My last boss had an obsession with eBay. And it framed a lot of his thinking about how online news could evolve. For me, the obsession is Netflix. I love it. Everywhere I look on Netflix, I see a newspaper of tomorrow.

Too many services misunderstand user feedback. For example, just because 1,000 people said an article was good doesn’t mean it’s good for me.

Netflix recommends based on what you like, based on your own feedback. Honestly, that might not be exactly how it works according to the algorithm. But that’s how it’s marketed and how it works in my head.

For me, when I say, “Netflix your news,” I mean users give feedback about their interests and then let the system find articles that match those interests.

On every story, readers should be able to rank whether it was of interest. A five-star rating wouldn’t say the story was well written or from a trusted source, just that it was really interesting.

Systems from companies such as Teragram can automatically assign taxonomy keywords to stories based on the content. For example, a story written about a new restaurant to open downtown would automatically be assigned keywords like:

Geography >> Florida >> Sarasota >> Downtown
Arts & Entertainment >> Dining

Let’s say a murder takes place at that same restaurant. Well, the system would probably assign the same keywords as above plus one for murder.

Anyway, with all of these keywords being assigned to stories already, it lets us interpret what exactly a person enjoys reading. If Joe reads that first story about the new restaurant opening downtown, we can guesstimate that it means he might be interested in Downtown Sarasota and Dining. Maybe he gave the story a five-star Netflix rating.

But maybe he read the murder story and found it wasn’t of interest to him. Let’s say he gave the story only one star. From that we would average the scores for each of the story’s three keywords.

Downtown would be left with three stars. Dining would be left with three stars. But murder would get the one-star rating.

Rinse and repeat, as they say. Joe can keep surfing the site and ranking his interest in stories. Maybe he’ll read a lot of stories about Downtown Sarasota and rank all of them high, bringing up the star rating for that keyword.

Eventually we’d get a good picture of what Joe says he’s interested in reading. It’s like we become Joe’s friends, who knows what he likes, so we tell him about news we’ve heard that he’ll find interesting.

Pretty simple, really. But it’s a huge change for news.

For the most part, personalization on news Web sites has meant the reader picks a very general category and then those stories display somewhere on the home page. More often than not, a lot of stories in that very broad category still aren’t interesting, at least not to me.