FEATURE1 January 2011

Making the research connection

sHow are market researchers using social media? Francesco D’Orazio, Sharmila Subramanian and Jess Owen of Face’s social media team find out.

Res_4004444_face_social_media_wordcloud

There’s no shortage of impressive stats on social media usage. Social networking now accounts for about a quarter of time spent online in the US, according to Nielsen. That makes it the single biggest category of online activity – bigger than email, gaming and instant messaging put together. If we see market research as a route to greater consumer understanding, it is going to become increasingly important that we as researchers develop an understanding of the social media environment in which consumers are increasingly immersing themselves.

In August, Face began the Social Media in Research project to explore how market researchers are using social media. This project is a first step in looking at where the research industry has got to.

The project began by analysing 345,819 tweets posted by 3,500 researchers on Twitter over a one-month period using Face’s social media research tool, Pulsar. The project also reached out to the research industry online to take part in a short survey. 530 people from the industry answered the survey and while no specific quotas were set for the sample, we closely monitored respondent demographics to ensure a spread of ages, geographical location and gender. The intention of the project is not to carry out a census of the market research industry but to highlight some emerging insights into how the industry currently defines the role of social media.

Here’s what we found (and there’s more detail over on the Face website).

1. Social media is still not used much to answer research briefs
Only 5% of those surveyed are currently using social media monitoring tools to answer research briefs – and that’s from a sample recruited mainly through social media. This low level of adoption implies that social media as a source of consumer insight is still in its infancy.

2. Social media is mostly used as a business intelligence tool
The primary use of social media for researchers is cited as professional networking and content sharing, mostly blog posts from people in the industry. This suggests that social media is viewed as a business intelligence tool rather than a consumer insight tool.

3. Social media is viewed as being distinct from other platforms
For the vast majority, social media is defined by platforms. Names of specific platforms, notably Facebook and Twitter, were among the top topics discussed, and were also the top words that emerged when we asked people what came to mind when they thought of social media.

4. Researchers still aren’t immersed in social media
As an industry we’re not yet living and breathing social media. Forty per cent of our survey participants indicated they spend less than one hour per day using social media, which suggests that social media is still viewed by many as just a tool rather than as an immersive environment. This makes it difficult to embed it as an integral part of the research process.

5. US-based researchers are leading the conversation
Looking at the most influential market researchers on Twitter, it’s clear that the US is still leading the conversation, as home to the majority of the top influencers. However, the engagement levels of the most influential researchers (based on the amount of interaction they have with others) are still quite far behind the engagement levels of influencers in other industries, such as advertising.

Patterns of engagement
We studied communication between pairs of users on Twitter to understand who is engaging with who. Tweets are generally viewable by anyone, but you can also communicate with specific users by including their user ID. The conversation is still driven by the US, with @TomHCAnderson (of Anderson Analytics) the single most ‘engaged’ person. But the US doesn’t simply set the agenda – it seems to be acting as a conversation hub and multiplier, a key means of connection between researchers and companies in other countries such as Canada, Belgium and Norway.

Most of the major influencers seem to be highly interconnected, which is not unusual in small industries. In this case the relatively low degree of separation is probably due more to the limited uptake of social media by researchers. This keeps the conversation space quite small, but at the same time facilitates connections between different branches of the industry.

Talking points
Looking at the various topics of discussion by volume, and using semantic analysis tools like OpenAmplify, we were able to identify the most lively areas of discussion.

The content of the conversations between researchers on Twitter is highly self-referential in that researchers talk a lot about social media itself. ‘Social media’, ‘Twitter’, ‘social media marketing’ and ‘Facebook’ were all major topics. This suggests that many people in the industry are still going through a learning phase when it comes to using social media.

Sharing and accessing content is the main reason researchers use Twitter, exemplified by the significant uptake among researchers of content-filtering tools such as the popular Paper.li (which aggregates daliy Twitter activity related to particular users or topics in the style of a newspaper).

Discussions about blogs appeared prominently too, showing that there’s still a place for longer-form writing in the industry as well as status updates and tweets, which are often used to promote more detailed content elsewhere.

