How to Check Internet Speed — Accurate Speed Test Guide

Last Updated: October 2026

Whether your video calls are freezing, your game is lagging, or Netflix is stuck buffering, the first step in diagnosing the problem is running a speed test. But not all speed tests are created equal, and the results can be misleading if you don't know what you're looking at. This guide covers everything you need to know about testing your internet speed accurately, understanding what the numbers mean, and figuring out what to do when your connection isn't performing as expected.

What Does a Speed Test Measure?

A speed test evaluates four key aspects of your internet connection. Understanding each one is essential for diagnosing performance issues correctly.

Download Speed

Download speed measures how fast data travels from the internet to your device, expressed in Mbps (megabits per second). This is the number most people care about because it determines how quickly web pages load, how smoothly videos stream, how fast files download, and how responsive cloud applications feel. When ISPs advertise "up to 500 Mbps," they are referring to download speed.

During a speed test, the tool downloads a large file (or multiple files) from a test server and measures the throughput. The test ramps up gradually and measures the sustained transfer rate over several seconds, discarding the initial slow-start phase to get an accurate reading.

Upload Speed

Upload speed measures how fast data travels from your device to the internet. It affects video calls (your camera feed is an upload), live streaming, cloud backups, sending large email attachments, uploading photos and videos to social media, and pushing code to remote repositories.

On most cable and DSL connections, upload speed is significantly lower than download speed. A plan advertised as 300 Mbps download might only include 10–20 Mbps upload. Fiber connections typically offer symmetrical speeds (equal download and upload), which is a major advantage for work-from-home and content creation.

Latency (Ping)

Latency, measured in milliseconds (ms), is the time it takes for a small data packet to travel from your device to the test server and back. This round-trip time is often called ping. Low latency is critical for real-time applications:

• 1–20 ms: Excellent. Ideal for competitive online gaming, VoIP, and video conferencing.
• 20–50 ms: Good. Sufficient for most online activities including casual gaming.
• 50–100 ms: Acceptable. You may notice slight delays in real-time applications.
• 100–200 ms: Noticeable lag in gaming and video calls. Browsing still works but feels sluggish.
• 200+ ms: Poor. Gaming is difficult, video calls are choppy, and everything feels delayed.

Latency is primarily determined by the physical distance between you and the server, the number of network hops in between, and congestion along the route. Fiber connections typically have the lowest latency; satellite connections have the highest (often 500–600 ms for traditional geostationary satellites, though Starlink achieves 25–60 ms).

Jitter

Jitter is the variation in latency over time, also measured in milliseconds. If your ping measurements over 10 seconds are 18, 19, 20, 17, 21 ms, your jitter is very low (about 1–2 ms). If they are 15, 45, 22, 80, 30 ms, your jitter is high. High jitter causes:

• Choppy, robotic audio in voice and video calls
• Stuttering and buffering in live streams
• Rubber-banding and teleporting in online games
• Inconsistent performance in cloud applications

Jitter below 10 ms is ideal. Jitter below 30 ms is acceptable for most applications. Anything above 50 ms will cause noticeable problems. High jitter is often caused by network congestion, an overloaded router, or interference on a wireless connection.

How Speed Tests Work

Understanding the mechanics of a speed test helps you get more reliable results and interpret them correctly.

When you run a speed test, the tool first measures latency by sending small ping packets to the test server and measuring the round-trip time. Then it tests download speed by initiating multiple simultaneous TCP connections to the server and downloading data in parallel. The test starts small and ramps up, measuring throughput over several seconds. The algorithm discards the initial ramp-up period and calculates the sustained speed. The upload test works the same way in reverse — your device sends data to the server.

Most speed test services use servers distributed globally. The test automatically selects a server based on your geographic location and lowest ping, though you can often manually choose a different server. The closer the server, the more your result reflects your connection's maximum capability. Testing against a distant server introduces more network hops and potential congestion, which can lower your results — but that more accurately represents real-world performance to that region.

Factors That Affect Speed Test Accuracy

Many variables can make your speed test results unreliable. Here are the most common factors and how to control for them:

Wi-Fi vs. Ethernet

This is the single biggest factor in speed test accuracy. Wi-Fi introduces interference, signal degradation, and shared bandwidth that can dramatically reduce your measured speed. A device connected via Ethernet cable directly to the router will nearly always show higher, more consistent results than the same device on Wi-Fi.

For example, if your plan is 500 Mbps, a wired test might show 480 Mbps (normal), while a Wi-Fi test from the next room might show 200 Mbps. That 200 Mbps result doesn't mean your ISP is underperforming — it means your Wi-Fi is the bottleneck. Always run your first speed test over Ethernet to establish a true baseline before troubleshooting.

