I've done a little work on Blog mining for a large conglomerate. One thing that blog posts have in spades is a remarkably high affect compared to news reporting

I've done a little work on Blog mining for a large conglomerate. One thing that blog posts have in spades is a remarkably high affect compared to news reporting. This enables high quality sentiment scoring to be useful. In one case we found an increase in reporting on a certain topic in the blogosphere coincided with a vast increase in negative sentiment about that subject. Basically the more people were talking about that subject, the more they hated it. Then suddenly a drop-off of activity and the sentiment abruptly changed to positive. Turns out there was a flood of discussion about some upcoming legislation on the topic and the feeling was generally that it was a stupid law that would cost millions of dollars and not benefit anyone. The law failed and discussion was quickly replaced with people back promoting the failed idea...hence the slowdown in activity and the newly positive sentiment. We used a similar technique on a large number of topic areas an fairly comprehensively mined consumer thinking for a fairly major branding campaign. The morale is that the corporate world is listening to what the bloggers say, just in aggregate and not in too much detail.