Talking to Your ISP About High Latency or Lag

Quick answer: contact your internet service provider (ISP) only after you have ruled out your own equipment, and go in with evidence. That means latency measurements taken over a cable, repeated at different times of day over several days, to several different destinations, plus the times it is worst and what you already tried. Describe the problem in numbers ("average 85 ms to nearby servers between 8 and 11 pm, normally 15 ms, with spikes above 300 ms") rather than "the internet is slow", and ask for specific checks. This turns a vague complaint into something a technician can act on.

When is it worth contacting your ISP?

Latency is the delay for data to make a round trip, measured in milliseconds (ms); see what is latency. Many causes of high latency are inside your home: Wi-Fi, a saturated upload, a failing router, a VPN. Your provider can do nothing about those, and a support call that ends with "it is your Wi-Fi" wastes everyone's time. Contact the ISP when your tests point to the line or their network, for example when:

• latency is high over a cable directly to the router, with other devices idle;
• it is high to many unrelated destinations, including nearby ones;
• it follows a pattern, such as every evening or after a certain event;
• it is accompanied by packet loss, disconnections or very uneven jitter;
• it started suddenly and a restart did not help.

Work through troubleshoot high latency first. It is designed to tell you which layer is at fault.

Step 1: rule out your own equipment

Before calling, do the checks a support agent will ask you to do anyway, and note the result of each. It saves a call and strengthens your position.

1. Connect a computer to the router by Ethernet and turn off Wi-Fi on that computer.
2. Pause other devices' downloads, uploads, backups and streams.
3. Restart the modem and router, wait until all lights are stable, and test again.
4. Turn off any VPN, proxy or traffic-inspecting security software on the test computer.
5. If possible, test with a second device, and with a different cable.
6. If the ISP's modem or router is separate from your own router, test with your own router bypassed, if the provider allows it and you know how, to see whether the issue is on their equipment.

Step 2: collect evidence that is hard to dismiss

Run repeated tests over several days

One test shows one moment. Run tests at the same times each day, for example morning, afternoon and evening, for three to seven days. With GlobeLatency, choose several destinations in different places, a few nearby and a few distant, and run a test. It records your average, best and worst pass for each, and shows how you compare with the world average for the same destinations. It then saves a report with a link you can keep; see saved reports and sharing. The browser also keeps a history of past tests on your device. Test from a wired computer, with the tab in the foreground.

Watch it live during a bad period

Leave the monitor running during the evening to see the pattern and how large and frequent the spikes are. The monitor shows average, best and worst and the no-reply percentage for each destination, and an optional alert can tell you when something goes down or gets slow. Live data stays in your browser; take screenshots with timestamps if you want to show them. Check accuracy and limitations so you understand what a browser-based measurement can and cannot show: it includes browser and operating-system overhead, and its timings are HTTPS timings, a bit higher than raw ICMP ping.

Add command-line output

A support technician will trust standard tools. Capture these from the wired computer:

• ping to your default gateway (your router) and to a public server, for a few minutes: ping -c 100 8.8.8.8 on Linux and macOS, or ping -n 100 8.8.8.8 on Windows. Note the minimum, average, maximum and the loss percentage.
• traceroute (tracert on Windows) or mtr to a destination that is slow, at a bad time and at a good time.

The ping to your router shows whether the problem is inside your home (high there) or beyond it (low there, high further out). The hop where latency first jumps and stays high in the traceroute shows roughly where the delay begins. Remember that a single intermediate hop with high values and normal later hops is usually a router that deprioritises its own replies, so do not present that as the fault.

Keep a simple log

Date and time	Connection	Destination	Average	Worst	Loss	Notes
Mon 20:30	Ethernet	Nearby server A	82 ms	310 ms	2%	Other devices idle
Tue 07:00	Ethernet	Nearby server A	14 ms	19 ms	0%	Normal

Two rows like these, repeated for several days, tell a clear story: it is fine in the morning and bad in the evening.

Step 3: describe the problem in a way that gets action

Support agents work from scripts. A clear, measurable description helps you skip to the right level.

