rationale.md

This document describes the rationale for developing GoatCounter, its goals, and a comparison with existing solutions.

tl;dr ("elevator pitch"):

GoatCounter aims to give meaningful privacy-friendly web analytics for business purposes, while still staying usable for non-technical users to use on personal websites. The choices that currently exist are between freely hosted but with problematic privacy (Google Analytics), hosting your own complex software or paying $19/month (Matomo), or extremely simplistic "vanity statistics" (Fathom).

GoatCounter attempts to strike a good balance between various interests. Major features include a free hosted version so people can easily add analytics to their personal website, an easy to run hosted option, an intuitive user interface, and meaningful statistics that go beyond "vanity stats" but still respect your users' privacy.

Background

I was working on a product idea and wanted to add some basic analytics to measure how many people are visiting the site, and I've also been wanting to add basic analytics to my personal homepage/programming weblog to measure if anyone is reading anything I write.

Even fairly simple analytics are useful to measure things like "what type of content is popular, and should I write more of?", "does it even make sense to distribute a newsletter?", or "how does the redesigned signup button affect signup rates?"

I tried a number of existing solutions, and found existing solutions are either very complex and designed for advanced users, or far too simplistic. In addition almost all hosted solutions are priced for business users (≥$10/month), making it too expensive for personal/hobby use.

What seems to be lacking is a "middle ground" that offers useful statistics to answer business questions, without becoming a specialized marketing tool requiring in-depth training to use effectively.

Furthermore, some tools have privacy issues (especially Google Analytics).

Goals

• Give useful data, while respecting people's privacy. For the most part, it should just "count events" (hence the name).

  There should always be an option to add GoatCounter to your site without requiring a GDPR notice.

• Easy UX; some existing solutions are surprisingly complex, to the point where I wasn't able to figure some basic stuff out.

• Make a web app that I like using (rather than tolerate it). This is a bit subjective, and perhaps my tastes are "old-fashioned", but I find a lot of "modern" UIs annoying. Google Analytics is a good example where pressing the "back button" will often break everything.

• Works well with any browser and assistive technology, whenever possible.

• Free hosted version. Almost all hosted solutions are exclusively oriented towards business use. This makes sense from a business point of view (better to support 100 customers paying $30 each than 1000 paying $3 each), but it does leave a lot of people without a good/affordable solution.

  I think it’s important to make software like this as accessible as feasible to make actual meaningful inroads to “de-Google-fi” the internet a bit, and make pervasive tracking less common. Making it freely available (for personal use) is part of that. In my own online purchasing behaviour I find that even a small $1 or $2 subscription is quite a barrier, especially for personal projects. From what I see, I don’t think my behaviour is an outlier.

  The only other options you have are to use GA, pay upwards of $10/month, or self-host – which also isn't free in terms of hosting costs, setup time, maintenance, etc. Never mind that Average Joe running his photography website probably doesn't have the know-how.

  If you want to make the internet a bit better, then the only real option is to offer a SaaS for free, at least for personal use.

• Easy to run your own. go install zgo.at/goatcounter should be all you need to get started. No need to set up web servers, PHP, MySQL databases, what-have-you (binaries should be provided, too).

  I feel this is an important feature, because "run your own" sounds nice, but it becomes a bit of a niche feature if you need to have a lot of knowledge and spend a lot of time setting everything up. Ain't got no time for that.

Existing solutions

Note: this is an overview that was, as far as I know, reasonable accurate in early 2019 when I was deciding if I wanted to work on GoatCounter, based on a limited set of criteria. It's not intended as an accurate long-term comparison (products evolve). It's retained here only for historical purposes.

Matomo (formerly Piwik)

• I found the UX exceedingly hard to use; stuff like "who is visiting which page from where?" are probably answerable with Matomo, but wasn't able to figure out how exactly.

• Large tracking script (192K, or 57K gzip'd).

• No cheap hosted option (cheapest: $19/month).

• Not "true" open source but "open core", although most premium features aren't really required for most users.

