When internet performance drags to a crawl every afternoon, the first instinct is to call the ISP. In small offices and home offices, that is often the wrong place to start.
There is a pattern to this. The slowdown usually lines up with a daily maintenance task or a local hardware limit. It might be a backup job kicking off at the same time, a switch loop adding traffic to the same network segment, or a router that is overheating after a long day.
The pattern is usually local
Bandwidth is not the only thing that matters. A 100 Mbps link can feel like a 2 Mbps link when a single device is flooding the network with duplicate traffic, or when a Wi-Fi access point is being asked to carry far more than it should.
Common triggers in Johannesburg offices include:
- Cloud backups starting around the same time each afternoon
- Unmanaged switches creating a loop
- Outdated routers losing performance as they heat up
- DNS lookups failing under load
- Wireless clients crowding the same channel
When you see the problem repeat at the same hour, the issue is usually not random. It is a schedule, a bottleneck, or a limit being reached.
What to check before you blame the ISP
Start with the facts. Check the router and switch logs for errors, packet drops, or connection storms. Look at what is using bandwidth at the time the slowdown begins. If a backup appliance starts copying data at 2 pm, the internet will feel slow even when the provider is fine.
Also check whether the issue is isolated to one room or all users. If everything on Wi-Fi lags while wired devices stay normal, the problem is probably local wireless congestion or a weak signal. If only one site is affected, and it repeats on a schedule, a local router or switch issue is more likely than a line issue from the ISP.
Another quick test is to compare DNS resolution and gateway response times. If websites resolve slowly or the gateway is timing out under ordinary load, the fault is inside the local network stack rather than in the upstream line.
Check the basics before you escalate
- Review router and switch temperatures and reboots
- Confirm if cloud backups or sync jobs start at the same time every day
- Check for unusual traffic on a single switch port
- Test wired and wireless users separately
- Review DNS and gateway logs for repeated failures
These are not complicated tasks, but they save you from wasting hours on a support call when the real problem is at your end.
A short diagnostic checklist
If the slowdown starts at 2 pm, ask: what normally happens at 2 pm? A staff member clocks in? A backup job runs? A security system begins syncing? Most performance problems have a human cause tied to a routine action.
That is why the first job is not to accuse the internet provider. It is to narrow the problem to the local path: edge hardware, Wi-Fi performance, or traffic patterns on the LAN.
Once you know what is saturating the link, the fix is usually straightforward. Move large jobs outside peak hours, separate the backup path from the user traffic, replace a failing switch, or add capacity where the bottleneck is really occurring.
In other words, the trick is not to fix everything at once. It is to identify the point where the chain breaks, fix that link, and then decide whether the supplier is actually the issue.