I understand the concern around the tracking of browsing history, as that's obviously one of the main drivers of ad blockers in the first place!
We definitely need to make sure we deliver on transparency around what data is tracked, how it is stored, and how it is used. We will be releasing this in more detail on our site when we launch.
One key point for you to be aware of is that we do offer the ability to turn off tracking of browsing history. In that case, we'll give you the option to support sites automatically based on sites everyone else spends time with.
Browsing history and/or time spent has to be tracked so the money can be distributed to the right sites. To protect the browsing history:
1) (on the client) Take the domain name (or URL, IP or data point) and add a salt
2) Hash it
3) Add it to a bloom filter (or another type of filter)
4) After enough URLs have been added, submit the filter and salt
5) (on the server) Run the list of participating domains through the filter (take the domain, add salt, hash, and filter)
Some thoughts about this type of scheme: No exact history data would be submitted. You may need multiple filters for domain, time spent, type of interaction, etc... The selection of filter and optimal number of domains (or encoding of data points) might be problematic. Submission can be done at any time (no schedule). The loss of a few submissions won't matter (no need for every domain or a complete history.) Submissions can be delayed until there is enough obscuring data, and filters provide plausible deny ability (about browsing history.) Submissions could be made to trusted 3rd parties (browser makers, large company networks, non-profits like the EFF, etc..) and only the results sent to a central authority. Some websites that were not visited will benefit, but the salt will stop gaming of the domains. Running the list of participating domains might be to demanding to be practical. Groups of domains could have their own payment/subscription filter (news domains, GNU endorsed domains, etc...) distributed to the clients like blocklists.
The biggest problem: Advertisers and ad networks will fight against loosing personally identifiable information and analytics data.
it would register access to the website, add an interaction token to the pay-to database and forget who it was that visited.
if you keep to an honorable "do no evil" philosophy, this should work.
i'm sure the temptation to use a large database of collected visit data to optimize ads is appealing from a financial point of view, but since your product is about avoiding tracking, it should fit and keep it popular enough to remain sustainable.
Expect browser extensions that make that time go to zero.