3 Tips to Improve Your Online Influencer Analysis

Analytics

3 Tips to Improve Your Online Influencer Analysis

By Ken Burbary ⋅ March 3, 2011 ⋅ Post a comment

Filed Under chris brogan, chuck hemann, copyblogger, jason falls, klout, radian6

Identifying influencers is an important part of social marketing, and becoming big business for some companies like Klout . Yet, influencer identification and analysis isn’t well understood, or easy to optimize. That’s because no two influence analysis efforts are the same. Like fingerprints or snowflakes, an influence analysis program is unique, specific to the goals and priorities of the organization executing it.

To date, automating influence analysis has relied heavily on quantitative measures (example: like followers/friends, RTs, mentions/replies). While these metrics can be valuable inputs into determining whether or not someone is influential, they aren’t worth much at all without their counterpart, qualitative metrics. If influence is a coin, then the 2 sides are quantitative and qualitative. Focusing purely on quantitative metrics tells you if someone is popular, not influential.

A recent and fitting example of this is the news of celebrity Charlie Sheen joining Twitter. Sheen amassed 1 million followers in approximately 24 hours. An impressive feat, no? Is he popular? Definitely. Is he influential? Well, according to Klout he is. Sheen currently has a Klout score of 88 (higher than most people, including industry thought leaders like Chris Brogan (81), Jason Falls (72), and Brian Clark a.k.a Copy Blogger (73). Does this mean Sheen is more influential than any of those 3? It depends on the topic.

Enter relative influence. Relevance brings the qualitative perspective necessary to complete the influence picture. Without it, you cannot be sure that you have an actionable list of influencers. The 3 individuals mentioned above are respected, trusted voices on topics such as marketing, social media and blogging. If you were to take a simple quantitative approach, and rely on Klout’s score, you could make the argument that Sheen is the more valuable influencer to focus on. Would you take his advice on any of those topics over Brogan, Falls, Clark? Absolutely not! Relative influence should be the focus on influencer analysis. You’re looking for individuals with deep, strong ties to the community on specific topics, not individuals with weak ties to many communities. Don’t confuse popularity with influence, they aren’t the same.

In addition to using automated tools to improve the effectiveness of your quantitative influencer analysis, here are some tips to gain a more robust qualitative view and gain a balanced perspective.

Define relevancy first - Before you begin, think hard about the topics, types of conversation or actions that are relevant to your business goals. Create a relevancy list to act as a compass, guiding your analysis away from anything that doesn’t meet those criteria. You want relative influence, not popularity (Example: Nike recently targeted absolute influence, not relevancy, including David Armano in a basketball outreach program, details here). Also, think critically about what qualitative metrics are relevant to you. Not all may matter, especially if you’re focusing on a single channel. As an example, if targetting Twitter, then exclude Facebook and/or LinkedIn data from your quantitative analysis.

Use Influencer Tools are a starting point – Don’t throw out the baby with the bathwater. Tools like Klout may not be where we want them to be, yet. However, they can be used as a starting point to help you find the general direction to move forward. A more comprehensive list of influencer analysis tools, can be found here at Jason Falls blog.

Integrate with other data to gain clarity – Social data is a great start but represents only a small slice of the digital data spectrum. Combine social data with other digital inputs to validate your influencer analysis. Two good types of data are search and web analytics. Pull web analytics data for the website/blog of a suspected influencer. Are there many inbound links? Traffic? Search volume? etc… The point is, you should find a corresponding level of activity in these other data types to support the social data findings and make the strongest possible case for them as an influencer.

Finally, if you’re interested in a rich, in-depth perspective on influencer analysis, take the time to read this eBook, authored by @chuckhemann and @radian6. It will enlighten you on the finer points of influencer analysis, including specific methods and metrics to consider in yours.

Radian6_JAN2011_Book

http://chuckhemann.com/ Chuck Hemann

Ken – nnFirst of all, thanks for the shoutout. I think that eBook really just scratched the surface. Truth be told, we could write an entire book on the subject… In fact, many people have about influence. Some of the folks who look at Klout as a great solution ought to check out some of those works (Cialdini and Rosen to name two). nnAnyway, on to the actual subject of your post…I think it’s right to shift the discussion from relative influence to just plain influence. The issue with just influence is that it lacks context. To your example, is Charlie Sheen more influential than Chris Brogan? Perhaps. We have no idea looking at just the numbers who is influential. Here are two points that I might take issue with from your post: nn1. Using tools like Klout as a starting point – There is probably SOME utility there, but I think you can do enough baseline research with Google, Compete, Yahoo! Site Explorer and a dream (ok, the latter being a joke, obviously). If you build a decent enough boolean query you’ll start to pull in a bunch of outlets that COULD be influential. These I would categorize as “more research needed” sites. To start that filtering process you can easily use compete.com and yahoo! site explorer (inbound links) to do that. n2. The difference between influence and popularity – This is such a tough nut to crack. I actually dont think they are that different. However, it’s very possible that popularity feeds influence. I dunno…this might be one of those in person debates. I’ve had it with matt ridings (@techguerilla) before and he’s persuaded me enough that they aren’t that different. Seems like we need an in-person influence summit. nnAnyway, great post. This is exactly the kind of discussion we need to be having. Not “Is Klout a Good Tool to Measure Influence?” kind of posts.
http://www.kenburbary.com Ken Burbary

