Bandwidth is the maximum capacity of your connection, usually stated in Mbps or Gbps. Throughput is the amount of data that actually gets through in real conditions. A 1 Gbps line often delivers 600–900 Mbps of throughput once protocol overhead, Wi-Fi, and server limits are counted. Confusing the two is why people think their ISP is lying when it usually is not.
Here is the short mental model: bandwidth is the width of the pipe, throughput is how much water actually flows, and latency is how long a single drop takes to arrive. A wide pipe with a bottleneck somewhere else still moves less water than advertised.
Bandwidth vs Throughput vs Latency vs Jitter
These four numbers get mixed up constantly, and each answers a different question:
| Metric | What it answers | Unit | Measured by |
|---|---|---|---|
| Bandwidth | How much capacity the link has | Mbps / Gbps | Plan spec, link negotiation |
| Throughput | How much data really moves | Mbps | Download/upload test |
| Latency | How long one round trip takes | ms | Ping |
| Jitter | How much latency varies | ms | Repeated pings |
If the complaint is "big downloads are slow," you care about throughput. If the complaint is "the game lags," you care about latency and jitter. A link can have excellent bandwidth and still be unusable for gaming, which is why a single speed-test number does not describe network quality on its own.
Why Your Throughput Is Lower Than Your Bandwidth
Every one of these eats into the advertised number:
- Protocol overhead. TCP headers, acknowledgements, and framing consume a slice of every second. A few percent loss is normal and unavoidable.
- Wi-Fi. Wireless adds retransmissions, interference, and shared airtime. Going from Wi-Fi to Ethernet often recovers 20–40 percent of throughput by itself.
- Single-threaded tests. One TCP stream may not saturate a fast link. Multiple parallel streams usually push closer to the real ceiling.
- Test server distance and load. A far-away or busy server caps your result regardless of your line.
- CGNAT and evening congestion. If many neighbors share the same upstream, throughput drops at peak hours even though nothing in your home changed.
- Background traffic. Backups, updates, and cloud sync quietly compete for the same pipe.
How to Run a Bandwidth Throughput Test Properly
The test setup matters as much as the tool. Do these six things in order:
- Use wired Ethernet for the device running the test. This removes Wi-Fi from the equation.
- Pause other traffic — updates, backups, streaming boxes, and cloud drives.
- Pick a nearby test server and note its location, so you can repeat the same test later.
- Run three times, at different times of day, and record the median rather than the best run.
- Test both directions, because upload is frequently the bottleneck on asymmetric plans.
- Repeat with parallel streams if the tool exposes that option, to see whether a single stream is the limit.
IPIPAI's speed test reports download, upload, latency, and jitter together, which is enough to tell bandwidth starvation apart from latency problems. If results swing wildly between runs, read why speed test results keep jumping before blaming the ISP.
What Counts as Good Throughput
Use these rough bands as a sanity check, always measured wired and against a nearby server:
| Advertised plan | Healthy wired throughput | Typical Wi-Fi throughput |
|---|---|---|
| 100 Mbps | 90–98 Mbps | 50–85 Mbps |
| 500 Mbps | 450–490 Mbps | 250–420 Mbps |
| 1 Gbps | 850–950 Mbps | 400–750 Mbps |
| 2.5 Gbps | 2.1–2.4 Gbps | 800 Mbps–1.5 Gbps |
If wired throughput sits below roughly 80 percent of the plan across three tests, the next step is to isolate the segment: test at the router, then at the modem, then compare with a second device. That tells you whether the limit is inside your home or upstream. Multi-location testing helps here; see global multi-location speed tests.
ISP Bandwidth Test vs Internet Bandwidth Test
People use these phrases interchangeably, but they answer slightly different questions. An ISP bandwidth test measures the link between your home and your provider's network, ideally against a server the ISP operates or peers with directly. An internet bandwidth test measures end-to-end performance to a general server on the internet, which includes more routing and more variables.
If the ISP test is fast and the internet test is slow, your line is fine and the problem is routing or the remote server. If both are slow, the bottleneck is likely your line or local equipment. That distinction saves a lot of pointless support calls.
Network Quality Is More Than One Number
Bandwidth and throughput describe capacity and delivery. They say nothing about whether the connection feels responsive. For that you need latency and jitter, and for routing problems you need a path trace. Run a ping test for round-trip time and packet loss, and a traceroute to see where delays pile up. The combined picture is what "network quality testing" actually means.
Test the pipe, then test the path. Start with the IPIPAI speed test on Ethernet with everything else paused. If throughput looks healthy but the connection still feels slow, move to ping and traceroute before you change anything on your plan.
FAQ
What is the difference between bandwidth and throughput?
Bandwidth is the link's maximum capacity. Throughput is the delivered rate after overhead, interference, and server limits. Bandwidth is the pipe size; throughput is the measured flow.
Why is my throughput lower than my bandwidth?
Protocol overhead, Wi-Fi, single-stream tests, distant servers, evening CGNAT congestion, and background traffic all reduce the delivered rate.
Is a bandwidth throughput test the same as a speed test?
A speed test reports throughput plus latency. A controlled throughput test adds wired Ethernet, an idle network, parallel streams, and a nearby server so the result reflects the link rather than the setup.
What percentage of my plan should throughput reach?
Wired and local, 85–95 percent is healthy. Wi-Fi commonly reaches 50–80 percent depending on band, distance, and interference.