• Say what is wrong in measurable terms: "Round-trip latency over Ethernet is 80 to 100 ms to servers in my own region in the evening, versus 12 to 18 ms in the morning, with occasional packet loss."
• Say it is not speed: speed tests may pass while latency is poor. Say so, so they do not close the ticket on a good speed result.
• State what you did: wired, devices idle, restarted, second device tested.
• State when it began and any change before then (new plan, new equipment, outage, move).
• Ask for specific checks: line signal levels and error counters, congestion on your local node or neighbourhood segment, a technician visit if the line is suspect, and a replacement of their modem if you use it.
• Ask for a ticket or reference number and a note of what was done.

What an ISP can and cannot check

Layer	Can the ISP see it?	What they may do
Line quality (signal, noise, errors)	Usually yes, for their own equipment	Test the line, send an engineer, fix or replace cabling or the modem
Local network congestion	Often, in their monitoring	Add capacity, rebalance the node (can take time)
Their core and peering links	Yes	Fix a fault, adjust routing; see routing, BGP and peering
A distant network or the destination	Not directly	Possibly contact the other network; often cannot resolve
Your Wi-Fi and home devices	Not normally	Advice only

What if support does not help?

1. Ask to escalate to a second-line or network engineer, mentioning your evidence by name ("I have a week of timestamped tests showing the evening pattern").
2. Reply in writing (email, ticket or online form) with your log attached so there is a record.
3. Request a technician visit if you have loss or error symptoms on a wired line.
4. Check your contract for any latency or service-level terms. Most consumer plans only state "up to" speeds, so you may not have a formal guarantee. Business plans sometimes do.
5. Use the formal complaints route your provider and your country's regulator offer, if the problem persists and is documented. Procedures differ by country, so look up the one for yours.
6. Consider alternatives: another provider, another access technology, or a backup connection. See Wi-Fi, Ethernet, fibre, mobile and satellite latency for baseline differences. Before switching, check what neighbours with the other provider experience.

Worked examples

Example 1: evening congestion

Over five days a user tests at 07:00, 14:00 and 21:00 over Ethernet. Mornings average 12 to 16 ms to a nearby server; evenings average 70 to 110 ms with worst passes above 250 ms. Speed tests during the day show the full plan speed. The user tells the provider: "Not a speed issue, evening-only latency increase to all destinations on a cable." Because the pattern is time-based and affects every destination, the support agent can escalate to network capacity rather than send the customer back to restart the router.

Example 2: the problem is only to one service

Latency to ten destinations is normal. A particular video service is 300 ms slower than expected from this user but not from a friend on another provider. The user attaches a traceroute showing delay starting at an interconnection between the ISP and another network. The ISP cannot fix the other network quickly, but now the ticket has a precise description, and the user can choose to test a VPN to see whether it avoids the bad path; see VPN and latency.

Example 3: it was inside the home

A user is ready to call after seeing 120 ms average on Wi-Fi. A wired test shows 18 ms. The ping to the router over Wi-Fi was already 60 ms, so the weak link is the wireless connection. No call is needed; moving the router solves it. Doing the wired test first avoided a pointless complaint.

Common mistakes and myths

• Calling with only a speed test. Speed (bandwidth) and latency are different; see latency vs bandwidth vs throughput.
• Testing over Wi-Fi and blaming the provider. Test on a cable first.
• Quoting one number. Provide averages, worst values and times over several days.
• Expecting a promised ping. Most plans do not promise one.
• Blaming the ISP for distance. A server 10,000 km away cannot give 30 ms. Compare with what distance allows; see what is a good latency.
• Being hostile to first-line support. They follow a script. Politely complete the steps, then ask for escalation.
• Forgetting to mention that you restarted. Say it in the first minute.

What to do next

Run the free latency test from a wired computer at several times of day and keep the report links. If you have not yet isolated the cause, follow troubleshoot high latency; if the problem turns out to be local, how to reduce latency lists the fixes. When you have your evidence ready, contact your provider with the log and the specific checks you want.