Server Distance

The physical and network distance between you and the test server directly impacts results. A server in the same city will give the best results. A server on another continent will show lower speeds due to increased latency and more routing hops. Choose a nearby server for baseline testing, but also test against distant servers to see real-world performance for services hosted far away.

Time of Day

Internet speeds on cable and DSL connections often drop during peak hours (7–11 PM) when everyone in your neighborhood is streaming, gaming, and downloading. This is because cable infrastructure shares bandwidth among all subscribers in an area. Run tests at different times — early morning, midday, and evening — to see if peak-hour congestion is affecting you. Fiber connections are less susceptible to this effect.

VPN and Proxy

A VPN encrypts your traffic and routes it through an intermediary server, which adds latency and can reduce throughput. If you test with a VPN active, your results will be lower than your raw connection speed. Disable your VPN before running a baseline speed test. However, testing with a VPN can be informative — if your speed is faster through a VPN, your ISP may be throttling specific types of traffic.

Background Activity

Any device on your network consuming bandwidth will reduce your speed test results. This includes: other computers streaming video, phones syncing photos to the cloud, game consoles downloading updates, smart home cameras uploading footage, cloud backup services running in the background, and Windows Update downloading patches. Pause or disconnect everything you can before testing.

Browser and Device Limitations

Older devices with slow processors, limited RAM, or outdated network adapters may not be able to fully utilize a high-speed connection. Browser-based speed tests in particular can be limited by browser performance, extensions, and other open tabs. If you suspect device limitations, test on multiple devices or use a dedicated speed test application.

How to Get Accurate Speed Test Results

Follow this step-by-step process for the most reliable measurements:

1. Connect via Ethernet: Plug your computer directly into the router with a Cat 5e or Cat 6 cable. This eliminates Wi-Fi as a variable.
2. Close everything: Quit all applications, close all browser tabs except the speed test, and stop any downloads or uploads.
3. Disconnect other devices: Pause streaming, pause cloud sync, and if possible, disconnect other devices from the network temporarily.
4. Disable VPN: Turn off any VPN or proxy software.
5. Choose a nearby server: Select a test server in your city or region for baseline measurement.
6. Run three tests: Run the test at least three times with a minute between each run. Average the results.
7. Test at different times: Repeat the process at different times of day (morning, afternoon, evening) to identify peak-hour slowdowns.
8. Record everything: Note the date, time, download speed, upload speed, ping, jitter, and test server for each run. This data is invaluable if you need to contact your ISP.

Understanding Your Results: Mbps vs. MBps

One of the most common points of confusion is the difference between Mbps and MBps. They look similar but represent very different values:

• Mbps (megabits per second) — This is how ISPs measure and advertise internet speed. A "bit" is the smallest unit of data (a 0 or 1). All speed test results are in Mbps.
• MBps (megabytes per second) — This is how file sizes and download progress are typically displayed. A "byte" is 8 bits. File managers, download managers, and torrent clients usually show speeds in MBps.

To convert: divide Mbps by 8 to get MBps. So a 100 Mbps connection can theoretically download at 12.5 MBps. If your ISP plan is 200 Mbps and your browser shows a file downloading at 22 MBps, that's actually 176 Mbps — perfectly normal with some protocol overhead.

Here's a quick reference table:

• 25 Mbps = 3.1 MBps (a 1 GB file takes ~5.3 minutes)
• 100 Mbps = 12.5 MBps (a 1 GB file takes ~1.3 minutes)
• 300 Mbps = 37.5 MBps (a 1 GB file takes ~27 seconds)
• 500 Mbps = 62.5 MBps (a 1 GB file takes ~16 seconds)
• 1 Gbps (1000 Mbps) = 125 MBps (a 1 GB file takes ~8 seconds)

What Internet Speed Do You Actually Need?

ISPs love to upsell faster plans, but how much speed do you actually need? It depends entirely on your household's usage. Here are realistic recommendations:

Basic Browsing and Email (1–2 users)

10–25 Mbps is sufficient for web browsing, email, social media, and standard-definition video streaming. This is the minimum tier worth paying for — anything lower will feel frustrating even for basic tasks.

HD Streaming and Video Calls (2–3 users)

50–100 Mbps comfortably handles HD streaming on multiple devices, Zoom/Teams calls, casual online gaming, and background downloads. This is the sweet spot for small households and couples.

4K Streaming and Gaming (3–5 users)

100–300 Mbps supports multiple 4K streams, competitive online gaming, work-from-home video conferencing, and large file downloads. If you have a family of four with multiple devices, this is the range to target.

Heavy Use and Content Creation (5+ users)

300 Mbps–1 Gbps is recommended for large households with many connected devices, content creators uploading large video files, gamers downloading 100 GB game updates, and smart homes with dozens of IoT devices. If symmetric upload speed is important (live streaming, cloud backups, running a home server), fiber with symmetrical speeds is the best choice.

