Related Tools
- Run a speed test to check your current download, upload, ping, and jitter.
- Ping tool to measure latency to specific servers in detail.
- DNS Lookup to check whether slow DNS resolution is adding to your page load times.
- Traceroute to trace the network path and locate where slowdowns occur.
- IP Geolocation accuracy if your speed test's auto-selected server seems geographically wrong.
Sources
WhatIsMyLocation Team
Our team of network engineers and web developers builds and maintains 25+ free networking and location tools used by thousands of users every month. Every article is reviewed for technical accuracy using real-world testing with our own tools.
Related Articles
Try Our Location Tools
Find your IP address, GPS coordinates, and more with our free tools.