Jim Tobin is president of Ignite Social Media, where he works work with clients including Microsoft, Intel, Nature Made, The Body Shop, Disney and more implementing social media marketing strategies. He is also author of the book Social Media is a Cocktail Party: Why You Already Know the Rules of Social Media Marketing.
Over the last three years, social media marketers have gotten a lot more sophisticated about the programs they deploy and how they’re measured. Platforms like Sysomos and Radian6 have become vital tools in understanding not only the social universe in which you operate, but how that universe responds to your brand.
But for all of our success, we’re still largely entering strings of Boolean variables into a tool and waiting for matching results to roll in. Most tools have added sentiment processing, but that clearly has a long way to go. Beyond sentiment, however, how are these tools going to evolve to provide more insights? The answer is with math.
Cluster Analysis Shows Promise
Cluster analysis is one of the features that many tools are adding. Basically, it involves complex mathematical computations that analyze how certain words are gathering — or “clustering” — relative to your search topic. It finds the words that are mostly likely to be associated with your search word. This may provide unexpected insight into what’s being said about you. In fact, some articles suggest it’s a way to predict record sales two weeks before they happen.
The downside to cluster analysis is that it’s complex. Nilesh Bansal, the CTO of Sysomos, likes cluster analysis, but worries it’s not that easy: “We very often find ourselves explaining to clients what these features mean and how to use them best. For a full blown unguided cluster analysis, a team of technical analysts will be required and it cannot be point and click.”
Semantic Analysis Will Also Be Key
When it’s not just where words appear but what those words mean that you’re trying to decipher, semantic analysis will become key. Semantic analysis strives to understand what words mean in context to provide deeper insights.
Marcel LeBrun, CEO of Radian6, likes to think of various “analytic lenses” to apply to the vast data they collect. LeBrun says he doesn’t see cluster analysis as a direction to move in, but as one of the tools in the analyst’s toolbox.
He explains that cluster analysis works on the basis of math, but it can’t differentiate apple (the fruit) from Apple (the company). Semantic technology, he says, can add additional insight because of its ability to identify entities and nuances in language.
Depth vs. Simplicity
For a while, the relative strength of social media monitoring tools was dependent on how much data they indexed, and how far back they had stored it. Today, major competitors have largely mastered indexing the stream (private Facebook updates being the obvious exception).
So the new race is on to display and interpret that data better than the next guy. Every monitoring tool is rushing to improve their dashboards, and competitors like RowFeeder are trying different approaches.
All of these companies are working with very smart people who can make data dance. Sysomos’ Bansal published a paper on cluster analysis applied to social data three years ago, indicating that they are ahead of the curve. The real challenge is to apply these complex lenses to the data in a way that lets us non-PhD holding marketers understand it at a glance — and to do it flexibly enough for different monitoring objectives.
Through cluster analysis, semantic analysis and enhanced dashboards, we’re likely to see an aggressive monitoring “arms race” over the next year. For the companies involved, nothing short of market share gains are at stake.
More Business Resources from Mashable:
- 5 Key Benefits of Monitoring Your Client’s Brand on Social Media
- 5 Tips for Managing Your Company’s Brand on the Web
- 10 Steps for Successful Social Media Monitoring
- 5 Tips for Aspiring Social Media Marketers
- Is Social Media Failing to Produce Business Leads?
Reviews: Facebook, Radian6More About: analysis, business, cluster analysis, monitoring, semantic analysis, social media, social media monitoring
For more Business coverage:
- Follow Mashable Business on Twitter
- Become a Fan on Facebook
- Subscribe to the Business channel
- Download our free apps for iPhone and iPad
![]()
![]()
![]()
![]()
![]()
![]()
![]()
![]()
![]()
![]()
![]()