Protect user privacy by using Matomo, dumping Google Analytics?

Hi,

I’ve been becoming familiar with the discourse community. I am considering setting up discourse for two completely different communities. I have slowly become familiar with it on other sites like letsencrypt.

I was disappointed that my tracker-blockers reported discourse is using google-analytics.com

Of course, I’m protected from that intrusiveness because I use a blocker, but not everyone will be. And not everyone wants to end up in a Google database, even if it claims that tracking data won’t leak into other divisions.


Has discourse never considered using f/loss Matomo (formerly Piwik)
instead of Google Analytics? Nothing came up in a search on any of these analytics before I posted this question. Matomo is a pretty decent product these days.

1 Like
8 Likes

Hi Michael, thanks for sharing this helpful post. I will again refer to it after I get my own discourse up and running.

What I was asking (sorry if I was not clear) is why discourse itself does not use Matomo on disccourse.org itself, instead of google analytics. That is why I asked this question in meta. Otherwise I suppose I should have asked in support or somewhere else? This is my first post so perhaps I asked in the wrong category?

It would be preferable, at least to me, to use open source GPL’d software on discourse.org, and to take this additional step to protect the privacy of users by not involving google.

Maybe I am wrong but it seems that many who would want to use discourse for their own communities also do not want to involve google collecting data on all their users, and discourse.org should be aligned with that.

Additionally, the quality of the analytics might improve, as google analytics is likely blocked more than Piwik/Matomo.

I have used Piwik/Matomo for 5 years on sites that I operate.

1 Like

Matomo is based on PHP, which means you’d have to prepare the environment and then care about this. Also, more people use Google Analitycs instead of Matomo. Discourse already has its statistics.

Responses so far are not to the question I asked. :anguished:
One non-response even received 6 likes.

I am disappointed after posting my first question here.
“Does your reply improve the conversation in some way?” is missing here.

Perhaps nobody wants to answer my actual question, or I have posted in the wrong forum.

Perhaps not as many people care about this issue as you do?

That does not mean you are wrong in any way, it just means that other people may not necessarily share your concerns or goals.

7 Likes

Maybe more people misunderstood your question…

1 Like

Perhaps. Anyway, the wonderful thing about discourse is that it is very configurable open source software and so none of us have to use it identically. I won’t have to take privacy away from my users. And I appreciate @michaeld for pointing out the howto. cheers.

3 Likes

Exactly, and it’s not like Google Analytics is forced on anyone… certainly not by Discourse.

4 Likes

@codinghorror, I’m not sure you understood my question?

Of course anyone who installs discourse is free not to activate g-a.
But google analytics is forced on everyone who does not actively defend against it on this website, discourse.org That was my concern as a new user and why I posted in meta, not support or suggestions.

And perhaps g-a’s not even that valuable, because discourse itself tracks a lot more about site users activity.

I suspect you’re right. I didn’t understand either until your latest post.

This probably still applies. Many people don’t care (me for instance) and those that do are free to protect themselves as you have.

I don’t know enough about the comparative pros and cons of Matamo vs. GA. Do you know of any good articles? GA is a pretty fundamental tool for a huge portion of our customer base, many don’t have the choice, so we won’t remove it but I guess we could potentially add another option if enough people would like it.

2 Likes

I’m don’t understand why this is a valid reason?

A lot more people use Facebook than Discourse forums. Why bother with Discourse?

1 Like

Hi! Coincidentally I just came across this older Discourse post of mine yesterday and a good comparison of Matomo vs GA at about the same time. (I didn’t find an answer to your question that I was happy with when you first asked and I originally looked.)

“The difference between cloud hosting Matomo Analytics vs Google Analytics is that when you choose Matomo, we acknowledge you own the data and we have no right to access it. This means we can’t on-sell it to third-parties, we can’t claim ownership of it, you can export your data at anytime (how awesome is that!) and you can migrate between cloud hosting and hosting on-premise for ultimate flexibility whenever you want.”

Why Matomo is a serious alternative to Google Analytics 360

And a comparison table: Learn about the differences between Matomo Analytics vs Google Analytics 360

GA is a pretty fundamental tool for a huge portion of our customer base, many don’t have the choice, so we won’t remove

I’m not sure where this extreme suggestion to remove it comes from!

a huge portion of our customer base, many don’t have the choice

I’d like to point out that Discourse itself isn’t offering analytics choice yet. Discourse is biased toward one analytics supplier, somewhat unpopular for its privacy-practices. Why is this necessary or even desirable?

I’d also like to respond to a previous post that suggested Matomo was a problem because it was written in php.

Websites interface with Google Analytics via Javascript tracking code. They can do so identically with javascript and a Matomo hosted solution: either that offered by Matomo, or their own instance of Matomo on their own server. See Matomo javascript

Discourse has no problem clearly declaring various php-based software on this page is integrated. So why raise php as a barrier for offering a choice for Discourse users to use open-source analytics?

Also, here’s an important reason self-hosted Matomo on your own server is beneficial. Google Analytics javascript is now widely blocked by anti-tracking browser plugins. Because Matomo is just js on your own website, you can easily configure it so it won’t be blocked and you get reliable analytics.

Can the way that GA code is inserted simply be generalized? I am thinking this could be as simple as a text box that is initialized with GA code that could be replaced, so that the user can simply supply the javascript analytics tracking code they want - be it Matomo or any other analytics software? Discourse seems to be wonderfully configurable in so many ways. User-definition of analytics javascript seems like a straight forward flexibility extension.

I would suggest that Matomo support should be added to Discourse as it is also open source software, not part of a monopoly trying to control the internet. Matomo allows a great deal of user customization and total control of data as Discourse does. Discourse should add support for Matomo because doing so supports the greater community of open source software, and user control of data. Discourse should add support for Matomo on greater principles.

I’m no ruby coder but I may take a stab at this if nobody else does, as it doesn’t seem to be a big project.

cheers,
Brett

The title, which says “dumping GA” :slight_smile:

Cool. Our roadmap is driven by a combination of demand and customer requests. So far there hasn’t really been either. Definitely not disregarding your arguments, but taking a stab at this yourself is going to be the quickest way forward here.

2 Likes

Fair enuf. If I can’t be bothered as a starting point…

It’s a good excuse to learn something new. :smile:

Brett

1 Like

If you go to your activity summary (accessed under your avatar dropdown) you’ll find everything you ever did here. :slight_smile:

2 Likes

Thank you for this link!

I do not think this activity summary is totally clear though. I looked at this an hour ago and I misinterpreted it. There is no title for the items, which seem to be posts by other people e.g. your response here, without a relationship. It wasn’t obvious what the relationship was to my activity. It was just a bit confusing, to be honest. A brief title of what the items are would be helpful IMHO.

I’m sure discobot taught me this many months ago and I forgot… :grin:

Has the rapid evolution of Google Analytics with v4 changed your mind at all? Specifically Google’s analytics product is capable of doing a ton more than it used to. Moving away from sessions towards events is a big deal.

Have you made any progress here?