This is a major problem. There's no guarantee the data is real, and most of the people who have to sell their old data clearly don't have much buying power regardless.

But, imagine if they were selling a real-time feed of what they're actually browsing. Imagine if authenticity verification was possible by closed software and hardware (totally not pointing android here). Then the data would definitely be worth something. And the more "clients" you have, the more they're worth.

I think you could easily train machine learning algorithm to do that.