[fsenart]

I hope you don't mind me if we stay focused on the original question even though all the subjects are interesting per se.

> I do think there is a HUGE difference...

This is off-topic and a matter of perspective.

> I understand that your main concern is not privacy, but the usefulness of such privacy-friendly stats.

It's not about my privacy concerns. It's about the purpose, legitimacy, and effectiveness of a feature. Showing a count of unique/returning visitors is simply a lie; privacy friendliness apart.

> Analytics is a very complex space...

This is off-topic and a matter of approach and experience.

> Privacy-focused tools already clearly state that they are a "simple" alternative and that they offer only basic stats.

This is the central subject of the discussion. They all promote simplicity and coolness. However, simple != erroneous. You can be simple and provide correct information or not provide them at all. Otherwise, there is a problem of ethics and liability.

How can we trust tools supposed to handle our online privacy while at the same time the same tools are pretending something that is not true?

And please, don't get me wrong. It's all about the metric on users. In effect, the same tools without these particular metrics may have an audience striving for such simplicity. However, with these invalid metrics baked in, it seems to be a more opportunistic move than a privacy-focused one.

[XCSme]

I am not going to contradict you regarding the opportunistic nature of many of the privacy-focused analytics tools showing up lately.

> simple != erroneous

What type of analytics are 100% accurate? I am almost certain almost ANY analytics tool is more accurate than the most popular Google Analytics (mostly because it is blocked by adblockers and being so popular it's used as a spam medium). So, I wouldn't really bash any of the simple analytics for being erroneous when GA is the "most" erroneous out of all, yet it is still the most used.

I still believe my initial point stands, in that having somewhat accurate analytics is usually good enough to take good business decisions in most cases, and if you can do that while being more privacy-friendly, why not?

[fsenart]

We may converge. Though, let's don't end up in an ideological battle. Thank you for the ride so far.

It's not about accuracy. In the original question, I've depicted a simple ladder of identity, and we both agree that the most accurate counter would be the counter of logged-in users. Anything after is by definition less accurate. In addition, the argument about adblockers is a matter of popularity and time, take, for example, the many contribs to open blocklist projects trying to blacklist the most popular privacy-focused tools. How ironic! Finally, I won't go into the GA bashing game at least out of respect for those who work on and with it, every day; an because it's out of scope here.

So what can we compare? How can we conclude? Let's focus on the very constituent of a unique/returning user metric.

It's all about semantics. We must first agree on semantic and then compare tools. And we both know what is the semantic of a unique/returning user and what cannot be.

Clearly, there is no privacy-focused tool, to the extent of my knowledge, that can or do provide a unique/returning users metric. Though, the problem is that all of them advertise the opposite, sometimes event viciously.

Any other discussion going beyond the semantic feels like I wanted orange juice; still, you provided me with a blend containing no orange while trying to either convince me that orange isn't that good or misleading me by advertising loudly that the blend contains orange.