I hate to play devil's advocate, but I see a couple of points worth making. First, IBM's numbers reflected sales that came directly from social media, not those that were influenced by it. Also, IBM's report included numbers from 500 large retailers, not SMBs.
Until a method for better downstream tracking is devised, it will be difficult to determine precisely how much social media influences sales. For example, if I go to Facebook, Twitter or other social network and make an inquiry about a product I'm considering and ask for input, that will likely impact my purchase decision. But, because I don't directly from Facebook or purchase through an onboard shopping cart like Payvment, for example, but instead perform a Google search and visit other sites before making my purchase, it's easy to assume social media had no bearing on the sale, when, in fact, it had great bearing.
To my second point, shopping cart provider Ecwid, which primarily serves SMB retailers, found that, on CyberMonday, sales referred by social networks rose 70% from the previous Monday. It has long been my contention, and Ecwid shares it, that SMBs are able to leverage the power of social media much more so than large retailers.
In my view, when you factor in the downstream impact of social media coupled with SMBs ability to leverage it more successfully, IBM's report is misleading. Perhaps not intentionally so, but at the very least its numbers don't tell the entire story.