The concept of influence in social media has gained relevance lately, particularly as marketing budgets and profit margins shrink. Social media has been touted as one of various strategies to successfully engage with potential customers and consumers. Since undertaking social media strategies wasn’t my day job (it isn’t yet, but I’m doing more applied work now), I didn’t pay a lot of attention to whether I was being influential in social media or being seen as “an influencer“. I will completely disclose that I am human, and sometimes I play along with the popularity contests along the lines of “measure your Twinfluence” or “you’re in the Vancouver Twitter Elite“. Heck, I sometimes even worry about how long I’m going to be able to hold on to my #1 ranking on UrbanSpoon. Ok, end of disclosures. My point is, I don’t claim to be a social media expert nor an influencer.

My perception is (and I could be proven wrong, but this is what my intuition tells me) that influence in social media is a very complex concept and there are very few reliable, quantitative metrics to measure it. As someone who has been trained to have a researcher mind, I find this complexity rather fascinating, but awfully disturbing. Given how much praise has been given to transparency in social media, and how important it is to be fully open and disclose everything in this realm, it comes to me as a bit of a surprise that methodologies to rank individual’s influence (be it on Twitter, on Facebook, across the country, etc.) are quite subjective.

I recently came across the website of Web Analytics Demystified and their analysis of Twitter Influence. When I first saw the site (hat tips to one of my followers who tweeted about my position on the rankings – at the time he informed me, around 4.30pm, I was ranked # 88 and only one of three Vancouverites on the top 100). By the time 1.45 am of the next day came, I was already # 91. My estimation is that this tool to calculate Twitter Influence will be rather influenced (pardon the cacophony) by individual followers and the early adoption of the tool to calculate your favorite social media degree of influence. That is – the more people use it and calculate their own Twitter Influence, the more the results will change. Now, will those additions affect the end result? Maybe. It could very well happen that, as more people know about the WAD TwitCal tool, more people will follow those who are in the top 10, 50 or 100 Twitter influencers in the world.

When I saw the results and came to the conclusion that WAD did a quantitative analysis of a variety of elements to calculate a particular score, quite frankly was somewhat taken aback. I highly doubt that it is possible for me to compare myself in degree of influence to people who actually DO social media for a living.

As a way of an example, several of my friends were nominated, ended up being semi-finalists and finalists in the Profection list of Canada’s most influential people in social media, and four friends whose work I know and admire ended up in the top 10 (Kate Trgovac, Kris Krug, Darren Barefoot and Rebecca Bollwitt).

These friends of mine certainly influence ME, and I am sure they influence a whole lot of other Canadian social media people. And their ranking in the Canadian influencers list came from a quantitative analysis. There are a number of challenges to using quantitative, crowd-sourced data as proxy for influence in social media. I’ll outline off the top of my head a few below.

1.- What happens if the person does NOT do social media for a living? Does he/she have high influence? I read a lot about it and write a lot of blog posts, and I use Twitter and numerous Web 2.0 tools. On a good day, I’m in the top 10 Twitter Elite. Does that make me a social media influencer? Not sure. Admittedly, I am capable of analyzing social media strategies and well, recently I did just that with my talk at Northern Voice 2009 on “How to Use Social Media Strategies to Increase Environmental Awareness”. I argue that influence in social media is more of a “reputation” game than a “numbers” game.

2.- What happens if the figures are crowdsourced? They become skewed towards who has more followership. That was my primary criticism of the calculation of Canada’s most influential figures (and I told this to David Forde from Profection). When influence figures are crowdsourced, the popularity contest begins. WHO you know and HOW MANY you know and how many KNOW YOU are all relevant variables.

The quote below comes from the Profectio site on the methodology.

# Voting was then done by the community for here and here and a score of 0 – 3 was given.
# Postrank rang a report on both list to determine their engagement, and a score of 0 – 3 was given. [Profectio]

The question that remained in my head was WHO is the community? I saw people whom I absolutely deem very influential in social media completely absent from the list. They probably weren’t nominated. AND people who are very influential didn’t end up in the top 10. The problem is always – the rankings game. It is, indeed, a zero-sum game. For every influencer who lands in the top 10, one influencer (legitimately influencer) is left out.

