Friday, April 03, 2009

Why Google Really Wants Twitter: Real Time Sentiment Analysis Scoring

Imagine the price marketers would pay for a tool that provided instant feedback from millions of people, opposed to the typical focus group. 

In my opinion, Google (NASDAQ:GOOG) wants Twitter's live data, which provides instant sentiment analysis via scoring tools such as Twendz.
Twendz mines Twitter conversations and alerts you to brewing trends, conversation topics and points of view. Most of all, it provides sentiment scoring with a keyword-based approach.

 

Meaningful words in each tweet are compared against a 'dictionary' of thousands of words that are associated with positive or negative sentiment; each word receives a score that, when combined with the other scored words, allows twendz to make an educated guess at the overall tone of a tweet. After Twendz scores a handful of tweets matching certain criteria, it extracts key terms, assigns a tone rating to each of those, and assembles them in a word cloud.
 

Google Trends offers a glimpse into popular search trends, but marketers want to know more than the phrase you're searching for, they want to know your take on a product or person.   Political advisers focus on real-time sentiment displayed during  debate highlights via linear graphs such as Fox News' Voters Voice. The graphs scores indicate the topics from each candidate that resonated with voters,. Imagine a text version of the Voter's Voice graph that provided real time sentiment scoring from everyone in the Twitter community. (Or anyone with a Google account)

You may recall my post about analytics company: Alterian and their survey of CMOs, titled, Calibrate how you operate, which surveyed 400 members of the Chief Marketing Council. Among its notable results, 94% of CMOs are utilizing platforms designed to better capture, qualify, communicate, convert and manage leads and customer contact information.
For demographic-specific sentiment analysis, PR companies such a Vocus offer detailed reports, but Twendz reveals the unflitered opinions of the online community as a whole. Imagine you're the movie producer for Bruno and you want to gauge the public's response after Sacha Baron Cohen appears on the Tonight Show with Jay Leno or the Late Show with David Letterman.

Producers at NBC and CBS could also utilize the data to see which guests generated the largest buzz. Advertising firms and CMOs like McDonald's Mary Dillon could monitor online reputation and detect the most effective segments of their commercials, etc.

Posted By Matt O'Hern at 10:26 AM
Permanent Link: Why Google Really Wants Twitter: Real Time Sentiment Analysis Scoring | Comments (4)

(4) Comments on Why Google Really Wants Twitter: Real Time Sentiment Analysis Scoring

Matt
What you wrote seems very aligned to what Nick Arnett wrote a while back. "massively parrallel self organization of point of view".
Google has mastered relevance which is one 1 criteria for influence (i.e the function behind remembering)..the other one being emotion..and knowing that people liked/didn't like something helps a lot. Combining both would be of tremendous power.

Comments by laurent : Friday, April 03, 2009 at 11:56 AM

If Google gets their hands on this website they might just change search entirely.

Comments by Nick Stamoulis : Friday, April 03, 2009 at 01:16 PM

Thanks for highlighting the value of RT sentiment analysis, an important area with a promising future. I'm afraid I have to disagree with one point though: I doubt the reason Google might be interested in Twitter is because of the real time sentiment analysis potential. Google sits on lots of data that can be extremely powerful for sentiment prediction, i.e. blogs, and news, yet they have shown little serious interest in the arena. I agree the tweet stream is different due to the immediacy of the data, but I doubt the difference is so large as to drive the giant's interest.

Comments by Deep Dhillon : Friday, April 03, 2009 at 03:31 PM

I realize Goodle is very good at what they do, but makes me wonder if mining Twits would be like mining for real gold in an ocean of fools gold.

Comments by Jeff : Saturday, April 04, 2009 at 11:23 AM

Post a Comment