The volume of big data being generated by social network sites (SNS) is increasing significantly. This study seeks to identify the market-applicable insights concerning the text-type big data generated by SNS and to suggest market reaction strategies for responding to signals emerging from big data. Since people can instantly access large amount of online word-of-mouth (e-WoM) contents due to mobile communications, movie sales are influenced significantly from various SNS contents. Based on this phenomenon, we focused on Twitter, one of the most prevalent micro-blogging services. This research conducted a sentiment analysis to determine consumer valences regarding products. This study finds that the extremity of sentiment - as measured by growth speed in the number of positive or negative tweets changed the direction of the tweets' positive or negative effect on revenue regardless of the valence of the word-of-mouth. The implication for SNS marketing professionals will be discussed.