“ From your 202 subscriptions, over the last 30 days you read 3,055 items, starred 28 items, shared 30 items, and emailed 0 items “ . This is what Google Reader trends say about me. What this means to me is that from what I read I consider valuable and worthy to be shared less than 1% of all the posts.

From the 202 subscriptions most of the noise comes from Digg, Reddit, Delicious, Scobleizer’s Twitter Favorites, my own alltop page feed and my own readtwit feed.

What annoys me most are the low percentages that I get from Readtwit and Alltop. I would have thought that if I choose my domains of interest (websites on alltop) and people I follow ( on twitter ) carefully I woud get a noise to signal ratio a lot bigger. It doesn’t seem to be the case.

From all the websites (services) that are trying to improve the situation (noise to signal ratio) it seems to me that only Postrank and Feedly are having a little bit of success (I do use these 2 services on a daily basis). The others such as Lazyfeed, Daily Perfect are still far behind.

I see a possible solution to this mess and I will try to describe it below.

I am pretty sure that Postrank filtration algorithm can be tweaked on a per person basis, but I am almost sure Postrank doesn’t have the resources right now to support this.

I am sure that the aggregation (clustering) technology that Google is using on the News site can be used on the Google Reader but I am also sure that Google has no interest in doing that since their scope is to have their ads displayed on as many pages as possible.

I imagine that Feedly which takes into account at the moment what I shared on Reader and what I twit in the last 30 days to decide what it’s more important to me could base their recommendations also on what my friends on Google Reader network and on Twitter promote.

Now if somehow someone could integrate all these 3 technologies together I know I would pay that company a small amount each week in exchange for a much better signal to noise ratio.