What latency and jitter mean for your internet connection

Download speed gets most of the attention, but interactive internet use is often more sensitive to delay and consistency.

That is where latency and jitter matter.

Latency is delay

Latency describes how long it takes for traffic to make a round trip between your Mac and a remote destination.

Lower is generally better, but the exact number is less important than context. Distance, routing, Wi-Fi conditions, VPNs, server location, and network congestion can all affect it.

A latency value should not be treated as a universal score. What matters is whether it is appropriate for the connection and whether it suddenly becomes much worse than its normal behavior.

Jitter is variation in delay

If one check takes a short time, the next takes much longer, and the next drops again, the connection has variable delay.

Jitter is a way to summarize that variation.

That inconsistency matters for real-time work because audio, video, remote desktops, and games prefer packets to arrive predictably. A connection with moderate but stable latency can feel better than one whose delay swings constantly.

Why a fast connection can still feel bad

Imagine a connection with high throughput but repeated delay spikes. A speed test may still transfer a lot of data overall, yet a video call can pause or a remote session can feel sticky.

That is one reason speed and reliability are different measurements.

Onlink keeps manual throughput tests separate from its Internet Reliability Score. The reliability model uses latency and jitter alongside availability, incidents, and failed lightweight checks rather than letting a big download number hide instability.

Look for patterns, not one isolated sample

One latency sample can be affected by a temporary route or server condition. A useful history shows whether delay is repeatedly elevated, whether jitter remains high, and whether those periods line up with outages or other incidents.

Longer history can answer questions such as:

• Does the connection become unstable in the evening?
• Did latency get worse after a network change?
• Are high-delay episodes isolated or recurring?
• Do they appear with other evidence that points toward Wi-Fi or the wider internet?

This is more useful than chasing a single “good” or “bad” number.

Loaded responsiveness is another question

Latency while the connection is idle is not the whole story. Delay can increase sharply while the connection is busy.

Onlink’s manual Diagnose includes a short loaded-responsiveness measurement that compares idle behavior with latency during a brief transfer. It is not a full speed test, and automatic incident diagnosis does not run it in the background.

That separation keeps passive monitoring lightweight while still giving you a deeper test when you explicitly ask for one.

Read latency and jitter together

A practical interpretation is:

• steady latency + low jitter: the path is behaving consistently;
• high latency + low jitter: the path is consistently slow to respond;
• normal average latency + high jitter: the connection is inconsistent;
• high latency + high jitter: both delay and stability are degraded.

Those are clues, not final causes. To find where the problem lives, combine them with the connection path described in Wi-Fi vs. router vs. DNS vs. ISP.

Latency and jitter are most useful when they stop being isolated technical numbers and become part of a timeline.