Tuesday 7 September 2010

Social media market research from an amateur perspective (1): some examples of online communities

I thought I'd just throw a few things together about some of the developments in social media market research over the last few months, based on a few articles and blogs I've been reading. It's not in any particular order, and is as much for my own reference as anything else, so don't get too excited if you happen to have stumbled across it!

Cecile Lux and Doron Meyassed, from Danone and Promise Communities respectively, gave a paper entitled The future of community based researchThey make the point that "the consumer is increasingly wanting to have their say". The sheer volume of online opinions support this statement; apart from all the obvious media for broadcasting opinions like blogs and forums, feedback on brands, services, and life experiences via all sorts of other channels - like reviews on Amazon or LondonEating, comments on online articles, and sharing titbits of knowledge through all kinds of crowdsourcing channels like Facebook group Secret London or, as in the examples given by Lux and Meyassed's paper, TripAdvisor or Wikipedia. They assert that "status is increasingly driven by what we share, not what we own" (handy for me - I don't own much, but love sharing secrets!)

They also assert that research communities will become more decentralised, with multiple business departments having direct access to the community. This sets my amateur alarm bells ringing. What good can be gained by HR and business development staff asking direct questions to a focus group? I can imagine knee-jerk reactions taking place following one forthright opinion on pricing or business ethics, for example. Would it not be better to leave the moderation and analysis of the community's opinions to the insight professionals? By all means there should be direct contact and input from other departments, but I would have thought it best to leave the overall control to the research team. Hopefully I'm missing their point.

The key point that is made, and the one which seems to be key to the whole concept of online communities as opposed to mere online panels, is the level of interaction and engagement with the consumers. Rather than simply being given a series of surveys to fill in, the participants are encouraged to come up with their own inspiration for the future of the product in question. Thus the brand can use crowdsourcing to generate innovation for their product, marketing, or anything else - everything will be much more open ended. Mircosites will encourage the community members to feel that they are making a difference and go out of their way to come up with insights and innovation in ways that they might not have if they felt they were being used. Research seems to be going the way of marketing where social media is concerned: the marketer cannot only take, but must give something back. In the same way that consumers will only follow a corporate Twitter account if they feel there is something in it for them, the community must fulfil some sort of promise to the consumer - giving them a reason to log on to the community and participate on a regular basis.

Tom Ewing in his paper "Cultures of Collaboration" (from which much of this post was derived) suggests that a consumer's default view of a research community could differ from person to person - so that their perspective is based on their own personal preferences, contacts and content, rather like the way a Facebook news feed runs now. This sounds like an interesting idea. Increasingly, since the push 

Can research learn anything from quantum mechanics?
It occurs to me that research could learn a lot from quantum mechanics. One of the most fundamental aspects of quantum mechanics is the double headed beast of the observer paradox, and the measurement problem (linked to Heinberg's Uncertainty Principle and Schrödinger's Cat). Basically, between them they say this: that you cannot determine a result until you take a measurement, but the very taking of the measurement affects the result itself (I've tried to explain it better with a pub analogy here). Whilst it's only noticeable when the scales are very small, these effects are universal.

The tenuous link to research, therefore, is that any measurement that a researcher takes will inevitably affect the result itself. Even in the context of making a passive observation of an online community, the very fact that the participants know that Big Brother is running their every word through text analytics software may have some implications on their responses to prompted questions, or even to open-ended creative thinking. In quantum mechanics, there's no "quick fix" - the errors are fundamental to the science - I wonder if the observer principle can be directly applied. I suspect that even the swishest of technologies can't escape the effects; in he offline world, I imagine that even those cameras which participants wear to allow researchers to analyse their eye movements in supermarkets are still liable to the observer effect as shoppers make conscious efforts not to look too hard at the bargains? Similarly, no matter how much online research community owners step back, the participants will always know why they are there, unless the environment is allowed to degenerate so much that the forum subject matter moves so far away from the guided topics that even the participants forget what they are doing logging on.

Facebook monitoring
In the last few months Facebook has allowed keyword searches on status updates of everyone whose privacy settings are set to public. This means that you can search for what people are saying on Facebook, across the world. For example, what is being said about social media monitoring at the moment? (You'll have to be logged in to Facebook for the link to work). In true stalker style, I now know that a Canadian girl called Erin Brown needs to learn about social media monitoring software for work ("oh god" is her reaction). How about something a bit more brand-friendly? Over the last few months, a brand who have taken a social media hammering at the moment are BP; a simple Facebook search on "BP" would reveal to them what is being said at the moment, in real time.

This is immensely powerful stuff. It would be easy to write this off as just another way of monitoring alongside forums, Twitter and blogs, but I would argue that being able to monitor Facebook statuses is more powerful than all of those, as it displays people in their most "natural" habitat, and perhaps displaying online behaviour closest to their offline ("real") personae.

