BrandSavant

Gaining Insight From Social Media Data

Six Degrees Of Social Media Monitoring

by Tom Webster on March 5, 2010

A few days ago I posted my thoughts on social media monitoring for market research.One thing that has become apparent to me over the past few months is that there are a lot of folks using social media monitoring tools to listen for brand mentions, but truly they are capable of so much more with a little forethought and planning. Jason Falls recently posted a survey to determine exactly how people are using these tools (take it!) while others seem to be questioning how the tools are used or constructed. One thing is for sure, however–social media monitoring for brand mentions is not a substitute for research. The mere fact that people are talking about your brand means little unless you can really be sure of the context of those conversations, and equally importantly, who isn’t talking about your brand and why.

Listening to the social web for brand mentions is really just the first “degree” of social media monitoring– a tactical tool for a tactical purpose. To really get the most out of your social media tools, you need to go a few levels deeper–and not all of it can be automated. With that in mind, here are my six degrees of social media monitoring from the consumer insights perspective:

  1. Brand Mentions. This is really the lowest common denominator of social media monitoring, but even without going any deeper, monitoring brand mentions allows your customer service reps to go out and solve problems, meet customers and understand their requirements. On a deeper level, monitoring brand mentions over time allows a company to measure their conversational “velocity” and even take a bankshot reading of the effectiveness of a social media awareness campaign by simply measuring the increase (or decrease) in  brand mentions over time. Monitoring for brand mentions also includes monitoring the specifics around those mentions–is someone having a problem with your product? Is someone having a great experience? All of those tactical interactions really fall under the orbit of this first “degree” of social media monitoring.
  2. Sentiment. Actually, I’m not a huge believer in sentiment analysis–yet–for two reasons: it isn’t yet as accurate as an intern would be, and even if it were–I’m not even sure what you do with it other than track it over time. There is certainly no correlation I am aware of between brand mentions and sentiment, or even “social media” sentiment and actual sentiment. Taking snapshots of sentiment is a lot like day trading–anecdotal events will “spike” sentiment one way or the other over the short term, and while you should never ignore a crisis, I don’t think you need sentiment analysis to tell you if you’re in trouble. Sentiment analysis over the medium and long term, however, may be a useful metric to track the effectiveness of your social media campaigns over time.
  3. The competition. Now we are just starting to scratch the surface of social media monitoring for strategic research purposes. Tracking your own brand mentions is only really interesting if you have other benchmarks with which to compare your data. If your mentions are going down while people are talking more about your competition, then you may have a social media awareness problem. But if mentions in your category overall are declining, you have a very different set of issues. Conversations about problems with competitive products are also extraordinarily useful to track, because you make money by solving those problems–whether they are with your products or someone else’s.
  4. Direct Consumer Needs. Surrounding all of those conversations about your brand and your competition are specific consumer needs and desires–the context that wraps around the mention. Brand mentions about a car, for instance, are generally wrapped around the features and benefits of those cars. These can be parsed out and coded (by humans, folks–is your goal to automate or to understand?) by category to come up with conversational segments. Once you’ve mastered the first degree of monitoring and you’ve reached out to individuals having problems with your seat belts or Von Flavin valves or whatnot, you are ready to take a more strategic look at all of these issues. One of the dangers of a focus group, for instance, is allowing one or two articulate and persuasive people to run roughshod over the other respondents, leaving the undisciplined observer with a false impression of what the groups were about, or what the mainstream opinions really were. In our first “degree,” above, you solve the problems of the squeaky wheels, but in this stage, you determine whether or not you really have a problem. I talked a lot more about this the other day, but once you have a segmented, coded set of “buckets” that your brand conversations fit into, you can take a more macro- view of the immediate issues and concerns people really have with your product or service beyond the simple fire-fighting stage. This step is work, but well worth it.
  5. Indirect Consumer Needs. Once you’ve fought the fires in the first step, and revised your offerings to social media participants based upon the last step, you are ready to steer the ship out into bluer water. Lexus did this when they launched with an extensive market research initiative that went beyond just asking prospective customers how many airbags they wanted. They listened beyond the “car”-focused conversations and really heard what their customers wanted to experience while driving. For the harried, affluent-but-time-starved prospective Lexus customer, that experience had nothing to do with anti-lock brakes, or acceleration or even lower price. Lexus launched their brand on the strength of one word: quiet. By going beyond the four degrees listed above and diving deeper into what the lives of their customers were really like, they determined that the potential Lexus buyer was looking for an oasis from their hectic work and home lives, and what they sold those customers was a quiet, peaceful place to park their tookus while they drove off to the rat race. Advanced social media monitoring is a great way to assess those indirect consumer needs–what else are the people talking about your category talking about? How can you provide the truest definition of value–to delight the customer profitably–by skating to where the puck is going?
  6. Social Needs. This is the stuff that really excites me–the macro-level trends of a population that transcend even the previous discussions and look at the disaggregation and re-aggregation of populations around new and ever-changing societal norms. Sociologists know all about Dunbar’s Number, for instance, but what happens when an entire lifegroup or psychographic cluster of people changes the denominator by shifting the definition of “friend?” When generations of children who met locally on the kickball field now meet virtually on the Arathi Basin, what are their social needs? What do their parents need to know? Why are there so many “Tweetups?” Why does Tony Robbins have almost two million followers on Twitter, but the Dalai Lama has ten percent of that? Why does over 1% of the world’s population play Farmville? The seeds of those answers can be found deep within the unstructured data of the social web. And those are the answers that will power the global economy for years to come.

