• Henley Wing

    “While that may lead to a correlation/causation argument, it might be considered a good beginning.”

    This, in essence is the main obstacle in showing any ROI in social media. Especially in offline businesses. How do you know whether a spike in social mentions resulted in extra purchases in your store? How do you attribute that to social, versus an offline ad campaign?

  • http://dragonsearchmarketing.com/ Ric Dragon

    Oops; sorry Henley – missed your comment! Thanks for the contribution. You’re right. It seems to me that it can happen when there’s enough data. In a company like Sony, or CocaCola – you’ve got enough data that when a trend in one place is followed by a trend in another place in multiple instances, the probability (!) that there is causation is much higher. In lower-data scenarios, anomalies can too easily skew the results.