Researchers say they've figured out how to analyze online consumer comments in a way that lets them beat stock-market averages. The economists analyzed nearly 350,000 Amazon.com, epinions.com and Yahoo! reviews of computers, cellphones, shoes and other products and managed a fictional $100 million stock portfolio of the manufacturers. The analytic software noted user ratings and the tone of accompanying text. In rare cases when the software couldn't get a handle on the tone, the text was personally evaluated. The best predictor of a stock-price increase was the volume of user comments. Numerical ratings had no predictive value, nor did positive comments per se, but negative commentary hurt, as did chatter about rival products. Using the data about user-generated content to guide simulated investments, the authors beat the S&P 500 index by 7.9 percentage points annually over the four years studied. "Does Chatter Really Matter? Dynamics of User-Generated Content and Stock Performance," Seshadri Tirunillai and Gerard J. Tellis, Marketing Science (March-April)
The best predictor of a stock-price increase was the volume of user
ReplyDelete