Jan 8, 2010

Going senti

Paul Greenberg is an analyst in the world of CRM tools, with a book on the topic, to his credit. In a recent ZDNet blog post, he added Cognizant's Social CRM ventures to his annual CRM watchlist, saying:
This is one of my most out-of-place and yet most appropriate picks for 2010.[...]The fact that they have a practice is well and good but they back the practice up with content and that makes a huge difference. They are actually developing a Social CRM framework for both technology and strategy and investing resources in products that can support them.
Further he adds:
For example, they have a sentiment analyzer that I saw not too long ago that is a solid entry into the tool belt. This is not a commercial product, but a tool for their work and their clients to use. Honestly, with not much more work, it could be a very salable sentiment analysis tool - competitive in most ways.
I quote these lines because my group is responsible for creating the Sentiment Analyzer quoted above. I've been working on this for the last couple of years and it is interesting to finally see some non-traditional applications of sentiment analysis come to the fore. (Usually, they have revolved around brand monitoring and reputation analysis.)

The jury is still out on what constitutes Social CRM and its benefits beyond simply adding another set of channels to existing CRM systems. People like Cognizant's A. Prem have been writing and talking about this fledgling area, but as someone involved in text analytics , I am interested in seeing how this evolves in the future.

4 comments:

Abhishek said...

So what you are working on is a way to get sentiment from text? Or is that an application of your work? And what do you mean by sentiment? Like a scale between "satisfied" and "dissatisfied"?

J Ramanand said...

This probably deserves a separate post to cover all the points, but simply, we try using natural language processing techniques to identify good/bad/ugly opinions from text. one way is to look at it as a text classification problem - this review is good, that is bad, that one was mixed.

Will try and explain this separately.

Seth Grimes said...

Abhishek, I wrote a number of articles on sentiment analysis that could be a good start for you --

http://www.b-eye-network.com/view/6744

http://www.b-eye-network.com/view/6897

http://clarabridge.com/default.aspx?tabid=137&ModuleID=635&ArticleID=722

J Ramanand said...

Seth: Wow, thanks for dropping by.

Abhishek: Seth Grimes is one of the leading analysts in Text Analytics and puts out a very useful survey of the domain each year. You could hardly find a better introduction to various topics in the area than those written by him.