Once again, I got one of these mails from WordpreSs — they’ve determined that this blog is a “commercial site”. They did this once before, and I said “er, no”, and then they went “OK, you’re right”.
Then a few days later, unrelatedly:
I no longer have a blog on wordprEss.com — I selfhost the blogs because their offering was just too restrictive for me — but I kept paying them because I thought I’d support their work. (And they were helpful with customer support for some stuff while I was hosted there, so I had some warm fuzzies.)
But the annoyance from their Jetpack “audit” just pissed me off, so now I’m letting it lapse. So there! *stamps foot*
Before the old stats become irretrievable (I’m running self-hosted stats now), I just downloaded them now to see whether I can do some er insightful er statistics…
Heh, the CSV files are corrupted — when there’s a quote mark in the titles, like the title Foo "Bar" Zot, it shows up in the CSV file as "Foo ""Bar" Zot". That is, as normal the entire field is surrounded by quote marks, so you have to quote the quote marks — which is done by doubling the quote marks (in this CSV dialect). But the second one isn’t, which just breaks everything. *slow clap*
But here’s some charts. Most viewed posts of all time:
Oh, right, the Epstein thing went viral this year. It seems like ages ago.
Where does the traffic come from?
Yes, Google wins here, but wow, Hacker News is really big. But also Twitter, but also because if that Epstein thing. There’s usually very little Twitter traffic.
Countries?
No surprises here.
What kind of stuff is clicked?
Hey, people really like clicking on images on this blog.
Since I’ve been running self-hosted stats the last year, I took the opportunity to do some quick comparison between the Jetpack Stats data and my own, and… yes, there’s big differences. Web stats is very touchy feely now, and there is no “right” way to do it (because of all the AI scrapers that you want to get rid of), but there’s some pretty odd differences.
Overall takeaway is that the Jetpack Stats is about 30% lower than the self-hosted ones… which is, I think, pretty natural, since people who run adblock aren’t counted on Jetpack, but are counted with the self-hosted one.
But there’s odd outliers like Jetpack saying this post has very few readers, while the self-hosted stats says that it’s the most popular post every day almost. 80% of the traffic has a Referrer that points to Duckduckgo, which is odd in itself? Perhaps 90% of people who use Duckduckgo also use adblock? Perhaps it’s all bot traffic that Jetpack successfully filters out? ¯\(ツ)/¯
OK, I just did some further ad-hoc statsing: Out of every fetch of a blog post, 98% are not counted by stats. Out of these 98%, 80% identify themselves as bots, and 20% pretend to be real browsers (or, indeed, are real browsers but just aren’t counted for various reasons).
98% is actually less than I would have guessed — I imagined we were up to multiple nines by now…
Anyway, here’s a list of the most popular blog posts, according to Jetpack Stats:
