New York – As investors continue to seek ever more primary and timely data the focus has shifted from channel checks, surveys and panels into the realm of social media. Here, firms provide analysis of blog aggregator feeds, Twitter, Facebook and Myspace (and even traditional media) to get a rough and ready heat map for particular products, services, brands or companies. This concept is compelling in that it is readily available information and is extremely timely. Still, it needs to be tested to a greater degree before investors can benefit in a meaningful way from this data.
To get at this data a number of text scraping and analyzing products have been developed. Of course, in most cases there is some learning for the assessment tools to do. The most rudimentary of these tools allows text to be interpreted and for specific terms and emotional words to be identified. These tools work well for some applications, especially since they can be amended to include or exclude any contradictory or ancillary terms or phrases. However, many of these textual analysis tools fall short, especially in less formal language prevalent with the online population. Imagine having screens set up for a word like “excellent”. This analysis would fall short wherever, for example, one used a Wayne’s World phrase like, Brand X is excellent. Not!
While we have not talked to everyone in the space, one firm, Peoplebrowsr is of interest, because their approach allows for more accuracy in determining the actual attitude and emotion in a description of the product or service. They explain that they are using a Mechanical Turk. The Mechanical Turk, you may recall, was a chess-playing automaton built for Maria Theresa in 1770. While it was able to beat nearly all who played it, it was found that is was in fact being operated by a person. Predictably, Peoplebrowsr uses people to interpret whether a passage is positive or negative and to gauge its emotional heat. The firm currently has over 10 billion conversations in its database, and has the longest running dataset of twitter conversations we know of.
PeopleBrowsr is currently building an insider database, which records the executive statements and assessments for the Fortune 1000 companies. Despite the large number of conversations in the database, founder Jodee Rich, estimates that there is just enough velocity of data to produce brand sentiment on the top 20 brands. As velocity increases, more brands will be added to the service, until it covers the Fortune 500 companies.
While it is still early days for the assessment of social media data, this is a trend that is picking up momentum. The early adopters are likely to gain an informational advantage in the world social media investment research.