Currently, there are a bunch of privacy-centric web analytics solutions out there. They all seem very similar. I would love to see a side-by-side comparison of the top 10.
seems disingenuous to keep calling things ‘alternatives to Google Analytics’ when there isn’t remotely feature parity.
It might be for a lot of people. Some claim there isn’t feature parity between LibreOffice and Excel and yet a lot of people use it without bigger problems :)
I’m with you here. I used Google Analytics for a few years before I finally managed to excise Google from my life (mostly) and honestly I used it for a blog and I looked at like 2 views ever, maybe 3.
I’m glad of that. Many GA users only desire a subset of its features. The superset of GA’s features are powered by user privacy violation.
These are indeed alternatives to GA and users can make their own choice as to whether the alternatives’ feature sets are sufficiently rich.
I guess it’s assumed that no one has access to the web server logs anymore? Is what why JS based analytics is so popular?
That, and there’s additional information available to JS that isn’t in server logs, such as the viewport size, and whether or not the user is currently still “on” a page.
To your point: yes, normal httpd logs are indeed more difficult to process in highly-distributed systems. My personal site is mostly statically-generated and served through a very thin layer in Lambda. It’s much easier to integrate JS here than to process the logs from Cloudwatch.
I think it’s really good for people to prefer log analysis as a first choice, but to @scoates point that doesn’t nearly cover all the angles and deep analytics that GOOG does.
That said, I think more people should be aware that they may not NEED real time deep analytics. I’m likely going with log analysis in my blog refactor from Wordpress.com back to Pelican/Kubernetes.
I really wanted to do install this but then running a Clickhouse DB + Postgres server dependencies didn’t make sense for my small static blog. I also explored Fathom which is a decent alternative but they do tracking via cookies for the free version but not for the paid version. Such kind of OSS projects just lose my interest, TBH. (Nothing bad, but I won’t personally like)
I eventually found shynet which I’ve been running since a month and pretty happy with it. Does it’s job and just bare minimum tracking.
I was considering Fathom and Matomo for my sites’ analytics as of late, today I found Plausible through this post, and now I found shynet thanks to you, and I find myself agreeing with your points, so I’ll probably try shynet first.
There’s one thing about Fathom/Plausible/shynet that I find interesting, and that is that all 3 designs are very similar, all three show similarly-distributed dashboards, with the same metrics, etc.
I’ve never used Google Analytics but, is this because GA pioneered that design and all these “alternatives” followed it?
Creator of Shynet here! The monitoring tools you mention do all have similar aesthetics—speaking for Shynet, it was mostly inspired by Fathom (although I designed its CSS library myself, so there are still some key differences).
Happy to answer any questions you might have about Shynet, either here or on GitHub.
Note that unlike Plausible and Fathom, Shynet is not a SaaS—there is no way to use Shynet except to self host it. So the self-hosted experience is the full experience, not a half-baked FOSS version!
That’s one of the main criticisms I had towards Fathom and their v1/v2 bs, but seems like Plausible is giving you the choice without taking out any features.
Ironically if you put github.com into Blacklight you’ll see it sends data to Google Analytics.
(Not commenting on the project, which is most welcome, just highlighting the problem)
Another (newish) alternative I don’t see mentioned here: Umami. (Note: alternative, not replacement. It’s much simpler than what GA/GTM does).
Previous discussions: 1, 2
What they meant by “lightweight” here? 1KB? resource usage is the major concern, right?