So get listening, already.

Be Sociable, Share!
  • Pingback: SOB Business Cafe 03-05-10 | Liz Strauss at Successful Blog

  • http://www.jodange.com Larry Levy

    Tom,

    Great post. The industry needs to recognize that the qualitative analysis is as important as the quantitative. Investing in who is influencing,what they are saying and how to best engage them based on what they have said in the past, will become a key theme moving forward.

    Larry

  • http://www.conversition.com Annie Pettit

    Ah…. you have much to learn about market research! Social media data is an astounding data source for research. Researchers (including conversition) have developed scientific methods for creating valid and reliable results just as you might see from survey research. Day after day, we produce social media research for clients who then tell us that the results map to their expectations and past research very closely. Social media data offers huge sample sizes which in turn creates precision.

  • Tom Webster

    Annie–I’m sure I do have much to learn about market research–don’t we all? But I am not sure what you are taking issue with here in this post. Other than noting that sentiment analysis is not quite the holy grail (which few people not involved in selling sentiment analysis products would dispute) I think my post spent lots of time detailing all the ways that social media is, as you put it, an astounding data source for research.

  • http://www.conversition.com Annie Pettit

    Completely agree that sentiment analysis is not the holy grail. Most (all?) market researchers will say that as well. However, some folks WILL call other types of research the holy grail – panel research, phone surveys etc. Every method had its weakness. Because sentiment analysis is perceived to be so new (it isn’t), I think people also perceive it to be less valuable than other types of research (it’s not). That’s where I end up speaking up. Love having discussions with people who know their stuff. :)

  • http://ci.biz360.com Maria Ogneva

    Thanks for a thoughtful post! Even though I work for a social media monitoring platform, which also uses sentiment monitoring as part of our analysis, I definitely agree that sentiment analysis is not a holy grail or a cure-all. It should be used in aggregate. As far as %s we tend to be pretty accurate. If you want to measure sentiment of an individual article, well, you should just read it :) Measurement in aggregate takes a quick pulse of yourself vs. others. However, the industry is still in infancy, so we have quite a few things to work out.

    @Larry – yes, qual is just as important as quant. If you can figure out what the major sub-topics are, that can help guide further research, help uncover emerging trends.

    Overall, I think there’s room for both types of market research – social and traditional, no? Let’s not throw the baby out with the bathwater.

  • http://www.symscio.com/ Mike Layton

    Hi Tom,

    You’ve presented some great examples on how to move beyond mentions and sentiment to pull some real insight from social media. Sentiment is a surface metric and offers limited value until we understand *why* sentiment is deemed positive or negative and how that compares to the competition. #4 and #5 illustrate some creative examples on the wealth of information that exists within social media. I really enjoyed your post.

    Thanks for sharing, Mike.

  • Tom Webster

    Thank you, Mike!

  • http://www.radian6.com Katie Morse

    I have to agree with Maria here, and agree that there are rooms for both types of research. As Annie noted each type of research has it’s own set of pro’s and con’s, and not all types of research are right for all companies.

    Tom – Great post! I think that companies who are wondering “how do we get started in all of this” will find a lot of value in how you’ve segmented out each type.

    Cheers!

    Katie Morse
    @misskatiemo | Radian6

  • Pingback: Social Media Monitoring And Human Business | BrandSavant

  • Tom Webster

    I was writing a response to you, Katie, and before I could catch myself it had become a bit more “ambitious.” So for all of you folks commenting here who are in the social media monitoring space, here are my additional thoughts on the role you have to play in human business.

    Thanks, everybody, for continuing a great discussion!

  • Pingback: The Hidden Bias Of Social Media Sentiment Analysis | BrandSavant

  • Pingback: Recommended reading – 14th March 2010 « Curiously Persistent

  • Pingback: 8 ways law firms can use social media | Creative Development

  • Pingback: ciberesfera » Blog Archive » links for 2010-05-10

  • Pingback: Top Seven B2B Marketing Tweets Of The Week | B2Bbloggers.com - B2B Social Media and Content Marketing

  • Pingback: links for 2010-11-23 « arts info

Previous post:

Next post: