Most people reboot the router when the internet feels slow, then go shopping for a newer one. A slow connection rarely has the cause people assume. Both steps skip the useful part.
The fix order below eliminates causes one at a time, cheapest first. It takes about twenty minutes if you work through it in sequence. It takes weeks if you guess.
Start with a cable, then a second device
Plug a laptop directly into the router with an Ethernet cable and run a speed test. This one step removes Wi-Fi from the equation, and Wi-Fi is the culprit far more often than the provider is.
If the wired result matches the plan you pay for, your provider is fine and the problem lives inside your house. If the wired result is bad too, stop here and call your provider with that number ready. Nothing you buy will fix it.
Now compare devices. Run the same test on a second laptop, or on a phone held in the same spot. One slow device points at that device. Every device slow points at the network.
This step matters because an old laptop with a tired Wi-Fi card makes an entire house look broken. People replace routers over one aging machine.
Then compare times of day. Run the test at seven in the evening and again at eleven in the morning.
If the evening result collapses and the morning result looks fine, you are seeing neighborhood congestion rather than a hardware fault. Your provider has oversold the local node, and no router fixes that. It only tells you when your large downloads should run.
There is also a quieter cause. Something in your house may be using the connection while you test it. A phone backing up photos, a cloud drive syncing, or a console preloading an update can eat most of the link without anyone noticing. Your router’s device list usually shows which client is busy, so check it during a slowdown before blaming anyone.
But there is a version of this problem that costs nothing to test. Before you change any hardware, change what feeds it. A damaged Ethernet cable or a failing power adapter causes more strange slowdowns than most people expect, and both cost a few dollars to swap. Replace the cable and rerun the wired test before you go further.
A slow internet complaint is often really a complaint about one website. If a single service drags while everything else is fine, the problem sits at that service, not in your house. Check your connection against a second site before you assume the whole line is bad.
Slow and laggy are two different problems
Bandwidth and latency are separate measurements, and mixing them up leads to the wrong purchase. Bandwidth is how much data moves per second. Latency is how long a single request takes to come back.
A 500Mbps connection with 90ms of latency feels broken on a video call and perfectly fine while downloading a file. A 100Mbps connection with 12ms feels instant for everything except large transfers.
So check both. Speed tests report latency next to the download number, usually labeled as ping. Anything under about 30ms is healthy. If ping sits above 60ms while the download figure looks fine, you have a latency problem, and more bandwidth will not touch it.
Latency usually comes from one of three places. A congested Wi-Fi band, a bad route to the server, or a device doing something heavy in the background. A console downloading an update in another room will do this every time.
What a speed test never shows
There is also a cause that speed tests never reveal. A slow DNS resolver makes every page feel sluggish while the raw connection underneath is perfectly fast, which is why the problem is so hard to spot. Pages take an extra beat before they begin loading, and downloads still run at full speed. If browsing drags while a large file flies, switch your DNS to a public resolver and test again.
Then look at where the router sits. Placement beats specification more often than anyone admits. A router behind a television, inside a cabinet, or against an exterior wall is doing worse work than a cheaper model sitting in the open.
The band matters as much as the box. The 2.4GHz band travels further and carries less. The 5GHz band is faster and dies sooner. If a device in a back room is stuck on 5GHz with two bars, moving it to 2.4GHz will often make it feel quicker despite the lower number on the label.
Channels your neighbors also use, and what to do about it
In an apartment, check the channel too. Neighbors on the same channel compete for the same airtime, and an analyzer app shows which ones are crowded. Switching your Wi-Fi channel to a clear one is a free fix that most people never try.
And if coverage is the real complaint, a mesh system with a wired backhaul beats both a bigger router and a wireless mesh. A cable between the nodes removes the half-speed penalty that wireless backhaul imposes on every device in the house.
Range extenders deserve a warning here. A cheap extender repeats an already weak signal and halves the throughput it passes along, so it often makes a bad room feel worse. A mesh system with a cable running to it is the version that works.
I once turned an old phone into a better router than the one I was paying for, which sounds like a stunt and turned out to be a lesson about placement and interference. The hardware was never the interesting part.
When to actually spend
Only after all of that should you shop. If the wired test is fast, the second device is fast, the evening test is fast, the ping is low, the channels are clear, and one room still has no signal, then you have a coverage problem. That is the point where mesh or a better access point earns its money.
Advertised speed is not delivered speed, and the gap is normal. Providers sell a tier as “up to” a number, and the real figure lands somewhat below it because of protocol overhead and the condition of the line. A gigabit plan that measures 850Mbps is working correctly.
One more measurement trick. Test against two different servers rather than one. A single distant server can report a low number while your connection is fine, because the bottleneck sits somewhere between you and that test node. If both servers agree, believe them.
So keep the total cost in view. A new router, a mesh kit, and a range extender can add up to more than a year of internet service, and most households need none of them. Fix the free things first and let the numbers tell you when it is time to spend.
One last setting to check before you spend anything. Log into your provider’s gateway and look for bridge mode or IP passthrough. Running your own router behind a gateway that is also routing stacks two networks on top of each other, and that configuration causes strange latency and broken multiplayer sessions. Bridge mode fixes it in one step.
So work the list in order. Cable first. Second device second. Time of day third.
Latency and DNS come after that. Placement and channels come before hardware. Most slow internet complaints in a normal home trace back to one of those, and none of them require a new router.
