Analytics Is King

Is the discussion around/practice of online influence actually regressing? Identifying and then activating influencers is an important topic that deserves informed debate and the implementation of only the best tools. I actually think we have quite a bit of the former, and the latter is the gigantic black hole that we're all getting sucked into.

David Armano wrote an excellent post examining Klout, followed by Rebecca Dennison writing her own review of some of Klout's new features. Rebecca also wrote a great post looking at mPACT that you should check out. I'm not being critical of Rebecca or David because both of their reviews are great, and obviously very helpful to the person that is coming to this space for the first time.

Before we launch into the meat of the post, lets debunk a few influencer misconceptions

1. The process can be automated - The extent of the possibilities around automation begin and end with identifying outlets and writers. The actual number crunching/qualitative analysis should not be eliminated. In my experience, clients feel better about the final output when it has been scrubbed by a human.
2. Influencers are bloggers - I can't tell you how many times I see examples of "influencer lists" that are comprised entirely of bloggers. As if the entire social web centered around blogs? Have we forgotten YouTube? Forum users? Twitter users? Mainstream press? Your influencer list needs to be all encompassing otherwise you have an incomplete list.
3. Because someone has written negatively about my brand they shouldn't be in my list - This one causes a bit of a headache for me. You can learn just as much from the negative comments as the positive ones. In fact, probably more. Why eliminate these people?
4. I have to reach out to everyone on my list - One of the things I advocate all of the time is trying to come up with a list that includes people you might only listen too. Another segment that you'd be reaching out to directly, and still another that might be receiving a mass pitch. A mass pitch?! In social media?! The horror! It can work. Trust me.

There are others, but I don't want to get bogged down in the misconceptions. Rather, I'd prefer to tell you what tools I'd use if I were starting to build an influencer list from scratch. And no, just to get it out of the way, the process would not involve Klout, PeerIndex, mPACT or any tool like it. So where do I begin?

1. Google - No duh, right? This is where the process usually begins and ends. I'll develop some sort of query (that involves my client's name, some industry keywords and competitors) to start identifying outlets. Ideally, I'll start with at least 6 months worth of results (though I'd prefer 12) and just start clicking through. Yeah, it's a manual process. Yeah, it's time consuming. Yeah, trust me, it'll be worth it in the end. Oh, don't forget running similar searches in Google Groups, and Blogs
2. Listorious/Twitter Search - Both of these are invaluable resources for Twitter users. Check them both out.
3. YouTube/Flickr searches - Self-explanatory, right? But you'd be surprised how few influencer lists actually involve video bloggers. There are ton of them out there with broad reach
4. Compete/Quantcast/Alexa - Several people, including my friend Dave Fleet, have written posts identifying issues with compete data. It's true. However, obtaining traffic data for sites you don't own is generally problematic. All you can do is attempt to verify the data on multiple platforms.
5. Yahoo! Site Explorer - An excellent tool for looking at a site's indexed pages and inbound links. Both excellent barometers of overall reach
6. Microsoft Excel - I'm assuming you have this on your computer, right? Ok, moving on...
7. Blogrolls - Pretty invaluable in identifying other bloggers stemming from your initial Google searches. Use them and use them well.

Anyway, those are the tools I use to start building my lists. Does this take more time? Sure. Are my results likely to be more accurate/relevant to the client? You bet your ass. Don't give in to easy. Conduct your due dilligence.

We're all a mixture of different skill sets. Some of us come from an analytical background. Others come from a pure communications background. Still others come from a political science background. The fact of the matter is that we all need to be a little bit of all things. What's the ideal make-up look like?

• Sociologist - Understanding how offline societies interact often helps us determine how online societies will perform.
• Psychologist - This goes without saying, but understanding human behaviors is critical. If I put out a piece of content, how is someone going to "feel" about it? (full disclosure: my dad was a forensic psychologist. I tried to major in psychology but, well, quickly changed after the first class. :) )
• Communicator - Obviously, right? You don't need to come from the traditional PR background, but you need to have strong interpersonal communications skills to succeed in social media (really any profession). This also includes the ability to write, and write well.
• Ambassador - You aren't going to come up with a solution that makes everybody happy. Building on the wisdom of Solomon, finding a way to split the baby is often how you survive in this business.
• Hard ass - I was looking for something creative to say here, but "hard ass" was the only thing that came to mind. Listen, there is going to be someone along the way that challenges you. When that happens, you have to be a mixture of all of these things, including hard ass, to survive.

