Data-mining the Solution for Newspapers?

My friend John Bergquist shot me this link and it got the wheels turning. Please add your thoughts and comments at the end:

http://newsosaur.blogspot.com/2011/04/who-needs-newspapers-one-couples-answer.html

I read this kind of article and see so much opportunity to offer product or service solutions to these "this is the way we've always done it" companies. All it takes is one executive that wants to be a hero for "saving the company" and all of a sudden there's budget for some wild and crazy solutions (That just might actually work).

We've seen this in the social media space- there's local consultants all over the country making a living on business owners that have finally relented and said "we probably oughta start doing this social media thing"...and then they throw money at it. The newspapers are clearly feeling the need to do "something"- that screams opportunity for those that think they have a solution, no?

When I mull over the newspapers problem, it seems to all come down to data...they have a hard time delivering value, because every reader is looking for a different reading experience so they flee to a variety of sources to sort of Jerry-rig their news consumption. What if there was a way the papers could start gathering and leveraging reader data, similar to how Facebook has...

From what I can tell, the user experience on newspaper sites is pretty static, you can search through different topics and sections, but ultimately, the user has to kind of bounce around and find what's interesting to them. 

Is it possible for these sites to start tracking UI data, like cursor behavior, time on page, topic choice, time per article, etc and also aggregating that data with ad clicking behavior...Maybe this is already happening, but it doesn't seem that sites have adopted ad programming that is customized to the reader behavior/preferences...

For example, if over the course of 10 days to 2 weeks, I've spent a majority of my time reading about national politics, gardening and photography, it seems like with just that data, you could compare that to other demographic data  and research to make some assumptions about the reader and other interests/needs they might have. With that info, you can not only play the "curator" role as far as the reading experience they get when they log in, but also more perfectly match them with advertisers. Currently what I see on my local paper and other sites, is just syndicated ads popping up based on keywords in the articles, nothing based on the fact that for the last 4 months on mondays, I've logged in to view the "business section" and "movers and shakers" only. 

Is this space too heavy with ultra high-end algorithm magic, or can techie newcomers with database savvy start to offer these solutions?

Look at the insurance agent getting all geeked out again...