What Wi-Fi Actually Costs You

Published August 29, 2026 10 min read

Wi-Fi rarely costs you bandwidth. It costs you consistency — and that is what breaks calls and games. How to measure the real difference on your own connection.

"Try a cable" is the most-given and least-followed piece of network advice, partly because the reason behind it is usually left unsaid.

It is not about speed. Modern Wi-Fi carries more bandwidth than most internet connections can supply, and if all you do is stream video or download files, wireless is genuinely fine. The problem appears only when an application needs packets to arrive on time, every time — calls, video meetings, competitive games, remote desktop.

That is the property Wi-Fi trades away, and it is worth knowing how much yours gives up before deciding whether it matters.

Measure the difference on your own connection
Run the connection test twice — once on Wi-Fi, once on Ethernet. Compare loss and jitter, not the speed figure.
Run the connection test →

Why wireless behaves differently

Three properties of the medium, none of which improve with a better router.

Airtime is shared. Devices on a channel contend for the chance to transmit rather than sending whenever they like. Wi-Fi 6 and later can serve several clients within one transmission opportunity using OFDMA, which helps considerably in dense settings, but it schedules contention more efficiently rather than removing it. Every device — yours, your household's, and every neighbour on the same channel — waits its turn. That wait is not constant, so delivery timing varies. This is jitter, produced structurally rather than by any fault.

Frames are retransmitted invisibly. When a wireless frame is not acknowledged, the adapter resends it below the application's awareness. If a retry succeeds, the packet arrives later than its neighbours and the application sees timing variation rather than loss; if the retry limit is reached first, it is dropped after all. Either way the delay is invisible from above.

Interference is everywhere and mostly invisible. Neighbouring networks, microwaves, Bluetooth, baby monitors, USB 3 enclosures. On 2.4 GHz especially, you are sharing with things that are not Wi-Fi at all.

The consequence: signal strength does not predict quality. Full bars means frames get through. It says nothing about how long they waited for airtime. A laptop beside the router on a congested channel can perform worse than one two rooms away on a clear one.

An example measurement

Two measurements on the same connection minutes apart:

Ethernet        loss 0.00%    jitter  1.2 ms    latency  14 ms
Wi-Fi (5 GHz)   loss 0.00%    jitter  8.7 ms    latency  17 ms
Wi-Fi (2.4 GHz) loss 0.30%    jitter 34.0 ms    latency  22 ms

The throughput figure barely moves across all three. Everything that matters for real-time traffic does.

These are illustrative, not a promise — your own numbers depend entirely on your environment, which is exactly why measuring beats assuming. A well-placed 5 GHz link in a quiet area can be nearly indistinguishable from a cable; a 2.4 GHz link in a dense apartment building can be far worse than the figures above.

How to test Wi-Fi vs Ethernet

The comparison only means something if you change one variable at a time.

1. Run the connection test on Wi-Fi. Note loss, jitter and latency. Ignore throughput — it is not the point.

2. Plug in a cable and run the identical test. On a laptop, disable the wireless adapter rather than trusting the OS to prefer Ethernet; many will keep using Wi-Fi if it is still associated.

3. Compare loss and jitter.

Reading the result:

Jitter drops sharply on the cable → Wi-Fi is a major contributor. Everything below applies.
Both are similar → your Wi-Fi is fine and the problem is upstream. Look at bufferbloat or the path itself.
Both are bad → Wi-Fi is not the cause, but that does not mean the problem is outside your house. A cable removes the wireless layer and nothing else: the router, the modem or ONT, a switch, the cable itself and the network adapter are all still in the path. Rule those out before documenting it for your ISP.

For a fuller picture, repeat each test while a large upload runs. Wi-Fi and bufferbloat compound, and a connection that looks acceptable idle can fall apart under load on wireless while holding up on a cable.

What your two results mean

On Wi-Fi	On Ethernet	What it points to
High jitter or loss	Clean	Wireless contention or interference — the rest of this page applies
Clean	Clean	Wi-Fi is not your problem; look further up the path
Bad	Bad	Not wireless, but not necessarily the ISP either — router, ONT, switch, cable and adapter are all still in play
Fine idle, bad under load	Fine idle, bad under load	Bufferbloat in the router, not the medium
Fine idle, bad under load	Clean under load	Wireless airtime contention under load
High latency	Equally high latency	Distance or routing, not your network — see what causes packet loss

The second row is worth dwelling on: a clean result on both is a genuinely useful outcome. It rules out your entire local network in one test, which is exactly what an ISP will ask you to do first.

Making Wi-Fi better when a cable is impossible

Realistically, most people cannot run Ethernet everywhere. In descending order of what actually helps:

Prefer 5 GHz or 6 GHz where coverage allows. Usually the largest single improvement. 2.4 GHz has three non-overlapping channels and shares spectrum with a great deal of non-Wi-Fi equipment; 5 GHz has far more room and much less contention. The trade is range and wall penetration.

