Web Analytics are helpful, but not helpful enough for me to warrant quadrupling my blog's bandwidth costs.
I like having a general idea of how people are interacting with my sites, and how many people see them. Until I made this blog (and had a slight focus on low page size), I didn't really care to check how bulky Google Analytics was. Now I know.
Adding GA to a website will cost you around 45KB of network bandwidth if you use the recommended Google Tag Manager way.
On the extreme other end of the spectrum, Plausible Analytics weighs in at under 1KB, and offers a lot (most?) of the same information, with slight stat skewing; allegedly Plausible reports slightly higher (15%-ish) unique viewers, probably because it's a little more privacy-oriented and not nearly as aggressive in its session/user tracking.
Plasubile sounds great, but I'm garbage at using Docker and didn't want to pay for the SaaS version, so I installed Umami, which is almost as lightweight (I think the analytics script ends up being around 2KB) and can use MySQL. The rest of those analytics softwares seem to use Postgres, which I don't have installed and configured (and didn't feel like doing so)