Gaming-Specific Requirements

Online gaming actually uses very little bandwidth — most games consume only 20–80 Mbps of throughput. What matters far more for gaming is low latency (under 50 ms, ideally under 20 ms) and low jitter (under 10 ms). A 50 Mbps fiber connection with 10 ms ping will feel dramatically better for gaming than a 500 Mbps cable connection with 60 ms ping. The big downloads are game updates, which benefit from higher speeds but don't need low latency.

ISP Speed vs. Actual Speed: Why They Differ

Your ISP advertises speeds "up to" a certain number, but there are legitimate reasons why your actual speed may be lower:

• Network overhead: TCP/IP headers, encryption (TLS), and protocol framing consume 5–10% of raw bandwidth. A 100 Mbps connection delivering 90–95 Mbps of usable throughput is normal.
• Shared infrastructure: Cable internet shares bandwidth among all subscribers in a neighborhood node. During peak hours, actual speeds can drop well below the advertised maximum.
• ISP provisioning: Some ISPs deliver slightly more than advertised to account for overhead; others provide exactly the provisioned rate.
• Distance from exchange: DSL speeds degrade with distance from the telephone exchange. If you're far from the exchange, you may only get a fraction of the advertised speed.
• Modem and router limitations: An outdated DOCSIS 3.0 modem may cap out at 300–400 Mbps even on a gigabit plan. An old Wi-Fi 4 router cannot deliver 500 Mbps over wireless regardless of your plan.

What's normal: Getting 80–95% of your plan speed over a wired connection is typical and acceptable. Getting 50–70% over Wi-Fi is common for a device in the same room as the router. If your wired speed is consistently below 70% of your plan, something is wrong.

Why Wi-Fi Speed Differs from Your Plan Speed

Many people test their internet speed over Wi-Fi and conclude their ISP is ripping them off. In reality, the gap between your plan speed and your Wi-Fi speed is almost always caused by wireless limitations, not ISP underperformance.

• Wi-Fi standard limitations: Wi-Fi 4 (802.11n) maxes out around 150–300 Mbps in practice. Wi-Fi 5 (802.11ac) reaches 400–800 Mbps for a single device. Wi-Fi 6 (802.11ax) can reach 1–2 Gbps under ideal conditions. If your plan is faster than your Wi-Fi standard supports, Wi-Fi is your bottleneck.
• Distance and obstacles: Wi-Fi signal strength drops rapidly with distance and through walls. Two rooms away, your effective speed may be 25–50% of what you get next to the router.
• Interference: Neighboring Wi-Fi networks, microwave ovens, baby monitors, Bluetooth devices, and cordless phones all cause interference on the 2.4 GHz band.
• Channel congestion: In apartments and dense neighborhoods, dozens of Wi-Fi networks may overlap on the same channels, reducing throughput for everyone.
• Device capabilities: Not all devices support the latest Wi-Fi standards. A phone with Wi-Fi 5 cannot take advantage of a Wi-Fi 6 router's maximum speed.

Speed Test Tools Compared

Not all speed tests produce the same results. Here is a comparison of the most popular options and their strengths:

• Speedtest by Ookla (speedtest.net): The most widely used speed test. Has the largest server network (15,000+ servers globally), provides detailed results including ping, jitter, and packet loss. Available as a web app and native apps. Some ISPs have dedicated Ookla servers that may give inflated results.
• Fast.com (by Netflix): Simple and clean interface focused on download speed. Tests against Netflix's CDN infrastructure, which accurately represents your streaming experience to Netflix specifically. Shows upload speed and latency if you click "Show more info."
• Google Speed Test: Search "speed test" on Google and a built-in test appears. Uses Google's own infrastructure. Quick and convenient but limited customization.
• Cloudflare Speed Test (speed.cloudflare.com): Tests against Cloudflare's global network with detailed latency measurements, jitter analysis, and loaded/unloaded latency comparisons. Excellent for advanced diagnostics.
• LibreSpeed: An open-source, self-hostable speed test with no ads and no tracking. Ideal for privacy-conscious users or for setting up your own internal speed test server.

Best practice: Run tests on at least two different services and average the results. This compensates for differences in test methodology and server proximity.

Troubleshooting Slow Internet Speeds

If your speed test results are consistently lower than expected, work through these troubleshooting steps systematically:

Step 1: Establish a Wired Baseline

Connect directly to your router or modem with an Ethernet cable and run a speed test. If the wired speed matches your plan (within 80–95%), the issue is with your Wi-Fi, not your internet connection. Focus on wireless optimization. If the wired speed is significantly below your plan, continue to Step 2.

