• by Erik Sass , Staff Writer @eriksass1, June 12, 2012

Following a moderately disastrous IPO and amid heightened skepticism about the efficacy of social media advertising, Facebook has responded to critics with a new study of social media ROI performed in collaboration with comScore.

The second part of the two-part report, titled “The Power of Like,” released Tuesday, draws on data on earned media exposure from comScore Social Essentials and ad effectiveness data from comScore AdEffx, along with internal analytics from Facebook, to show advertisers how to go beyond simple and less-meaningful metrics like fan acquisition by breaking down social media marketing into three core elements: “fan reach,” “engagement,” and “amplification.”

These three elements work together sequentially to drive sales. For example, in the “fan reach” phase, an advertiser might reach fans with brand messages in their news feed. In the second phase, “engagement, “ the fans talk about this news feed content, and in the third phase, “amplification,” the news feed content spreads to friends, resulting in earned media. Earned media can then be correlated with, among other things, e-commerce or in-store sales.

To illustrate this process, in “The Power of Like 2: How Social Marketing Works,” comScore looked at data from Starbucks, which has one of the largest social media followings, to determine how earned media related to in-store purchase incidence over a four-week timeframe. The results, according to comScore, showed a “statistically significant positive lift in purchase incidence in each of the four weeks following exposure, and an increasing cumulative lift during those time periods.”

The study compared the purchase behaviors of a test group of Facebook fans who were exposed to earned media with Starbucks brand messages, and a control group which wasn’t exposed to the messages. Over a period of four weeks, the nominal lift in the purchase incidence of the test group (compared to the control group) increased from a 0.17 bump in the first week to a 0.36 bump by the second week, 0.5 by the third week, and 0.58 by the fourth week.

In addition to demonstrating that social media messages can indeed drive actual purchases, comScore noted that these results show there is “a latent branding effect that continues to drive increasing lift in purchase behavior weeks following exposure.”