I hesitate to rank any of those in order of importance because I honestly believe it is a blend of all, but my intuition says, particularly in social media, that being a part-time sociologist and psychologist are most critical (probably followed by communicator). We have to be able to answer the "why." Why is someone talking about our brand online? Why do they have such a negative opinion about a member of our management team? Why are they turning to Facebook instead of a forum where the rest of our community lives? Said another way, we have to understand the motivations of our online community. Nowhere is understanding the underlying motivations more critical than when developing and then implementing an influencer program. This week, I will be participating in the Radian6 Social 2011 User Conference on a panel about the value of influence. We're going to be talking about everything from the metrics used to define influence, to how to utilize an influencer list to even something as fundamental as whether influencer lists are necessary. There has been a lot written on the subject lately, and I'm under no delusion that we're going to solve all of the issues associated with influencer analysis, but hopefully we can make a dent. One of the issues we'll likely be tackling is determining the influencer analysis output. What I mean by that is understanding how we want the list to be used, and what the heck do WE actually mean by influence. In a post a couple of months back, my friend Tom Webster pointed out that one of the ways influencer analysis could be improved is understanding motivations. I couldn't agree more. We have to know what influences people, why they are talking, what messages they respond to and how they want to interact with brands. If you can't answer those, your influencer analysis output is going to be flawed. You'll never completely answer the "why" using listening data, as Tom noted in his post. The only way you truly know is by asking (and yes, I know, people can lie). So, as you start working with influencers what can you do to fully understand the "why" when asking your community may not be practical at that stage? A few things:

1. Use your listening data - Yes, I know I just said you can't rely on this solely, but it can be helpful in gaining a basic understanding of why people do what they do online when referencing your brand.
2. Asking the influencers themselves - In theory, you should know through your research what types of content communities react to from these influencers. But, why (there's that word again) not ask them? You might find a totally different answer from the blogger/forum poster/mainstream press member, etc... then what your research is telling you. That's valuable intelligence.
3. Surveying the online community - No, social media listening is never going to fully replace surveys. They have value. No, they may not be a truly representative sample of your online community. Yes, there is valuable intelligence to be had from using a survey instrument.
4. Knowing your motivations - Why do you want to interact with influencers? Is it to increase awareness in the form of share of voice? Is it to improve sentiment about your brand? Is it to build long-term brand value. Not only should we understand the motivations of our community, but we need to understand why we're doing what we're doing. Only then can you truly grasp the entire landscape.

We have to be one part psychologist, one part sociologist, one part communicator, one part ambassador and another part hard ass to succeed in this social media world. Whether that's participating in social media more broadly, or developing an influencer identification program answering the "why" is critical. We're not going to solve this dilemma during our panel at the Radian6 conference, but likely through more informed conversation online and off. Think back to the programs you've built using social technologies... Have you sufficiently answered the why internally and the why within your community? If not, it isn't too late to course correct depending on your findings.

The world of social media never ceases to amaze me. Just when you think you have things figured out, you come across something that is totally mind blowing. Like, for example, announcing a Christmas party by issuing a release that indicates you are throwing a "first-of-its-kind party" with people specifically chosen by their "social media influence and engagement." Listen, I couldn't care less if someone throws a Christmas party. In fact, I think they should. Being around people during the holidays is important. I'm hoping that we can all agree announcing your Christmas party via release is pretty over the top ridiculous, right? Yikes. I'm not going to spend a lot of time talking about the event, but I have to get a couple of things out of the way. A large network on Facebook means you are influential? Listen, any moron with a dream can create a Facebook page in about 30 seconds. There are over 500,000,000 people on Facebook. If you create a page that's mildly interesting you'll gather fans. If you don't have a fan page how does having a large Facebook presence mean anything other than you have pretty loose standards when accepting Facebook friend requests? Or, better yet, if you've created a Facebook fan page for yourself how ego-maniacal are you? Seriously... get a hobby. Yes, there are legitimate reasons to create a Facebook fan page. A fan page dedicated to your own ego aint one... Twitter and Facebook represent influence on the social web? This is one of those things that drives me nuts over almost anything else. THERE IS MORE TO THE INTERNET THAN FACEBOOK AND TWITTER! Because you have 5,000 followers on Twitter DOESN'T mean you are influential on the Web. It means you could be influential on Twitter! If you really wanted to have an event of the "social media A-listers" how about trying to define influence from participation across a wide variety of channels. Perhaps that's the way it's actually happening and the release is too narrowly written, but it sure doesn't seem that way. Anyway, if you take those two things away, my larger issue is with how these people are being selected. I think my record is pretty clear about off-the-shelf influencer algorithms. Generally speaking, I think they suck (and not just a little). Where do they go wrong?

