How to fish for insight in the sea of social media

Image courtesy of Tim Parkinson

Too much social media research is limited to automated aggregation and quantification of sentiment.  Its progress is restrained by the misconception that algorithms can replace human analysis.  The real value is hidden in the depths of social context and a more qualitative approach is necessary to find it.  It supplements rather than replaces traditional research.

Listening to what people say online is vital to building more mutualistic relationships. The better brands understand people’s agenda and what has social value, the more engaging and responsible they can be.

The problem with social media research…(and market researchers)

I gave the presentation below ‘Fishing for Insight in a Sea of Social Media’ at the MRS Online Research conference (January 2012).  I muse that if a conversation were a fish, then that makes researchers fishermen (and gourmet chefs): sailing the seas of social media to catch the most insightful conversations before serving them up to nourish clients with naturalistic, socially-enriched insights.

My main point is that automated sentiment analysis is a bit like fish fingers:  A bland amalgamation of conversations that lacks any real flavour (or nutritional value).  You would never come out of a group discussion or survey based research and say ‘well, when you put everything that was said together, I’d say it’s about 40% neutral’….where’s the value in that?

I also suggest market researchers should not ignore social media.  To me, it’s essential if you want to understand people on their terms.  It’s natural, spontaneous and socially loaded.  So it gives you a window into what people value and the positions they’re willing to publicly adopt.  Too many researchers see it as a threat to the status quo, rather than a potential complement to what they already do.

I’m certainly not saying traditional research is fish food:  far from it.  Traditional research gives you control over what you ask and who answers.  I see social media as a complement rather than replacement.

The problem is too many social media research oversells itself. The debate is positioned around conversational coverage and the accuracy of sentiment analytics.  You get automated charts with misleadingly meaningful diagrams (often based on mis-ranked sentiment or  irrelevant sources).  This gives clients the misconception that their in-house or agency teams can make sense of it.  Everyone ends up underestimating the true effort and time necessary to get real value from it.

However, I do believe that bespoke tools are necessary (like Sysomos which is my preferred one).  They effectively let you cast the net out wider than free-aggregation tools and search engines.  They help you manage large volumes of conversations and empower you to shape and refine your hypotheses. This requires trial and error, detailed investigation, qualitative research skills and ultimately time.   If you need quantitatively meaningful data (but don’t have the resources to gather and score it yourself) then companies like Wavemetrix and Integrasco can be necessary (both provide human analysis which increases accuracy).

[slideshare id=11610491&doc=fishingforinsightinaseaofsocialmediaslideshare-120216070208-phpapp02]

I’d love to hear from any believers or non-believers in the potential in social media research…so please do share your thoughts….

5 Responses to “How to fish for insight in the sea of social media”

  1. 40deuce

    Great post Tom.
    What I see as a problem is that people rely a little too much on the tools today. They think that they can type a couple of words into a tool and then have it spit back the big answers, like are we getting ROI from our social program? In reality though, finding real answers takes a real human mind to analyze it. Like you said, tools can be super helpful in gathering, organizing and looking at the data, but it takes a real person to look at that data an make sense of it.
    Also, thanks so much for recommending Sysomos as your tool of choice!

    Cheers,
    Sheldon, community manager for Sysomos

    Reply
    • Tom Woodnutt

      Thanks Sheldon…I agree people often underestimate the necessity of people…machines don’t complain…but they don’t really think either…!

      Reply
  2. LoveStats

    Are you sure that neutral data isn’t helpful? On their own perhaps, but in comparison to something else, they are massively important. If 40% of people are complacent about your brand, but 60% are complacent about a major competitor, I care a lot. On top of that, if we know that 40% of opinions about recommendations are complacent versus 60% of opinions about purchasing, that’s another important piece of information. But i’m just picking at nuances. I’m a firm believer in the right tool for the right job and I think THAT is where we are weak.

    Reply
    • Tom Woodnutt

      I take your point Annie – especially when analysis goes to that more nuanced level that you describe…

      …My – somewhat sweeping – criticism was aimed at analysis that stops short of that level and is based on inaccurate algorithms, an absence of human scrutiny, quality control or due consideration of factors like context.

      The fact many companies and their agencies are using weak tools and / or over-representing their accuracy, is damaging perceptions of its overall potential….

      Reply

Leave a Reply

  • (will not be published)

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>

More posts: