Quick answer: A good VPN should keep your base speed within ~80–90% on nearby servers and ~50–70% on distant ones, with stable latency and minimal jitter. If you see worse results, double-check protocol (NordLynx/WireGuard-family usually wins), server load, and whether your device/ISP is the bottleneck.

What a VPN speed test actually measures

Most people look at a single “result” and call it a day. In reality a proper test is a set of measurements taken under controlled conditions. Here’s what matters:

• Download throughput (Mb/s): sustained data pulled from the internet. Impacts streaming, downloads, updates.
• Upload throughput: pushing data outward — video calls, cloud backups, sending large files.
• Latency (ping, ms): the round-trip time for a tiny packet. Crucial for gaming, remote work, SSH/RDP.
• Jitter (ms): the variability of latency. Even with a low average ping, high jitter wrecks calls and gaming.
• Packet loss (%): lost packets trigger retransmissions that tank real-world speed.
• Consistency (stability over time): one-second peaks look pretty but don’t help your Netflix stream.

How to run a realistic VPN speed test (step by step)

1. Baseline first: disconnect the VPN and run 2–3 tests on a reputable test target. Note down download, upload, ping.
2. Pick a nearby VPN server: ideally within your country or neighbouring region to minimise distance and hops.
3. Select the modern protocol: on most providers that’s WireGuard or a variant (e.g., NordLynx/WireGuard family). Avoid old PPTP/L2TP.
4. Run multiple passes: three tests per location. Average the values. If they vary wildly, note jitter/server load.
5. Test distance properly: repeat with a mid-range server (e.g., another EU country) and a far one (e.g., US↔EU).
6. Mirror real usage: if you care about streaming, test to a streaming CDN; for work, test during your meeting hours.
7. Record conditions: device, OS, adapter (Wi-Fi vs Ethernet), time of day, and server label. This helps reproduce results later.

How to read the numbers

Don’t chase a single “top speed” screenshot. Use these sanity ranges instead (assuming a decent base line and a modern VPN):

• Nearby server: 80–90% of your baseline download is normal; 60–75% is acceptable; below 60% suggests congestion or a sub-optimal protocol.
• Mid-distance: 60–80% with a moderate ping increase (e.g., +20–40 ms).
• Long-haul: 50–70% and a much higher ping; jitter becomes the deciding factor for calls and gaming.

Throughput is only half of the story. If latency spikes by 150 ms during a stream or frames drop in calls, switch servers even if the raw Mbps looks “fine”.

12 proven ways to make your VPN faster

1. Use WireGuard-family protocols (e.g., NordLynx). They have leaner crypto handshakes and kernel-level implementations.
2. Switch to a nearer server. Physical distance ≈ more hops ≈ more latency and potential bottlenecks.
3. Change server cluster, not only the country. Providers often list multiple cities; try a less busy one.
4. Prefer Ethernet over Wi-Fi or at least 5 GHz Wi-Fi with strong signal. Your adapter can be the bottleneck.
5. Close background syncs (cloud drives, Steam/console updates) that chew upload and inflate latency.
6. Reboot the router/modem to clear bufferbloat. Consider a QoS/smart-queue setup on supported routers.
7. Disable battery/performance savers that throttle the CPU or NIC on laptops/phones.
8. Try another port (e.g., UDP vs TCP) if the ISP shapes or blocks traffic on specific ports.
9. Split Tunneling: route heavy but latency-insensitive tasks through VPN, keep voice/game outside when appropriate (learn more).
10. Pick the right server type for your goal — some providers expose “Streaming” or “P2P” labels (which server to choose).
11. Check for DNS leaks & slow resolvers — bad DNS can “feel” like a slow VPN (DNS leak guide).
12. Update the app/OS — performance patches land often, especially for new protocols.

Video: quick setup & best practices

Watch on YouTube (NordVPN official).

Common pitfalls that ruin speed tests

• Comparing different targets. Run baseline and VPN tests against the same test server/CDN.
• One-off screenshots. Always average multiple runs and include variability (stdev) in your notes.
• Testing on crowded Wi-Fi. Saturday evening in a flat full of devices ≠ fair network conditions.
• Ignoring upload & jitter. Video calls die from jitter far more often than from raw download deficits.
• Old protocols. If you’re stuck on OpenVPN-TCP everywhere, you’re leaving a lot of speed on the table.

When a slower VPN is still the right choice

Security features like Double VPN or multi-hop add overhead by design. For journalism or high-risk scenarios, the privacy gain outweighs a 20–40% speed hit. For streaming or gaming, choose single-hop/WireGuard-family and a close server instead.