Chuck – Thanks for the comment, always appreciate and respect your opinion, especially on this topic.nnI’ll elaborate on the point about using tools as a starting point. Sure, you can use those others that you mentioned as an alternative. That said, my hope and opinion on the pure-play influence tools right now is that they can provide that baseline more efficiently through their automation than if we manually build it through many other tools. We’re probably splitting hairs on this one. I recognize both ways can do it, one faster, one more accurate. Either can help get you started though, no?nn2) We definitely need to have that influence vs popularity discussion. I’m not sold yet so let’s drag @techguerilla into this! nn
Donbar2001

Hi Ken,nnBuilding on your good post and Chuck’s comment, here are a couple of quick points:nn- I disagree with Brian Solis’ definition of influence in that I believe influence can occur without an associated action. I might define influence as the ability or process of causing, changing or reinforcing attitudes, opinions, beliefs and behaviors. If one changes how another feels about a particular issue, you have influenced them without an action (i.e. body part moving).nn- Influence absolutely is contextual. I might have a little bit of influence in my field – no one cares about my opinion on carsnn- Popularity is a different concept than influence. Popularity can certainly be a platform that may be used to influence. nPeople were fascinated by the train wreck that is Charlie Sheen, but does he have any real influence? (Well, maybe a bad example, I think he might have influenced a few people to stay off drugs and take their meds). Popularity more often leads to entertainment than it does true influence. nnCheers, Don B @Donbart
http://www.techguerilla.com/ Matt Ridings – Techguerilla

Well, consider myself dragged in I suppose nnAlthough I don’t know that either of you will necessarily like my response. I don’t remember the exact conversation with Chuck, but my guess is that it was the one where we were discussing *defining* influence. I have a very literal, exacting view of influence. Anything that causes you to think or do something. That’s it. Period. So in the context of a simple question like “is charlie sheen influential”, well hell yeah he is. Look at all the people taking actions and thinking thoughts based upon what he’s done or said. In that context there’s not a lot of disparity between influence and popularity.nnHowever, and it’s a big one, that does not make them close to being the same thing. A insane dictator of a country is incredibly influential, yet he certainly would not be considered popular.nnInfluence is such a nuanced beast when you bring it into the marketing context. People try and apply their own relative definitions of influence to items they shouldn’t. Context, relevance, type of influence (and there are a lot), etc. are all critical when crafting a influence program. Two things I would say about a lot of these articles. One, the front end of Klout where you see that nice shiny klout score is simply a gaming mechanism to keep people engaged. It’s the *back end* of Klout where the data is and where the real magic lies. You can buy some pretty nifty stuff back there if you’ve got the cash. Two, some of the best influencer campaigns being run at the moment in my opinion are highly complex, successful, and the ‘influencers’ are people with relatively low follower counts and folks you’d never expect.nnSo the consultant answer about influence should always be, “it depends” nnCheers,nn-Matt
http://www.dayngrzone.com/ Dayngr

Great tips Ken and thanks for including our eBook with Chuck. I liked Matt’s brief yet all encompassing definition of influence “anything that causes you to think or do something.” Short and sweet yet, gives one pause for thought.nnTrish (@Dayngr)nCommunity Manager | Radian6
http://www.kenburbary.com Ken Burbary

Thanks for the comment Trish. Simple definitions like that make it easy, but may not always work (depending on the circumstances), don’t you think?
http://www.kenburbary.com Ken Burbary

Matt,nn”It depends” is always the consultant answer (that much they burned into my brain). nnMore seriously, I do agree with your answer. One of the things we’ve been doing for clients now is to focus the influence analysis on the “mass influencers”, which I think meets your definition above (see Forresters Influence Pryamid – http://blogs.forrester.com/interactive_marketing/2010/02/my-first-forrester-report-tapping-the-entire-online-peer-influence-pyramid.html). They are responsible for ~80% of the influence impressions and posts about products and services in social channels. Despite this, they have relatively lower following/follower counts, especially compared to the a-list/celebrity social personalities that get most of the attention. nnWhile not literally matching the Forrester definition, it was a starting point for some of our mass influencer thoughts. Does this align with the type of influencer you were thinking of? If not, what are the differences? I’m keen to understand how our thinking may align on this.
http://twitter.com/JGoldsborough JGoldsborough

