Page scroll depth

[nelsonic]

At present our chosen Analytics platform https://github.com/plausible/analytics does not support page scroll tracking. 😢
There is an open discussion: plausible/analytics#121 but there hasn't been much progress from the author/maintainer: plausible/analytics#121 (comment)

I offered to contribute the code up-stream: plausible/analytics#121 (comment)

We are paying customers (€11/month) and GitHub sponsors ...

We need this feature.
As I wrote in my comment on their discussion forum, we are either going to fork and add the feature
or we will gladly submit it up-stream for everyone to benefit from.

I realise this is a non-trivial amount of work.
There needs to be a new DB table that allows streaming events.
And they need to maintain a full copy of each page to show the page-scroll.
This has important security/privacy considerations too.
Like where will the copies of pages be stored? how is this managed?
If the user-agent is viewing something sensitive like medical, finance or other personal data, how is that treated? 💭
So it would need to be opt-in i.e. not enabled by default.

[nelsonic]

Had a look at https://github.com/adieuadieu/serverless-chrome last night. 👀

Context:

In my mind there are 3 potential approaches to tracking page scroll:

1. Capture capture the page URL, viewport dimensions for the useragent and then the scroll offset.
   This is the most basic approach and will work most of the time for public pages.
   But it won't work for single-page-aps (SPAs) with
2. Take a snapshot (.png) of the page using headless Firefox / Chrome
   with the dimensions corresponding to the screen of the user-agent.
3. Download the html, css and js of the whole page.
   And then render it in an <iframe> to show the interaction with scrolling.

For public pages that are openly accessible on the interwebs there is no privacy concern.
But for Web Apps with personal/confidential data we may be need to add an extra step to anonymise the data. 💭