There are quite a few players in the blog measurement space. I have mentioned a few in previous posts. I spoke with Kevin Roe from Umbria and he showed me some of their latest thinking.
Their Buzz Report " sifts through a collection of 14 million blogs, some number of message boards and opinion Web sites. They look for meaning in the sentiment of the poster and even in the design templates chosen. They use this to determine age and gender and offer a type of cluster analysis. Yes, they have a white paper to substantiate their methodology.
Here's how they summarize their Buzz Report:
Determines attributes of speech
3 layers of topical anlaysis
Supervised and unsupervised cluster analysis
Determines attributes of speaker
Age group determination
New and Unique voices assessment
Here's how they describe their technological methodology:
Umbria uses a combination of natural language processing methodologies, machine learning algorithms and business rules to produce its analysis. Natural language processing (NLP) is used to decompose the text analyzed into its various components. For example, NLP is used to identify the paragraphs,sentences, nouns, verbs, adjectives, adverbs, phrases, term clusters, and other textual components of the analyzed text. Machine learning is used to build mathematical models of the relationships between and among the various textual components to accomplish the speaker and content analysis. This includes Bayes classifiers, neural networks, decision trees and other capabilities. Finally, business rules are employed to aid in textual analysis to better understand the nuances of the speech and address many of the challenges of turning the non-grammatically correct online community into actionable research.
Like a lot of "big" players in this space - Cymfony, Biz360, Intelliseek and, even PR Newswire, they have a pretty compelling promise. Big clients who already have traditional media monitoring packages from these guys are ready for these new services. But some will balk at price and scale - "too much for what I need". Umbria's offering is quite reasonable.