Abstract: System and method for measuring the performance of brands across online and offline consumer conversations. Raw conversational data of online and offline conversations is received and analyzed using proprietary algorithms to extract, for each type of data (online and offline), relevant metrics that reflect the performance of the brand. In a non-limiting example, these metrics include: volume, sentiment, brand sharing and social influence. One of the problems with using raw data is the extreme volatility that will cause the results to sharply vary between a week and the other depending on the most recent activity or the lack of activity. To solve this problem, the embodiments compute for each metric a moving average (MA) formula specific to that metric to reduce the volatility of the results. The formula is computed so that the metric values produced by the formula would have the best/most correlation with factual data/business outcomes (e.g. sales/sale figures, stock price, purchase or dumping of stocks etc.