What are the key benefits of using social media in research?
Open text comments on the benefits of using social media for research revealed a clear perception – social media offers a window into naturally occurring behaviours and in this sense is seen as an invaluable observation tool to get an understanding of consumers (or ‘people’, which we were pleased to see cropped up more often than ‘consumers’). We also see the words ‘trends’, ‘information’, ‘quick’ and ‘fast’ appearing a lot. Having quick access to valuable information is still seen as the key benefit of using social media in research. Social media seems to be all about keeping up to date with fast moving trends, shape-shifting segments and real-time flows of content. Beyond gathering and sharing content, social media is seen as a useful tool to keep in touch with clients, or seek out new ones. In this respect many see social media as something they have to be seen to be involved with, or at least show an understanding of in order to impress clients.

When you think of social media, what are the first words that come to mind?
The above wordcloud (click to embiggen), based on the words that respondents said came to mind when they thought of social media, shows that the mainstream social media platforms, rather than the whole web environment, still dominate the conversation. There seems to be a feeling in the industry that social media is about the big mainstream platforms of Twitter and Facebook and less about blogs or forums, which hints at quite a recent uptake. There’s also a focus on professional networking in the industry. ‘Friends’ is smaller than ‘connections’ and only just bigger than ‘LinkedIn’ and ‘networking’, suggesting that social media for many researchers is a business intelligence tool and something they may have come to later in life for professional reasons, rather than as part of their own personal social life. ‘Communities’ makes an appearance, presumably referring to online communities as a research method. However, it’s interesting to see that there is almost no mention of social media monitoring, web/behavioural analytics or netnography.

Still early days
The findings indicate that the research industry is still only at the start of the journey when it comes to the use of social media. For an industry that is focused on working out what lies ahead, it appears we may be behind the curve in this regard. Face is planning to carry out the Social Media in Research study on an annual basis, so it will be interesting to see if this continues to be the case over the next few years.


The benefits of social media in research

?A selection of verbatim comments on the benefits of using social media in research

Finger on the pulse
“An immediately available feel for what’s going on out there; it allows me to gain a quick perspective on all sorts of different targets and markets”

Targeting youth
“Connecting with younger audiences, who are less responsive to traditional research methods”

Industry networking
“As a freelance researcher I find it a great way to connect with other researchers. [Twitter] is like a surrogate office for me”

Getting stuck in…
“Need to be aware of the tools people are using in their lives, understand options for communicating with people, learning is extremely important – constantly learning about everything”

…or not
“We try and stay on top of the social media trends so we understand what new things our consumers are up to (e.g. Foursquare) but I don’t necessarily sign up to them myself”

Client demand
“Clients want to know about it so I make sure I know about it”

Lingering scepticism
“Unconvinced as of yet, appears to be just a clique of people talking to each other”

Every man for himself
“I am very interested in getting work through contacts and spreading malicious rumours”

Head over to Face’s website for more from the Social Media in Research study

6 Comments

13 years ago

Really interesting article, thanks Francesco. It seems like a lot of non-research agencies are getting involved in social media research – PR, creative, media planning, web / social media – and if researchers don’t catch up they could end up losing out!

Like Report

13 years ago

The problem is worse than this. The limited social media research that is occurring is focussed on really basic volumetric activities, such buzz tracking, word frequency counts or automated sentiment analysis. These things barely brush the surface. Social Media monitoring in its current, automated form is like standing next to a road and monitoring the traffic. You can record the noise of the engines and get a sense of how busy that road is at different times of the day. You can write down the colours, and makes, and speeds and numbers of the cars. This will give you a whole raft of data, which you can cut in lots of different ways. But unless you are a traffic policeman, its pretty much completely useless. To know anything useful about the traffic, anything that you can affect, or influence, you need some context. You need to see where the road is on the map - where the cars are coming from, and going to. You need to know who’s in them, why they’re travelling, why did they choose the car over the train, etc etc. In social media terms, that would be a full social media audit of audience, influencers and themes... and an analysis of implications, opportunities and challenges. That's real research. That's what we, as an industry, are good at . Sure its harder than just counting cars. But its worth it.

