Social Media

4 Social Media Mining Metrics

Posted by Daniel Flamberg on June 15th, 2010 at 9:51 am

Social media potentially offers clues about customer mindsets, sensibilitities and brand affinity and awareness. The trick, in mining the flood of conversations, is to separate the abundant noise from the not-so-obvious signals and to analyze the yield in ways that offer marketers actionable intelligence to identify competitive strengths or vulnerabilities, shape messages, identify informal opinion leaders and influencers or suggest the best choice of media channels.

In undertaking this labor-intensive effort you need two key things; tools and skillful analysts who can sift through the conversations, validate or invalidate the built-in assumptions that are baked into many of the tools and to find the insights that will be valuable for brands and their marketers. Mining social media cannot be fully automated.

But the trolling can be. An array of free tools can get you started. My favorites for scanning the broad waterfront are Social Mention, Sency, YackTrack and Addictomatic. For peering into the blogosphere Technorati and Google Blogs are the most reliable tools with the deepest reach, though results are mostly a function of the key search terms and phrases used. Since each one is built using a different logic, savvy social surfers use several so they can get a broad sweep of the social graph, expose different facets of the social media universe and rely on each tool compensating for the shortcomings of the other.

In assessing and using the intelligence gained, I am evolving four key metrics or dimensions to help clients understand and use whatever we turn up.

VOLUME. Count the total number of conversations and the relative size of the conversations to calculate baseline awareness. By comparing this number relative to conversations about competitors or about the business vertical, you can understand the relative positioning of your brand. By analyzing the volume by audience segment or over time we can plot the impact of marketing campaigns, promotions, social media activity or news coverage. With a clear sense of awareness and positioning, all kinds of marketing strategy and tactics can be brought to bear.

SENTIMENT. Do they care and are they for you or against you. This vector seeks to understand receptivity to and perceptions of brands. Most of the free tools use baked-in business rules about word proximity and phraseology to determine sentiment. This is useful but not necessarily accurate since it depends on tables of words and phrases pre-determined to be positive or negative.

As a result most tools return bell curve results with the hump in the middle labeled “neutral” which means that the data scanned doesn’t have a preponderance of good or bad words associated with your brand. And while its true that most consumers are ambivalent about most brands; it would be a mistake to accept a “neutral” rating on its face. Sifting thru conversations is required to separate “machine neutrality” from genuine neutrality.

INTENSITY. Borrowing technique from signals intelligence analysis, this dimension seeks to understand where the conversation originates, who responds and if there are changes at specific times or over defined time periods. Gathering this data can suggest seasonality; trigger-events or begin to identify opinion-leaders and market-makers.

IOLs. Informal opinion leaders are bloggers, tweeters, videographers, uploaders, commentators, friends or frequent site visitors who direct, distract, side track, explain or enrich the online conversation. Understanding who they are, what opinions or perspectives they represent and gauging their reach, their relative circle of influence and the consistency of their POVs guides marketers in shaping media and PR outreach. IOLs are the poor man’s focus group to test new initiatives, float trial balloons or drive instant feedback.

These four vectors shape the utility of social media mining and transform raw data into useful intelligence that can be nimbly applied to messaging, marketing and media. All four are evolving as the tool sets and the analysts using them become more familiar with the social universe and how it impacts brands and business categories. Over time, norms and best practices will emerge. But for the moment, these four pathways provide the best lines to understand and utilize what customers are saying about us and our brands.

2 Responses to “4 Social Media Mining Metrics”

  1. Rob Diana says:

    First, thanks for using and mentioning YackTrack! I like the intensity metric you present, as I think many people tend to overlook the power of that data. Social sites have a very heavy time component, but people want to look at aggregate trends. The big problem is that sites like Twitter use a stream of data, so when you do something is just as important as what you did.

    Opinion Leaders are also a controversial topic. Some people want to go after big influencers (or people heavily followed on Twitter), but you really need to look at your target audience and determine who to engage. I think that frequent visitors or heavy users or too often overlooked as influencers.

  2. I agree with your statements and theory. But your measurements are all pretty complex and hard to calculate and measure. I think it is also worth to gather and analyze the stuff, that is much easier to track.

    We are creating a social media metrics application at that will centralize all your web stats in one place. We only have twitter, facebook and so far. But I would love to get feedback from you guys about it, as soon as it gets live in the next weeks.

    Isn't that also something, that can add value to your path?

Leave a comment