Of course, arguments still rage over whether whether Facebook's rather fast-and-loose approach to privacy settings are infringing on human rights - the goalposts have moved significantly since I joined in 2007, when it was basically a way of sending messages and updates to your closed group of friends, and nothing was shared with people I didn't know. Furthermore, many people have issues with the ways monitoring companies go about harvesting data, and most importantly how that data is used: this reached the Mail on Sunday recently. There's also an interesting riposte here from a Alex Brown of Virgin Media. My own take on it is that since those Facebook goalposts have started moving around quicker than a bat in a dustbin, users feel a  little disgruntled when their content is going further than expected - particularly when it's rubbed in that they're being followed. As far as pure monitoring is concerned, I would say that as far as possible, the guidelines set down by the MRS Code of Conduct and ESOMAR should be adhered to as far as possible; the first principle of the Code, relating to "informed consent", may need rewriting or adapting for monitoring situations, as clearly the explicit consent of Facebook users cannot be sought, although it could be argued that the fact that the users have opted to make their content private implies consent for companies to put it to good business use. I'm not so sure about this, but feel it's acceptable so long as all data is anonymised, and, where research agencies are concerned, is only used for research purposes and not passed on to the client for marketing or other purposes (like customer services in the case of BT in the Mail article), but only used to provide aggregated research data - in keeping with the spirit of the code as it stands.

See Jeffery Henning's summary and Ray Poynter's post on social media monitoring ethics for more in-depth views on the subject.

I digress. Whatever the ethics of listening to Facebook conversations, the data is now available to be viewed, although my understanding that the API does not allow this data to be automatically exported or cached (perhaps someone can put me right on this? I'm not a developer and got a nosebleed when I stuck my nose in the developers' section of Facebook the other day). These are publicly available opinions, which, as far as Facebook are concerned, users have elected to make public for anyone to "listen" to. And yet I would speculate that attitudes towards Facebook, as opposed to Twitter or forums, say, mean that Facebook personae are closer to "real life" than any other online profiles.

Let me explain what I mean. On Twitter or LinkedIn, one sets up a profile attached to one's real name and image, and the default - and ethos of the networks - is that information is viewable by anyone. To a certain extent, therefore, there could be the online equivalent of "good behaviour for the mother-in-law"; polite, measured, intelligent, but not necessarily spontaneous, as many users will bottle up true opinions for fear of jeopardising their chances of impressing the mother-in-law (or, perhaps more pertinently, potential employers). On the other hand, it's well documented that people posting under anonymous profiles on forums and social networks can often give overly aggressive, controversial, speculative, or simply wrong opinions.

Someone famously described the difference between Facebook and Twitter thus: "Facebook is for people you went to school with. Twitter is for people you wished you went to school with." (I don't know who came up with it first...possibly this guy.) It's a nice analogy.

Co-creativity and innovation by collaboration
One excellent example of co-creativity is Coproducer. The brainchild of Yougov boss Stephan Shakespeare, it is a very simple concept: creation of an entire film by means of creative collaboration. Every aspect of the film is determined by means of surveys and collaborative suggestions: plot outlines, twists, characters, moods. For some reason I haven't been getting their emails recently, so I'm not sure what stage the project is at right now, but I understand (well, hope) that the collaboration effort will be further applied to any business decisions that are made. I understand that the film will actually be made; it'll be fascinating to find out whether the quality of the product (the film) will be improved by the collaborative aspect.

Co-creation of products is something that consumers are actively enthused by, too. The classic example that gets trotted out as a case study wherever you look is My Starbucks Idea. Others far more qualified than me have said it all before so I won't embarrass myself, but the power of having brand advocates all giving their solicited and unsolicited opinions on what could improve the brand, is frightening really. Yet this rather unstructured qualitative data isn't a million miles from the sort of thing they could achieve by monitoring Facebook and Twitter; except that there's much less noise.

Noise is a real problem with social media monitoring - particularly as there are so many automated bots retweeting content all over the place. The day someone comes up with a comprehensive database of Twitter contacts, arranged geographically, and divided up by their area of "expertise", could be the day that monitoring gives more insight than noise. Putting together such a database is something I'm currently working on in my day job - it's time consuming but I think we've really nailed a method for saying with some confidence that on any given industry or subject area, we can pull together a pretty big list of the people that matter.

Having worked through Steve Rappaport's rather plodding book on social media monitoring (it does, however, have a lot of case studies which, most of them being American, I hadn't heard of), I'm currently getting started on Ray Poynter's new book, The Handbook of Online and Social Media Researchwhich is a hefty work and comprehensive. It's not a sexy book but I think that's the point; more a work of reference, which it seems to me to succeed at brilliantly. I'll probably throw a few words together about it when I get through it (which won't be for some time).


  1. Interesting post.

    I really like your point about co-creativity, I think this is hugely important and one of they key areas which has been facilitated and cultivated by and through social media over the years. I do think that engagement also plays an important role when it comes nurturing that input - but companies need to be smart about it.

    You mentioned noise being an issue which can affect effective social media monitoring, this is true however there are tools out there, both paid for and free, that can make cutting through this noise a whole lot easier in order to bring the most value from social media to brands.

    I look forward to reading more.

    Olivia Landolt
    Marketing and Community Manager

    @6Consulting | UK authorised Radian6 partner

  2. I read market research article, this article very informative and interesting..I refer your blog to many of my friends as well.
    Thanks for sharing knowledge..