Wi-Fi Speed Test: Find Out If Your Router (Not Your ISP) Is the Problem
Run a Wi-Fi speed test the right way: compare bands, rooms and ethernet baselines, decode the numbers, and fix weak coverage with the cheapest fix first.
Quick answer: A Wi-Fi speed test tells you what your wireless link delivers where you actually sit — usually 20-60% of what ethernet shows. Test both, compare, and you know in two minutes whether your problem is the line or the Wi-Fi.
Wi-Fi speed vs internet speed: two different numbers
Your internet plan is a promise between your ISP and your modem. Your Wi-Fi is a separate radio network between the router and every device. A speed test over Wi-Fi measures the weakest link in between — radio interference, distance, walls, band congestion, an overloaded router CPU. This is why "I pay for 300 Mbps but get 90" is, in most homes, a Wi-Fi story, not an ISP story. Measure both: ethernet run (what the ISP delivers) and Wi-Fi run (what you actually experience) with Toolfyra Internet Speed Test.
The 2.4 GHz vs 5 GHz vs 6 GHz reality
| Band | Real-world speed ceiling | Range | Best for |
|---|---|---|---|
| 2.4 GHz | 50-150 Mbps | long, walls no problem | smart home gadgets, far rooms |
| 5 GHz | 300-900 Mbps | medium | laptops, phones, streaming |
| 6 GHz (Wi-Fi 6E/7) | 900 Mbps-2 Gbps | short, loves line of sight | new devices, gigabit plans |
The trap: 2.4 GHz travels further, so phones glom onto it near the edge of coverage and crawl. If your router broadcasts one merged network name, force a band check — stand where you usually sit and test; if the result looks like 2.4 GHz numbers, split the SSIDs (e.g. MyWifi and MyWifi-5G) and pin speed-hungry devices to 5 GHz.
The room-by-room testing method
- Baseline on ethernet (or sitting next to the router on 5 GHz) — call this 100%.
- Test in the kitchen, bedroom, and workspace. Note Mbps and ping each time.
- Any room below roughly 50% of baseline has a coverage problem, not an ISP problem.
- Fix in order of cost: move the router (center of home, off the floor, away from metal/kitchens), then add a wired access point or mesh node in the weak zone, then consider powerline/MoCA adapters before blaming the plan.
Wi-Fi numbers decoded
- Wi-Fi vs wired gap over 40% — interference or old standards. 802.11n (Wi-Fi 4) gear tops out around 100-150 Mbps real-world; Wi-Fi 5 reaches 300-600; Wi-Fi 6 more.
- Great near router, terrible one wall away — classic 5 GHz attenuation or mesh backhaul running over a weak signal.
- Fast in the morning, collapsing at night — neighbors’ networks filling the same channels. A free Wi-Fi analyzer app shows channel crowding; pick channels 1/6/11 on 2.4 GHz, or move to 5 GHz DFS channels.
- Speed fine, ping all over the place — power-save mode on the radio, or bufferbloat. See bufferbloat.
Router settings that quietly halve your speed
- Legacy compatibility mode (b/g protection) drags every client to the slowest device’s pace — disable legacy modes.
- 20 MHz channel width on 5 GHz — double it to 80 MHz unless you live in an apartment tower.
- Band steering too aggressive — devices stick to 2.4 GHz; split SSIDs to control it manually.
- Old firmware — several router lines recovered 20-40% throughput from vendor updates.
When to blame the router, not the ISP
A router CPU saturating under NAT load is the hidden tax on fast plans: a 2016 router happily passed 100 Mbps, collapses at 500+. If ethernet into the router is fast but every Wi-Fi device struggles, and rebooting only helps for a day, the router is the bottleneck. Wi-Fi 6 routers are the current price-performance sweet spot for plans above 300 Mbps.
FAQ
Why is Wi-Fi always slower than ethernet? Radio shares airtime, retries collisions, and pays encoding overhead — 60-70% of wired throughput is normal for a good link.
Does a speed test damage my Wi-Fi? No — it is just traffic. Run as many tests as you like; they move tens of MB.
My phone shows 866 Mbps link speed but tests at 200? Link rate is the radio’s theoretical ceiling under perfect conditions; real throughput after overhead and retries is much lower. Trust the speed test, not the badge number.