Fathom

• Open version is no longer maintained in favour of a complete (closed source) rewrite, although there is a promise to start maintaining it in the future again (see).

• Hosted is expensive ($14/month).

• Doesn't give referrers per-page (just globally).

• Back button doesn't work.

Open Web Analytics

• No hosted version.

• Somewhat complex UI with more learning curve than needed. Kind of hard to get meaningful data out of it.

• Large-ish payload (73k, 21k gzip'd). While not at large as Matomo, it still more than doubles the size of most pages on my website (which are about 25k, or 8.5k gzip'd, depending on article length).

• It's a large and IMHO somewhat dated PHP codebase. From a quick inspection, it doesn't seem like the kind of codebase I would enjoy working with at all, which means that improving it isn't really an option (unlike e.g. Fathom, which has a much more workable codebase).

• It's still maintained, but seems to be considered a "finished product". Based on the commit log from the last few years I don't expect major changes or additions.

KISSS

https://github.com/kis3/kis3; new project, first commit at around the same time as GoatCounter.

• No hosted option.

• Didn't investigate in-depth, but seems very simplistic from looking at the code?

Ackee

https://github.com/electerious/Ackee

• No hosted version.

• Doesn't display refers per-page; charts don't give a lot of insight; dark UI is annoying.

• IMHO dubious tech stack (NodeJS, MongoDB).

Countly

https://count.ly/

• Has hosted option, but with unclear pricing.

• Seems to be "open core", with quite a few things missing in thet "community edition".

• 32K tracking script (which doesn't seem compressed?)

• Loading the backend means loading almost 12M of data(!!!) It took 2 mins and 45 seconds to finish loading everything(!!!)

• Find it kind of tricky to get useful data out of it.

• IMHO dubious tech stack (NodeJS, MongoDB).

Analysing log files

There are many tools for this, and there have been for decades. There's great options for many cases but all have the same downsides:

• Needs server access.

• Need server logs in specific format (which is not always trivial to set up).

• Hard to get some common data (like screen size).

Non-open solutions

While I prefer an open source solution, I'm not fundamentally opposed to closed solutions.

Google Analytics

By far the most popular way, not in the least because it's one of the few solutions that's free or charge.

• Google is large and collects a massive amount of information for almost every internet user. Many people are uncomfortable with this.

  It's kind of tricky to determine what role Google Analytics plays exactly in this. Google has a single privacy policy for all its products and services; there is no "Google Analytics Privacy policy", just the general one which is rather vague and open-ended on what kind of data it collects exactly. This policy allows Google to "combine the information we collect among our services and across your devices", but it's unclear how this applies to GA.

  There is an option to disable "data sharing" for "Google products & services", which allows Google to combine data. This is enabled by default. It's hard to determine how this data is combined exactly; at least one help page mentions that this enables "Google Analytics collects additional information from the DoubleClick cookie (web activity) and from Device Advertising IDs", although the inline help gives fraud detection as an example.

• Large-ish payload (73k, 28k gzip'd).

• There is no way to add Google Analytics to your website without displaying a GDPR notice. The GA TOS specifically states that you need to have a privacy policy and tell people you're using GA to collect data (in practice, many sites don't do this though).

• The UI is complex (although better than Matomo), and stuff like back button, open in new tab, etc. often don't work. It's not terrible, but also not optimal.

• It's free up to 10 million hits/month, which should cover almost all personal and small-business use.

statcounter

• Cheapest plan is $9/month (the "free" one is

• Lacks per-page referral view. Various UX items seem clunky.

• Tracking script is 32k (10k gzip'd).

Simple Analytics.

• Hosted is expensive (cheapest: $19/month).

• Doesn't give very good/useful data.

• Very small tracking script.

getinsights.io

• Cheapest plan is $12/month

• Didn't look at in detail; was only recently started (the founder emailed me, which is the reason I know about it).

• Very small tracking script.

Other

• I didn't look at quantcast.com, mixpanel.com, similarweb.com in-depth. Both seem very complex and enterprise-y products that don't really overlap much with the intended usage/audience of GoatCounter.