You don't need analytics on your blog
TL;DR for everything below: the only way I track traffic trends on this blog is with vbnla, a hacky little Ruby script that summarizes my Nginx access log. This has been more than sufficient for assessing post popularity and the general geographic and link origins for incoming traffic.
[...]
I originally added analytics (in the form of Google Analytics) to the blog back in 2015, when I was still in college.
My original rationale for adding analytics was to to motivate myself to write posts on a regular schedule: I thought that having a pretty dashboard of (ever increasing, in my natural human desire for attention) visitors would induce additional writing from me each month.
The opposite turned out to be true: like most blogs, nobody really read mine, and my visibility into that demotivated me and made me feel like giving up on a blog entirely. Looking back, this was inevitable, but having analytics made it worse: instead of only evaluating my own desire to write and improve my blogging, I became (resentfully) dependent on a small trickle of statistics.
Ultimately, there’s nothing wrong with deriving some motivation from the attention that can come with blogging (I still do, as noted below). But it’s also the cheapest source of motivation, in the senses of being the easiest to track, least reliable, and ultimately not tied to a passion for the act of writing itself. Turning off the blog’s browser analytics helped me decrease the overall degree of motivation that came from external attention, while increasing the overall degree I felt from both habituation (writing at least one post a month) and post quality (feeling that I was actually expressing something interesting to myself, rather than attempting to track readers’ interests).