I’ve taken to calling social media “the world’s largest focus group,” since I believe it holds great promise as a source of qualitative research for brands, companies and social scientists. In that sense, mining social data is not necessarily the best tool for coming up with answers, but it has the potential to be stellar at identifying better questions, which is the first step to finding those better answers. And, like traditional focus groups, social media can also be a great source for isolating language and ideas.
I don’t believe we are near the day when mining tweets will replace the good-old-fashioned focus group. I wrote about the death of focus groups in 2009, and what I wrote then still holds true today—social media gives you a richer slice of information about that percentage of people who share their experiences with brands online, but not necessarily a richer slice of information about your current or prospective customers. It’s different data, not categorically better data.
I am a trained qualitative researcher and have practiced my trade for nearly two decades. When I walk into a focus group, it is my job to size up the room, identify the personality types I am dealing with, and ask the precise questions to elicit the richest responses from that particular assemblage of personalities. I am trained to notice subtle cues: hesitation, furtive movements, equivocation, averted eyes, flinches and swaggers. All of these verbal and non-verbal cues help me adjust my questioning, my handling of a given individual, even my own body language, in service of getting to what people are really thinking, in addition to what they want to be perceived as thinking. Often, the cognitive disconnect between the two turns out to be the most important insight for my clients.
In the early days of social media (what, four years ago? I was such a boy then) I focused on what social media couldn’t give me—those verbal tics and non-verbal cues that exposed the inner workings of the respondents’ thinking. I believed strongly that social media mining could never replace the focus group precisely because of all that missing metadata.
Today, I think I was wrong about that. In a focus group, you need that metadata, because the 8-12 people sitting around that conference table eating M&M’s are strangers to you, and you have an hour or so to assemble a story based upon characters you’ve never met. All of that metadata that lies outside the transcripts provides essential clues to “cold read” the room, and construct a model in a short period of time with no prior information.
With social media, though—as I’ve come to realize—that metadata is all there: it’s in the public history of the persons tweeting about the brands I’m studying. In a focus group, there are roughly six personality types, and I have to assemble “evidence” on the fly to figure out who fits into what model. I think this is also true on the social web, as well. And I think that crafting those kinds of models is an essential next step in deriving useful insights out of social data.
It’s here that I think the current crop of influence measures might help, if they focused on the right questions. I’m fascinated by the model Klout uses to identify “Klout Styles,” for instance. I’ve been skeptical (not cynical) in this space about various aspects of influence measures, but here is one aspect that I think offers great potential for researchers and marketers. My Klout Style hasn’t changed in six months—my activity pattern identifies me as a “Broadcaster.” I might quibble with the name, but the descriptor associated with that name is pretty accurate, I think. I’d like to think that I engage with people enough not to be a strict “broadcaster” in the pejorative sense, but I am certainly not a curator, or a “celebrity” or a “syndicator.”
I think there is great promise in things like Klout Styles (more than in the “score,” which I continue to see as the worstification kindification of gamification.) I’m not sure there are really 16 kinds of influence styles, though that certainly makes a nice, even square grid to look at. Instead, after giving this some thought, I looked to the focus group—and the styles of people I typically encounter there—and came up with some “Webster Styles” (and you know how stylish I am). If I could categorize content by its source—in other words, as emanating from one of these specific styles—I would have a better context for the data, and have more clues to form better constructs.
So here are the “Webster Styles” of social media brand conversations, a work-in-progress. Feel free to add your own in the comments!
These are the “default behaviors” in a focus group—those persons who do not appear to push an agenda, but do not shy away from questions and will participate freely and easily without needing to lead the discussion.
Every focus group has one of these—the peer group dominator whose opinions are so strongly held (and strongly voiced) that failure to identify and mitigate the effects of this person early on can completely derail the group. We all know people like this—they are lovely people, and often aren’t even aware they are lapsing into this behavior—but they are the ones who speak first after every question, and have the potential to change other people’s answers.
We see this sort in social media, as well, right? Some may see them as bloviating, but that’s a judgment. These personality types have just as much value as the quieter types, but need to be categorized slightly differently for research purposes than they might be for “influence” purposes. These people might answer first in every Tweetchat with strongly held beliefs which then get retweeted and syndicated, which tags them as “influencers” by services like Klout—but (just as in a real focus group) often the group strategy for such persons amongst the other respondents is “appeasement;” i.e., getting them to finish as quickly as possible even if you disagree with them. Which means that “influence” is a bit more temporary than you might think. Let that one roll around your noggin for a while.
Some people are naturally confrontational. This is another shade of “The Dominator,” but in this case, these people don’t just need to be “First” or “Heard,” they need to win. In online conversations, they become lightning rods for contentious debates, which all play a role in things like sentiment analysis around the topics (like brands) that arise in such conversations. They can also bring out other quarrelers, and discourage the less-confrontational from participating in a conversation. In a focus group, I can “fix” this—but online, it just happens. Knowing whether or not a particular brand conversation was being engineered or hijacked by a quarreler is key in parsing and contextualizing the data from that conversation.
You know what is not all that common an activity? Talking about brands and products online. Look at your own conversations—I bet less than 1% of your tweets/Facebook posts reference a product or service. The Promoter is different—they talk a lot about brands and services, and either enrich or instigate conversations about them. But they aren’t mainstream consumers—they are generating disproportionate amounts of content (in any form) about specific brands. Why would I want to know this as a social media researcher? Well, I’m more interested in the opposite type, actually—the person who rarely promotes or discusses specific brands. Because when they do? It means something.
Every focus group has 2-3 of these—the people who go-along-to-get-along, agreeing with what others have said (especially The Dominator—The Quarreler usually instigates either disagreement or silence). There is a clear analog here on the social web—the inveterate retweeter.
Again, you might wonder why I would care about this—but imagine the difference between these two scenarios: A Dominator tweets about a product, sparking a hundred Assenters to spread the idea. That’s a model of the power of the person to spread a message. But imagine the opposite scenario: an Assenter tweets about a brand, and that message is spread by Dominators and Quarrelers alike. Now we are talking about something more powerful—the power of an idea, or the content, or the brand itself. If I’m a brand manager, I’m more interested in that.
It’s tempting to just call these people “quiet”; they contribute relatively less content in both a focus group and online, and they are generally drowned out by the volume produced by the types above. But in qualitative research, it is incredibly dangerous to overlook these people—they often are quiet because they are actually thinking about the conversation and mulling it over before they contribute. A raw mining of social data will miss or de-emphasize these people because they get drowned out, and their Klout Score will reflect that. But I never fail to get a key insight from these people, and being able to extract the opinions of the quiet Contemplators into their own bucket means I won’t miss the ideas they offer.
I’m sure there are others. “The Expert” comes to mind, as well, but functionally they tend to represent as Dominators or Quarrelers, and even that can be contextual depending on their actual level of expertise with a given topic. And I certainly invite you to submit your own—though I think there’s a Ph.D. thesis in here that I don’t have time to write .
Please do not miss this point, however: my interest in these “styles” is not to profile or segment people—it’s to spot conversations that occur out-of-type. What are the brand conversations that turn Responders into Dominators? The ones that silence the Quarrelers? Or the ones that prompt the Assenters to Dissent? In those conversations, qualitative gold lies.
I’m eager to get the tools to suss all this out. And I’m eager to read your comments, below.