3.- Are we using the right metrics of engagement as a measure of influence? According to Dave Forde from Profectio, they used PostRank, which hypothetically measures engagement by virtue of analyzing how engaged you are in the Web 2.0 world. Truth is, I argue that nobody can really measure engagement in an objective way. For example, according to PostRank you are more engaged as a blogger if you write a blog post in reply to other blogger’s post. But what if I don’t want to link to that blogger’s post because I may have higher traffic and I don’t want to drive traffic to said blogger’s site? This is quite a real possibility.

I have previously written asking the question of whether we are losing the bi-directionality that is supposed to characterize the Web 2.0 landscape. I fear that it’s happening at least on Twitter. In fact, my thoughts on bi-directionality resonated a lot with what Ian Capstick presented at Northern Voice 09 arguing that Obama hadn’t really won thanks to social media (because well, Obama wasn’t bi-directional on Twitter). Heck, Canadian politicians really didn’t do much in terms of bi-directionality! But I digress… back to influence.

At the end of the day, I have to agree with what Nate Elliot said at his Northern Voice ’09 talk – there is a shift in the way influencers behave (Nate calls someone who pro-actively seeks to influence people a ‘new influencer‘). In his slides (which he generously shared with me) he indicates that a new influencer proactively seeks to influence people by providing advice. Nate kindly mentioned me during his NV ’09 talk making reference to me as a trusted advisor in regards to food (thanks Nate! I hope you enjoy the restaurant reviews!).

One point where I don’t agree with Nate (and we’ll have to discuss this in person) is that new influencers “broadcast” their recommendations. I think that, (assuming that people do perceive me as an influencer), the way in which I influence people is through engagement, beyond broadcasting. One of the reasons why I get unfollowed frequently is because I am way too chatty on Twitter. But I find that if I don’t engage with my readers and my followers, the value of the conversation diminishes.

So, am I an influencer in the social media realm? Maybe, maybe not. But one thing is for sure, I certainly am proud of what I’ve accomplished in the blogging world. I would like to be seen as an influencer in the environmental field (particularly in my two primary fields of focus, water governance and environmental policy). But in academia, influence is very much a quantitative game. How many citations do your papers have in the Social Science Citation Index? Citations are usually the proxy for influence. My view is that, in social media, we are trying to measure influence in a quantitative way when it also has substantial qualitative overtones. It’s not only how many people you know (and influence) but who you know (and influence). Also, I *really* want to be perceived as an influencer in my research fields. But that’s the subject of another post.

With that in mind, what do YOU think would be good quantitative metrics of influence?

Related posts:

1. Measuring influence in social media (the Klout score)
2. Is measuring real influence in social media possible?
3. My recent Social Media Club Victoria talk: “Towards an Action-Focused Agenda for Social Change Using Social Media” (#smcvictoria) @smcvictoria

My personality traits work to my advantage when using qualitative research methods. I am pretty good at interviewing people, analyzing textual data, coding using Strauss and Corbin’s axial coding methods, undertaking ethnography, etc. I am considered a specialist in institutional analysis because, well, I know how to study rules and routines. And the majority of these studies are undertaken by observing and interviewing.

The thing is, during the course of my PhD, I became REALLY quantitative. Since my advisor’s training was quantitative, he impressed it upon me. I’m quite grateful to him for doing that because thanks to his sage advice and training, I have examined wastewater governance and policy using quantitative methods (something that is rather atypical in this body of literature). And of course, there’s my love of game theory and econometric methods. WOWSA.

So, it shouldn’t come as a surprise that I drooled (just as Greg tweeted earlier this morning) when I saw StatsCan’s RSS feeds. Given the kind of research I do, economic, government, population and environmental indicators are the RSS feeds I grabbed. You can grab any others as you may see fit.

Unfortunately, and this is quite sad, I can’t get any quantitative data on water through those RSS feeds (or at least I haven’t been able to get any so far). I do know where to find some water-related statistics in Canada, but the state of the art in regards to accurate water stats in Canada is (as mentioned in Karen Bakker’s edited book) rather appalling.

If I manage to get my hands on some good data, you’ll see some pretty graphs here sometime soon.

Related posts:

1. I need help with my RSS feed
2. Statistical improbabilities?
3. The echo chamber in water research (social sciences)