Ah, love a good influence debate. I think the way we define influence — which Matt and Don brought up — is integral to understanding how we view the topic. As it pertains to brands, I see influence as someone who causes people to do something the brand wants them to do. IMO, that last part is the key. Universal influence doesn’t exist. You are influential within a certain context. Charlie Sheen is not influential for any brand I know except for maybe the TV stations airing his interviews because they want people to watch those interviews. nnI think Chuck may be right — if I read your comment right, Chuck — that popularity may not be far from influence in that it is also contextual. As popular as Oprah or Bieber are, they are not popular to all people. And I’d argue that Sheen isn’t really popular, at least in the way I think of it because that word has a positive connotation to me. Nobody is watching and following him because we admire or like him. People are watching him because he’s a train wreck they enjoy making fun of. I wouldn’t classify that as popular as much as I’d use a word like relevant.nnGreat discussion, Ken. Thanks for starting it. Best, JG.
http://www.convinceandconvert.com jaybaer

Winning
http://blog.ecairn.com Laurent Pfertzel

Kenn”Does this mean Sheen is more influential than any of those 3? It depends on the topic….” – I really like this. And I think it summarizes the issue the industry has with influence. Clearly, someone influencers somebody else about something. Our brain has a relevance cog that filters out what it doesn’t need. nFinding influencers is hard and can’t be 100% automated. To find relevant ones, keywords (topical signals) aren’t enough. nOur world has become a world of specialists. We are experts in one area, may be 2, not many more. As a token, social media is really a network of niche tribes (1000s of them) where influencers occupy a strategic position. Influencers network with eachother and their community a lot. I see people forgetting about the concept of the Network, though it’s the big difference between the world seen by Google (pages and keywords) and Facebook-Twitter-like (people and network). nSo to find influencers, it’s key to mix topical and network signals.nThat’s what we do at eCairn (a little plug here).nLaurent nnn
http://www.techguerilla.com/ Matt Ridings – Techguerilla

It aligns with one type of influencer. Although there are a lot of what I would call “influential communities” that few people are taking advantage of. That’s where most of my personal focus is. It’s more about influence development via relationships to evolve influential *advocates* vs. simply getting my message into the hands of an influencer. It takes time, it’s extremely analysis heavy on the front end to define your target pools, and more of a continuous effort vs. a campaign style mentality.nnIn other words, I want to reach that soccer mom with 500 friends who look up to her. And not just any 500 friends, but 500 of the most secondary influential friends. I want to drop my stone into that core group and let the ripples move outwards, not inwards. Her opinion on her awesome minivan carries a great deal of weight, her friends who are now influenced heavily by her frequent mention of her awesome minivan, tend to automatically recommend that model when someone else mentions looking for a new car, those purchases build new advocates which reinforces the original soccer mom’s expertise on choosing an appropriate vehicle for that lifestyle, etc, etc, etc. Finding, and then ‘influencing the influencer’ takes a great deal of effort, reinforcing that influence takes time, and ultimately the ‘trusted advocate’ you help build is what you want. It’s certainly a more impactful way of moving the needle than getting someone with a million followers to send out a single tweet, no matter how ‘influential’ they are. End of the day, it’s still about people and relationships, and it still takes time, as cliche’ as that may be.
http://www.kenburbary.com Ken Burbary

Laurent – Thanks for the comment, I appreciate you stopping by and adding to the conversation. Would like to learn more about how eCairn does just what you described (adding topical relevance). Let’s take this offline, ok?
http://www.kenburbary.com Ken Burbary

Great point JG. Which leads to another question. If popularity has a positive connotation, what word do we use to describe the negative (Sheen scenario). Train wreck is great but isn’t quite it.
http://www.kenburbary.com Ken Burbary

Thanks for providing the detailed response Matt. I don’t disagree with the cliche. There are no shortcuts when it comes to community and influencers.nnCurious about the up-front identification and analysis. This can be a painstakingly manual process, at least in my experience. Which methodologies/tools of the current influence analysis landscape do you find useful, if any?
http://www.techguerilla.com/ Matt Ridings – Techguerilla

The initial collection stages are done via well defined (and then continually tweaked) filter profiles in tools like Radian6. The secondary stages of refining that pool are almost completely manual. A high level view of the process looks very much like this: http://www.brasstackthinking.com/2010/05/social-hunter-gatherer-programs/

M	T	W	T	F	S	S
« Feb				Apr »
	1	2	3	4	5	6
7	8	9	10	11	12	13
14	15	16	17	18	19	20
21	22	23	24	25	26	27
28	29	30	31

Search

Analytics

3 Tips to Improve Your Online Influencer Analysis

Calendar

Archives

Twitter Updates

Find me on the web

Also Find Me At:

Analytics