Wi-Fi vs Ethernet for speed tests: which one tells the truth?
If your speed test result is far below what your ISP promised, the bottleneck is almost always between your router and your device — not between your router and the internet. Here's how to tell the difference, and when each connection type is the right one to test on.
- Ethernet shows you what your ISP is actually delivering. Test there first.
- Wi-Fi shows you what your router is delivering to a specific spot in your home.
- If the two numbers are far apart, the line is fine and the router, the band, or the distance is the problem.
- 5 GHz Wi-Fi is roughly twice as fast as 2.4 GHz at close range and about the same at long range (because 5 GHz falls off faster through walls).
The two questions a speed test can answer
A speed test reads the throughput of one path: from the device you're using, through your home network, to a measurement server on the internet. Whether you're testing over Wi-Fi or Ethernet decides which part of that path you're measuring.
- Ethernet to the router → internet. This is the closest your device can get to the raw ISP line. If this number is below your plan, the problem is your ISP, your modem, or the cable from the street.
- Wi-Fi to the router → internet. This adds your router's radio, the wireless band you're using, distance, walls, interference from neighbours and other devices, and your laptop or phone's own Wi-Fi chipset.
Both are useful. The mistake is assuming a slow Wi-Fi test means your ISP plan is slow.
How much speed Wi-Fi actually loses
Even on a perfectly tuned modern setup, Wi-Fi rarely matches a wired connection. Here's a realistic look at what you can expect from each common Wi-Fi generation, on a 1 Gbps home plan, with a modern phone or laptop:
| Wi-Fi generation | Same room (close) | One wall away | Two walls / floor away |
|---|---|---|---|
| Wi-Fi 4 (802.11n, 2.4 GHz) | 40–80 Mbps | 20–50 Mbps | 5–25 Mbps |
| Wi-Fi 5 (802.11ac, 5 GHz) | 300–600 Mbps | 150–350 Mbps | 40–150 Mbps |
| Wi-Fi 6 (802.11ax) | 600–900 Mbps | 300–600 Mbps | 80–250 Mbps |
| Wi-Fi 6E / Wi-Fi 7 (6 GHz) | 800–1500 Mbps | 400–900 Mbps | 100–400 Mbps |
These numbers vary wildly with your specific router and device. A budget mesh node, a phone with a single antenna, or a laptop with an older Intel wireless card will all bring the top end down. Old smart-home devices (especially ones that only join 2.4 GHz networks) tend to drag the band down for everyone else on it.
When to test over Ethernet
Use Ethernet when you actually want to know what your ISP is delivering. That includes:
- Verifying a new install on the day it's switched on.
- Filing a complaint with your ISP or regulator (Ofcom in the UK, the FCC in the US, your national equivalent elsewhere). Most regulators will only act on wired test results.
- Working out whether a slow speed is the ISP's fault or your home network's fault.
- Testing whether an upgrade from, say, a 300 Mbps plan to a 1 Gbps plan would actually do anything for you. (If your Wi-Fi already caps below 300, paying for more bandwidth changes nothing.)
How to test over Ethernet, properly
- Plug your laptop directly into a free LAN port on your router with a Cat 5e (or better) cable. Cat 5 caps at 100 Mbps — anything labelled Cat 5e, Cat 6 or higher will do gigabit.
- Turn off Wi-Fi on the laptop. Otherwise some operating systems will quietly use both.
- Close other tabs, video calls, cloud-backup tools and downloads. Background traffic eats into the measurement.
- Run the test two or three times. Take the median, not the highest reading.
If the wired result is at or near your plan, your ISP is doing its job. If it's far below, that's worth a call to support — ideally with the screenshot saved.
Run a wired test now
Plug in, close the noisy tabs, and see what your line is really doing.
When to test over Wi-Fi
Use Wi-Fi when you want to know what a specific room or device actually gets. That's arguably the more useful number in daily life — you don't use the internet at the router; you use it on the sofa, in the bedroom, in the garden. So:
- Test in the spot where you'll actually use the connection.
- Test on the device you'll use (a phone Wi-Fi test isn't a laptop Wi-Fi test).
- Test at the time of day you care about — evening congestion is a real thing.
If the Wi-Fi result is well under the wired result, the action is on your side, not the ISP's. Common fixes:
Quick wins (free)
- Move the router off the floor and out of cupboards.
- Centralise it — not behind the TV in the corner.
- Separate the 2.4 and 5 GHz SSIDs so you can choose.
- Reboot the router monthly; it really does help.
- Move smart-home devices that only do 2.4 GHz off your main SSID.
Bigger fixes (cost something)
- Upgrade to a Wi-Fi 6 or Wi-Fi 6E router if yours is pre-2019.
- Add a wired-backhaul mesh node in the far end of the house.
- Run an Ethernet cable to a desktop access point (faster and steadier than mesh).
- Replace the cheap powerline adapter; modern ones are wildly better.
A note about "ISP-supplied" routers
The router that came in the box with your broadband install is engineered to a price. Most do 5 GHz Wi-Fi 5 or, on newer plans, basic Wi-Fi 6, with two or three antennas and modest CPU. They're fine for an average flat. They are commonly the bottleneck on gigabit and multi-gigabit plans, especially when the connection has to travel through brick, concrete, or floors.
If your ISP installed a 1 Gbps fibre line but you're seeing 200 Mbps in the bedroom, the line is doing what it should — the issue is the radio in the router. A modern third-party access point typically pays for itself in a few months of fewer buffering spirals.
The honest verdict
Both numbers matter, but they answer different questions. Test over Ethernet to argue with your ISP. Test over Wi-Fi to plan your next router upgrade. If you only ever look at one, look at the wired number — everything downstream of that is something you can fix.