How Accurate Are Speed Tests? Report 2026

Factors that affect speed test accuracy

Key findings

• Wi-Fi is the single biggest accuracy killer: Testing over 2.4 GHz Wi-Fi can reduce measured speeds by 30–40% compared to a wired Ethernet connection. Even 5 GHz Wi-Fi introduces 10–15% variance due to interference, signal attenuation, and protocol overhead. For any accurate measurement, use Ethernet.
• VPN usage invalidates speed tests: A VPN routes all traffic through an encrypted tunnel to a server, adding encryption overhead and routing through the VPN provider's infrastructure. This can reduce measured throughput by 30–50%, making it appear your ISP is throttling when the issue is the VPN itself.
• ISP throttling is real and detectable: Some ISPs throttle specific categories of traffic (streaming, gaming, P2P) while allowing speed test traffic to pass unthrottled. Testing via a VPN and comparing results to a direct test can reveal this — if the VPN test is faster, the ISP is selectively throttling.
• Peak-hour results reflect real experience: Testing during the 7–10 PM congestion window intentionally captures what users actually experience during peak usage. An ISP that delivers 500 Mbps at 2 AM but only 200 Mbps at 8 PM is providing a lower real-world experience than the marketing speed suggests.

How to run an accurate speed test

1. Use Ethernet — connect your computer directly to your router or modem with a Cat5e or Cat6 cable.
2. Disable the VPN — turn off any active VPN client before testing.
3. Close background apps — pause cloud sync (OneDrive, Dropbox, Google Drive), close streaming apps, and pause any downloads or updates.
4. Use a native app — browser-based speed tests cap at approximately 900 Mbps due to JavaScript limitations. For gigabit connections, use a native desktop app.
5. Run 3 consecutive tests — average them to reduce variance. A result consistently below 80% of your plan speed warrants further investigation.
6. Test at different times — compare peak-hour (8 PM) and off-peak (2 AM or morning) results. Large differences indicate network congestion at your ISP.

When your ISP is actually at fault

If you test via wired Ethernet during off-peak hours with no background traffic and consistently get below 70% of your advertised plan speed, the issue is likely with your ISP's delivery, your modem, or the line between your home and the network node. Contact your ISP with documented test results and timestamps. Request a technician visit to check the line signal levels.

Methodology

Speed impact percentages are based on controlled testing methodology comparing baseline wired Ethernet results against the same tests performed under each condition. Results represent typical median impact rather than worst-case scenarios. Individual results may vary by hardware, distance, and ISP infrastructure. Run a speed test on SpeedTestHQ for accurate, unbiased results.

Frequently Asked Questions

Why does my speed test result look so different from my plan speed?

The most common culprits are Wi-Fi and background apps. Testing over 2.4 GHz Wi-Fi can reduce measured speeds by up to 35% versus a wired connection. A VPN running in the background can cut throughput by 30–50%. Close background apps, disable the VPN, and connect via Ethernet to get a result that reflects what your ISP is actually delivering to your home.

Is my speed test result accurate if I run it in a browser?

Browser-based speed tests cap at approximately 900 Mbps due to JavaScript overhead — meaning if you have a gigabit plan, a browser test may underreport your speed by up to 8%. For plans under 500 Mbps, browser tests are sufficiently accurate. For gigabit or multi-gig plans, use a native desktop app to measure the full connection speed.

How do I know if my ISP is throttling my connection?

Run a wired, off-peak Ethernet test and note the result. Then test again at peak hours (7–10 PM). If peak-hour speeds drop 25% or more, that may indicate node congestion at your ISP. To detect selective throttling of specific traffic types (video, gaming), compare a speed test result with a VPN active versus without — if the VPN test is faster, your ISP is likely throttling that traffic type. Some ISPs can throttle speeds by up to 50% on targeted traffic categories.

How many speed tests should I run to get an accurate picture?

Run at least 3 consecutive tests and average the results to reduce variance from momentary network conditions. For a full assessment, test at three times: off-peak morning, afternoon, and peak evening (8 PM). A result consistently below 70–80% of your plan speed across all three windows — on a wired Ethernet connection with background apps closed — is grounds to contact your ISP with documented evidence.