Pick a clear channel. Most routers default to auto-selection and many choose badly. A Wi-Fi analyser app will show which channels your neighbours occupy. On 5 GHz, prefer whichever channels your neighbours are not using. DFS channels are often the emptiest, but they carry a trade: the access point must monitor for radar and is required to vacate the channel if it detects any, which causes a brief interruption. Not every client supports them either. Worth trying, worth reverting if you see unexplained drops.

Move the access point, not the device. A router in a cupboard, behind a TV, or on the floor performs poorly regardless of its specification. Central, elevated, and in the open beats an expensive router badly placed.

Reduce channel width if the band is crowded. Wider channels raise peak throughput and occupy more spectrum, which makes them harder to keep clean where neighbours are close — more interference, more retries, more jitter. Narrowing from 160 to 80 MHz, or to 40 on 2.4 GHz, trades peak speed for consistency, which is the right trade whenever calls matter more than file transfers.

Disable adapter power saving on laptops. Radios that sleep between transmissions add inconsistent wake-up delay. It is in the adapter's advanced properties on Windows.

Consider a mesh node or access point near where you actually work, ideally with a wired backhaul. A wireless-backhaul mesh node shares airtime with its own clients, which limits how much it can help.

Check roaming if the trouble follows you around the house. With a mesh system or several access points, moving between them makes the client reassociate, and calls can drop a second of audio at the handover. Devices also cling to a distant access point long after a nearer one would be better. If problems correlate with where you are sitting rather than what you are doing, that is the thing to look at — and a call taken sitting still is a quick way to confirm it.

Enable SQM on the router if latency climbs under load. Bufferbloat and Wi-Fi compound each other, and fixing the queueing helps on both media.

When Wi-Fi is genuinely fine

It would be misleading to suggest wireless is always the problem. It is not.

Streaming video, downloads, web browsing and file transfers all use buffering or reliable transport, and neither cares about a few milliseconds of variation. A 4K stream buffers seconds ahead; jitter is irrelevant to it. The applications that suffer are the ones covered in why calls drop on a fast connection and why games lag.

If your complaint is that video takes a moment to start, or that a large file transfers slowly, Wi-Fi is very unlikely to be the cause and you should look at your connection speed or the far end instead.

The rule of thumb: if the application can buffer, Wi-Fi variability is usually invisible. If it cannot — voice, video calls, competitive games, remote desktop — Wi-Fi is the first variable worth testing.

Frequently asked questions

Is Ethernet faster than Wi-Fi?

Usually not for the reason people expect. Modern Wi-Fi is often fast enough to saturate the internet connection behind it, so raw bandwidth is rarely why you would choose a cable — though on a multi-gigabit line a cable can still be meaningfully faster. Ethernet wins on consistency: lower jitter, no airtime contention, and no invisible retransmission. That matters for calls and games and not at all for streaming.

Why do my calls drop on Wi-Fi but not on a cable?

Because Wi-Fi introduces variable delay. Packets that miss the receiver's playback deadline are typically discarded, and most applications count them as loss even though the network delivered them. Ethernet removes the variation the wireless medium adds, which is usually enough for the buffer to cope. It is not a guarantee — congestion further along the path, bufferbloat in the router, or a loaded CPU can still deliver packets late over a cable.

I have full signal bars — why is my Wi-Fi still bad?

Signal strength determines whether frames get through, not how long they wait for airtime. A strong signal on a congested channel still produces high jitter. Contention and signal strength are independent problems.

Should I use 2.4 GHz or 5 GHz?

5 GHz or 6 GHz for anything real-time, provided the signal is good where you sit. They carry far more non-overlapping channels and much less interference from non-Wi-Fi equipment. But this depends on coverage rather than being a rule: 2.4 GHz travels further and through more walls, and a clean 2.4 GHz link at range can beat a weak 5 GHz one. 6 GHz has the most spectrum and the shortest reach.

Will a better router fix my Wi-Fi problems?

Sometimes, but placement and channel selection usually matter more than hardware. A well-positioned mid-range access point on a clear channel outperforms an expensive router in a cupboard. If you do buy, buy for queue management support rather than headline speed.

How do I test the difference between Wi-Fi and Ethernet?

Run the same test on both, minutes apart, and compare loss and jitter rather than throughput. Disable the wireless adapter when testing the cable, since many systems keep using Wi-Fi if it remains associated.

Does a Wi-Fi extender help?

It can improve coverage but often worsens consistency, because a wireless-backhaul extender uses the same airtime for both its uplink and its clients. A wired access point, or a mesh system with a wired backhaul, is a better answer where it is possible.