Implementing Plausible on our websites - need your input!

Social
1

Transparency is key at the foundation, and because of this I’d like to solicit your feedback on an implementation we’re planning: collecting analytics from our websites in the most privacy-focused way we can.

If you want to skip my overview and get straight to the post to provide feedback, you can see the full blog post draft here. Please share your feedback in this topic.

We’re looking to understand if anything is unclear in why we’re implementing this and what we’re collecting. All feelings are welcome - whether you’re unsure about or supportive of this, we definitely want to hear how it feels for you.

I’m gonna try to be more brief than the blog… :sweat_smile:

The ‘What’

We will be using a self-hosted version of Plausible, an open source analytics platform (meaning we control the versioning and can fork it if we need to). We will have a publicly available dashboard when we publish the blog that will show you exactly what we see, just like our Home Assistant analytics page.

Here is what we will see as described by Plausible. There is a table in the blog draft that details explicitly what we track and will not track with links to the Plausible documentation with definitions for their metrics.

It’s important to note: this is only for our websites and not any of the projects we maintain.[1]

The ‘Why’

We want to be able to make better decisions about how to update our websites (and, for the merch store, identify items people are visiting but are not purchasing). We cannot do this without analytics. Yes, feedback widgets and community polls are valuable, and we do use them for certain things. But they only capture responses from engaged users who choose to speak up and participate, not those who stop by for a visit.

We chose Plausible because they align with our value of privacy and with the strictest data regulations; this will allow us to identify the needs of all website visitors in an aggregated manner and not rely on data from our active community members.

Share your concerns or questions

We have a few expectations of concerns from people which will be published in a FAQ, but the purpose of this topic is to find out a) if the blog answers most of the concerns and b) if we are missing some things that should be in the FAQ.

Please read the blog post in full.[2] Your concerns or questions will help shape both the blog and FAQ. :folded_hands:

  1. One thing I’d like to call out, which is in the blog: opting out is a manual task. Because there are no cookies, there is no way to do a consent banner; we simply wouldn’t be able to save the request. I’d love to hear your thoughts on this. ↩︎

  2. There is a reference at the very end of the post that calls out where to give us feedback without a link; this will point to a separate forums topic here so that we can have a place to gather feedback because this will be posted on the OHF blog, not the HA blog. ↩︎

13 Likes
3

Damn, i clicked the url above(to the "blog post",before i realized it was a docs.google ... so now Google most likely "tracked" me :laughing: ... so much for privacy :thinking:

2 Likes
4

I don't like this.

However it also means that if you prefer not to be part of the aggregated statistics, you’ll need to block Plausible using your ad blocker rather than clicking a preference box. Many ad blockers already filter it automatically.

I will block any analytics with my uBO.

2 Likes
5

Unfortunately there are some things we just cannot escape - I can guarantee we aren't looking at any of that data to track people, though. :wink:

6

I will block any analytics with my uBO.

From reading the doc, you won't have to; uBO's default lists already block the "Plausible" analytics via a generic URL pattern.

I should note that it is possible to do analytics based on server-side logging of requests, which cannot be blocked (if you access the site, the server knows what you accessed, your IP, and which browser you used). I'd guess that the main reason HA is going with a JS-based solution like Plausible is that the method they're using to host the site uses third-party services that don't provide detailed access logs?

I personally have no problem with HA collecting analytics, but I hope the HA folks are aware that their userbase tends to be on the tech-savvy side, and so JS-based analytics are likely to be biased significantly due to undercounting users with adblockers.

5 Likes
7

Yeah i figured , with your Hidden Profile ... however Google and i.e FaceBook etc already "got You covered" ,Wake up and worry about i.e things you can control, instead of things that really doesn't matter,or infact can contribute to a better software ( on the surface )

8

I block Google and FaceBook too :squinting_face_with_tongue:

1 Like
9

as i said ... In Your Dreams

1 Like
10

Yep! We're aware and it's why we're transparent about knowing it's gonna happen. In general, we hear from those of you who do use those tools when our websites aren't fitting your needs. :grin: Like I stated, this is more to find insights from those who don't speak up.

1 Like
11

There doesn't seems to be anything "controversial" in what you will "collect" ... And then in the end it all boils down to "how do you interpret it" :grin:

1 Like
12

I am fine with tracking of the data.
Just please please don't make me click yet another "I accept cockies/tracking" popup each time I open the site. I get so annoyed by all those popups and extra clicks.

4 Likes
13

I do not want yet more data collected about me. It's your choice to be ok with it, butI if they collect they have to ask for consent.

They don't have to ask, they can just not collect so if you don't want popups, you need to object to collection in the first place.

1 Like
14

Personally, I do not care... but I really do. I never like the idea of being tracked, thus my firewall rules are a little over the top. In this day and age it is not reasonable to believe that you will never be tracked, unless you are a literal cave man... And a caveman would not have HA in his cave!

15

You obviously haven't read the blog.
It is NOT about you!

All that Plausible does is to aggregate the data ALREADY ON THE SERVER!

@MissyQ Mention analytics and the paranoid come out of the walls. Go for it. You have probably already been collecting some of this data manually or in Google Analytics.

1 Like
16

Actually, we disabled Google Analytics on our websites years ago*, and the server-side data we could collect we do not look at. We find the latter to be an invasion of privacy because it holds IP addresses and that's not something we're very comfortable with.

Since Plausible only looks at the IP address for geolocation and then discards it 24 hours later, it's an approach we're much more comfortable with.

*I'll caveat this with yes, we can see some referral traffic (meaning what websites people are coming from when landing on our domains) from Google but it's not very good data and often unreliable.

3 Likes
17

Very good approach + choice yes. :100:

Especially having it exposed publicly for anybody interested to see and have transparency both ways!
Kudos moreover for not using a proxy to try to bypass ad-blockers. :mending_heart:

If it is the case, then it is opt-in by default and super great.
I was about to recommend having a subdomain with something like analytics so that people do not need to opt in to their ad blocker but that solves the issue at the source just perfectly tbh!

It is doable yet not super reliable + quite more complex to implement as a whole. Here we do leave the door open for people to block them, which is also a win-win. :+1:t2:

Moreover, the team probably just want some signal rather than the full picture with minimal implementation/maintainability burden and ease of use.
Publicly exposing server-side analytics would be an entire different can of worms to deal with.

Precisely why Plausible is better than Google Analytics, no need for this one. :hugs:

18

nothing wrong with it, and nowadays you cant do without when streamlining a company. I think home assistant/ohf is reliable enough to not have any concerns about privacy things, unlike other companies. Gonna make things better...

19

You won't have to, don't worry :slight_smile:

20

Just to clarify, the IP is never stored.

The IP is used for doing a geo-location check and then discarded immediately.
The only other time an IP is used is for the unique identifier for a user session that is stored for 24 hours. This is hashed, unable to be reversed and is only used to work out the "unique visitors".

From Plausible: "Raw IP addresses and User-Agent data are never stored."