Step 2: Reboot Your Equipment

Unplug your modem and router for 30 seconds, then plug the modem in first and wait for it to fully connect before plugging in the router. This clears the device's memory, resets connection state, and often resolves transient issues. After the reboot, run the speed test again.

Step 3: Check for Bandwidth Hogs

Open Task Manager (Windows) or Activity Monitor (macOS) and check for processes with high network usage. Check your router's admin panel for per-device bandwidth usage. Cloud backup services, streaming on other devices, game downloads, and OS updates are common culprits.

Step 4: Test Latency and Packet Loss

Raw speed is only part of the picture. Use our Ping & Traceroute tool to check latency and trace the network path to common destinations. Look for high latency (over 50 ms to nearby servers), packet loss (any percentage above 0%), and specific hops with large latency jumps that indicate a bottleneck.

Step 5: Check Your Equipment

Verify that your modem supports your plan speed (DOCSIS 3.0 for cable up to ~400 Mbps, DOCSIS 3.1 for gigabit). Check that your router supports the speed you're paying for. Inspect Ethernet cables for damage — try a different cable. Check the lights on your modem for any error indicators.

Step 6: Test at Different Times

If speeds are normal during the day but drop in the evening, you're likely experiencing peak-hour congestion on shared cable infrastructure. Document the speeds at different times — this data is useful when contacting your ISP.

Step 7: Contact Your ISP

If wired speeds are consistently below 80% of your plan despite all troubleshooting, the issue is on your ISP's side. Call their support line with your documented speed test results, timestamps, and the steps you've already taken. Ask them to check signal levels on your modem, look for line issues, and potentially dispatch a technician.

For a deeper dive into diagnosing slow connections, check out our guide on why is my internet slow.

When to Contact Your ISP

Not every speed issue warrants a call to your ISP. Here's when you should and shouldn't contact them:

• Do contact your ISP if: Your wired speed is consistently below 80% of your plan, you experience frequent disconnections, speeds drop dramatically at certain times despite no changes on your end, or your modem shows error lights.
• Don't contact your ISP if: You're testing over Wi-Fi (fix your wireless first), other devices are consuming bandwidth during the test, you haven't rebooted your router recently, or your speed is within 80–95% of your plan (that's normal).

Use our What Is My IP tool to check your public IP address and ISP details before calling — this information helps their support team locate your account and diagnose issues faster.

Frequently Asked Questions

What does a speed test actually measure?

A speed test measures four metrics: download speed (how fast data arrives, in Mbps), upload speed (how fast data leaves, in Mbps), latency/ping (round-trip time in milliseconds), and jitter (variation in latency). Download speed affects streaming; upload affects video calls; latency affects gaming and real-time apps.

Why is my speed test result lower than what I pay for?

Common causes include testing over Wi-Fi instead of Ethernet, other devices consuming bandwidth, peak-hour congestion, a distant test server, an active VPN, or outdated router/modem hardware. Always test wired with nothing else on the network for an accurate baseline.

What is the difference between Mbps and MBps?

Mbps (megabits per second) is how ISPs advertise speeds. MBps (megabytes per second) is how file downloads are measured. There are 8 bits in a byte, so 100 Mbps equals 12.5 MBps. If your plan is 200 Mbps and a file downloads at 22 MBps, that's 176 Mbps — perfectly normal.

What internet speed do I need for streaming?

SD streaming needs 3–5 Mbps. HD (1080p) needs 5–10 Mbps. 4K Ultra HD needs 25–35 Mbps per stream. For a household with multiple simultaneous 4K streams plus video calls, aim for 100–300 Mbps.

How do I get the most accurate speed test results?

Connect via Ethernet cable, close all apps, disconnect other devices, disable VPN, choose a nearby test server, and run three tests spaced a minute apart. Test at different times of day and average the results.

Why is my Wi-Fi speed so much slower than my wired speed?

Wi-Fi is slower than Ethernet due to radio interference, signal loss through walls and distance, shared bandwidth among wireless devices, and protocol overhead. A device two rooms from the router may get 25–50% of the wired speed. This is a Wi-Fi limitation, not an ISP issue.

What is jitter and why does it matter?

Jitter is the variation in latency over time. Consistent ping (e.g., 20±2 ms) means low jitter. Fluctuating ping (e.g., 15–80 ms) means high jitter. High jitter causes choppy calls, stuttering streams, and laggy games. Aim for jitter below 30 ms; below 10 ms is ideal.

When should I contact my ISP about slow speeds?

Contact your ISP if your wired speed is consistently below 80% of your plan, you experience frequent disconnections, or speeds drop dramatically at certain times after you've already rebooted equipment and ruled out local issues. Have speed test results with timestamps ready.