Like Report

13 years ago

Thanks Aya, you're absolutely right, quite a lot of (fairly basic) research on social media is being carried out by non-research specialists. Not surprisingly most of the tools available on the market are very poor in terms of research methodology, especially when it comes to semantic analysis, influence assessment and content "weighting". But the most worrying thing for me is the fact that a big part of the research industry doesn't live and breath social media (and a substantial number of people in the industry is still debating whether this is a fad or not...) @Alistair, thanks for commenting here, I tend to disagree on a number of levels: in my experience and from research carried in the field, the few companies in the mrx industry who are switched on about social media research are much better equipped to do nethnography than social media analytics. (and the fact that in our survey only 5% of the respondents uses any social media monitoring tool is quite telling). On your volumes vs quality point, we see social media research as a phased approach where the mere quantitative analysis is only part of an iterative layered process where software analysis, human content analysis, statistics and ethnography complement each other. Whatever research you're doing in social media, there's no way that a purely quantitative (volumetric) assessment is going to take you anywhere. Having said that, luckily enough, semantic analysis, although still massively flawed, allows us to gather slightly more complex insights than the colors of cars, for example hints to the reasons why people are driving somewhere. In "qualitative" research there is a historical prejudice towards more quantitative methodologies and this is not helping the industry to fully grasp social media, where qualitative and quantitative patterns are completely intertwined. More generally I think we're heading towards a new model research where the neat distinction between qual / quant will become irrelevant and damaging. Think about the variety of dimensions of human behavior that we are increasingly capable of tracking, or simulating and think of the wealth of social data increasingly available through all sorts of API... Not integrating this richness into our processes and methodologies would be a missed opportunity and a big mistake.

Like Report

13 years ago

Thanks, Francesco, Sharmila and Jess! This is great information - and @Francesco I agree "the neat distinction between qual / quant will become irrelevant" Social, mobile, augmented reality and more offer researchers a wealth of insights that would be missed if forced into the predefined mold.

Like Report

13 years ago

Hey Francesco. Good thorough reply, thanks. Er... either I'm missing something, or we're violently agreeing on most points... Yes - MR companies need to get a better handle on social media, living it, breathing it, using it for research Yes - if they do, they'll be better at it than purely automated SM analytics. That's exactly my point Yes - when you use any tool as part of a multi-staged approach, then the whole is greater than the sum of the parts. Approaching every problem in exactly the same single way is a recipe for failure. and YES - qual and quant don't really help as distinctions in the social media world. We're on the same side, brother! Where we might disagree is the role that humans have to play in this. I firmly believe that only human minds have sufficient sophistication, subtlety and comprehension to extract meaning from data and assign some parts of that information an 'importance' score. And the smarter, more experienced and better trained the human mind, the more effective the extraction of meaning. That's good news for researchers. We should be at the forefront of the social media revolution. Let's go!

Like Report

13 years ago

@Alistair Given the volumes of content, I think we have to automate the first level of analysis of 'meaning' and 'importance' because it's simply impossible to apply human analysis to every message. For example, sentiment analysis gets it right maybe 80% of the time, so that's a first guide to emotional tone - and we we can write algorithms to say that news articles tend to have greater reach than tweets, or blog posts with lots of comments are more influential than those with fewer. "Mechanical" analysis of this sort also has the advantage of being entirely consistent in method over time - rather than influenced, say, by the researcher being in an especially good mood and consequently analysing everything in a rose-tinted light that day. Algorithmic ranking of key topics or influencers is certainly imperfect, but it also offers a valuable element of surprise or disruption - foregrounding everything that's numerically important even if the individual researcher might instinctually ignore half of it. It forces us to confront our biases and make more rigorous analytic decisions about what is most important. But then of course this needs to be refined with human coding to handle irony, complex sentiment, the outlying cases where a tweet goes viral or a news article isn't actually in an important source... And to put all this data into the wider social & cultural context, and what we know about the brand in question and how it's developed over time. Qual + quant, human + algorithm, each working with its (supposed) opposite to build & refine a richer understanding. Social media analysis as Hegelian dialectic?

Like Report