What is the difference between bandwidth and latency?
Bandwidth is how much data can move per second. Latency is how long a single piece of data takes to make the trip. They are independent, and confusing them leads people to buy capacity that does not fix their actual problem.
The usual analogy is a road. Bandwidth is the number of lanes; latency is the speed limit and the length of the journey. Adding lanes moves more cars per hour but does not make any individual car arrive sooner.
Which one matters depends entirely on the activity. Downloading large files, streaming high resolution video and backing up data are bandwidth-bound; more capacity genuinely helps. Video calls, online games, remote desktop and even ordinary web browsing are far more sensitive to latency, because each one involves many small round trips and every trip pays the delay.
This is why a connection advertised as very fast can still feel sluggish. A page load may require dozens of sequential requests, and if each costs a significant delay the total is dominated by waiting rather than transferring.
A third measure matters more than most people realise: consistency. Jitter is the variation in latency, and it is what makes calls stutter even when the average looks acceptable. Related is buffer bloat, where a heavily loaded connection queues traffic aggressively and latency rises dramatically during large transfers. That is the mechanism behind a video call degrading exactly when someone else starts a download, and it is fixed by queue management on the router rather than by more bandwidth.
When diagnosing, measure latency while the connection is busy, not while it is idle. The idle figure often looks fine and hides the problem entirely.