Start by defining slow
“The network is slow” can mean many things, and the first job is to make it specific. Ask:
- Slow to do what: load web pages, open files on a shared drive, join video calls, or download large items?
- Who is affected: one person, one room, one department, or everyone?
- Since when: always, since a particular change, or only at certain times of day?
- Is it constant or intermittent?
Writing the answers down turns a vague complaint into a lead. A problem affecting one laptop points in a very different direction from one affecting everyone at 3 p.m.
Step 1: Check the obvious on the affected device
Before blaming the network, rule out the machine itself. Look for a full disk, a runaway background process, a large update downloading, or a browser with dozens of tabs. Restarting the application, then the device, is cheap and sometimes enough. Checking background services and startup items can show a sync tool quietly saturating the connection.
Step 2: Compare with another device
Try the same task on another device in the same spot. If it is fast, the problem is probably specific to the first device. If it is slow too, move on. Then try the same device in a different place, ideally on a cable. The comparison between wired and wireless is one of the most informative tests available.
Step 3: Separate wired from wireless
If a wired connection is fast and wireless is slow, the issue lies in the Wi-Fi: distance, walls, congestion from neighboring networks, a crowded channel or an overloaded access point. The factors involved are described in the article on Wi-Fi channels. If both are slow, the cause is further in.
Step 4: Test the connection quality
Run a few simple measurements and note the results:
- Ping to the local router, then to an external address. High or erratic times to the router suggest a local problem. Fine times to the router but poor times beyond suggest the provider or the path.
- Packet loss during the ping run. Any consistent loss deserves attention.
- A speed test repeated a few times, ideally from a wired device.
- Traceroute to see where in the path the delay jumps.
What these numbers mean is covered in the explanation of network latency, bandwidth and packet loss. Remember that a single reading proves little; look for patterns.
Step 5: Check name lookups
Sometimes the connection is fine but the first step of every request, finding the address for a name, is slow. Symptoms include pages that hang before starting to load, then finish quickly. Testing with an address instead of a name can reveal this. The mechanism is explained in the piece on DNS lookups.
Step 6: Look at the local infrastructure
If the problem spans many users, inspect the shared parts:
- Check lights and port statistics on the switch for errors or a port negotiating at a lower speed than expected. A damaged cable can drop a link from gigabit to a slower rate or cause errors.
- Look at the router’s CPU and traffic graphs. A device that is overwhelmed or running out of memory will slow everything.
- Check whether the internet link is saturated. A large backup, cloud sync or software update can consume most of the bandwidth.
- See whether one device is generating unusual amounts of traffic. That may be harmless or a sign of malware types worth investigating.
- Look for loops or duplicate addresses, which can cause strange, widespread slowdowns.
Your system monitoring data, if you have it, is invaluable here, because it shows what normal looked like.
Step 7: Consider the outside world
If the local network looks healthy, the cause may be beyond your control: the internet provider, a remote service that is itself slow, or a congested peering path. Test several unrelated destinations. If only one site is slow, the fault is probably theirs. If everything is slow at certain hours, check with the provider and share your measurements, since evidence tends to speed up the conversation. This is also where service level agreements may be relevant.
Step 8: Change one thing at a time
When you have a theory, test it with a single change: move a device to a cable, swap a patch lead, change a channel, pause a sync. Then measure again. Making several changes at once leaves you unsure which helped, or whether one made something else worse. This discipline is the heart of troubleshooting method in general, and it applies to every technical problem.
Step 9: Record and communicate
Write down what you tested, what you found and what fixed it. A short note in the knowledge base helps the next person, including future you. When explaining the result to non-technical colleagues, avoid jargon and say what changed and what to expect; the guidance on technical communication covers this. If the fix involves changing network settings or replacing equipment, follow the usual change management routine so changes are approved and reversible.
Common culprits at a glance
- A weak or crowded wireless signal
- A faulty or poorly made cable
- An overloaded or aging router
- Bandwidth taken by large uploads, updates or streaming
- Slow name resolution
- A provider-side fault
Most slowdowns turn out to be one of these. Working from the device outward, one layer at a time, usually finds the cause faster than guessing.