1. There are so many intricacies of influencer identification that a computer (no matter how smart you make it) can decipher that I'm generally distrusting of the output. Maybe there are some smart humans on the back end of these tools, but how much do they know the brand/topic area? About as much as your dog Fred.
2. The tools that are available typically only measure one channel -- usually Twitter. We have this love affair with defining influence on Twitter. Why? It's a big channel, but it's only one. What does 5,000 followers really get me? The CHANCE to reach 5,000 people? Excellent. Guess what? The first person to talk about "opportunities to view" on Twitter is getting a knuckle sandwich. We don't need the issues of traditional PR measurement seeping into social media, thank you.
3. Similarly, platforms like Klout tend to overvalue the volume of posting. Yes, someone that's relatively inactive on a particular channel would likely not be considered an influencer, however there have been enough analyses by people showing how important frequency is to your Klout score. I'd much rather contribute a smaller amount of content and have it be useful to people then just be a chatter box on Twitter. Check out my man Kasey Skala's post from Monday about the inflated importance of volume on influence if you don't believe me. For the record, I probably post too much on Twitter but it isn't because I'm hoping my Klout score goes up.
4. These are all big concerns, but the biggest has to be relevancy. Now, that's less of an issue here because the folks putting on this event are only looking for people like them, so obviously they are highly relevant. What I worry about, though, is that some unsuspecting brand gets a hold of that release and tries to identify influential people through services like Klout. Instead of taking a list of people who have a lot of followers and dropping them into a service like Klout, how about spending some time with a cup of coffee, a search engine and a list of key terms and try to identify some people who are RELEVANT! I'd never say reach doesn't matter. However, it takes two to tango and reach and relevance go hand-in-hand.

By the way, this isn't a reflection on Klout. Do I think they are one of the better off-the-shelf solutions? Yes. Would I utilize them to define influence in the current form? No. Anyway, I'm hoping these folks have a great time at their "over-the-top" holiday bash. I hope they have a great time in New York City. However, the event only serves to confirm two thing: If these people are influential, it's likely only influential to each other; and their "score" is based on a flawed system of popularity and NOT influence.

I’ve been amazed watching the sheer volume of conversation around the Fast Company Influence Project. Can’t we all agree once and for all that it’s a really poor representation of influence online? I mean, spreading a unique URL in the hopes that someone clicks on it, signs up and starts the project themselves is an appropriate metric for influence? In the words of my friend, David Mullen, “hogwash!” Amber Naslund wrote a truly excellent post (please check out Brass Tack Thinking if you haven’t already) in response to the project that received a lot of fantastic comments and sparked a question in my own mind: Are we making the definition of online influence harder than it needs to be? I tend to think the answer is an emphatic yes! I’ll agree that it’s very difficult to make sweeping generalizations on who (or what) is influential across verticals. I’ll also agree that it’s difficult to make assumptions on influence given the diversity of goals that every brand has online. However, I think what we’re talking about can be boiled into a relatively simple equation: Individual Influence + Outlet Influence = Overall Influence. Let’s hit each of those individually… Oh, one quick note – I’m only providing metrics for demonstration purposes. These shouldn’t be used in your campaign unless you feel that they align with your program’s goals. Please don’t cut and paste otherwise I’ll hunt you down :-). Anyway…

1. Individual Influence – Each brand is going to have a different definition for how influential a person within their micro community is on other members of that community. What we’re really wanting to know is how relevant is their content (related to a particular), how many people are they (or could they) reach, how much content are they producing (without spamming us) and how often their content is picked up by others.  There are metrics you can obtain pretty easily that underlies each of these categories, but I think it makes sense for you to start at that level and match metrics to each category that also align with your goals.
2. Outlet Influence – Again, every brand is going to place varying weights of importance on the different channels. Some will place more emphasis on blogs, while others will think forums are most important. Ideally, to define influence within your online community, you’re taking into account a variety of channels, but understand that most aren’t going to go to that level. Whether you’re looking at inbound links for a blog, or total visits for a mainstream news outlet, or total authors on a forum, your goal is to quantify (ideally you’re looking at a bunch of metrics) how strong that Web site's strength

The trick, in the end, is bringing those two areas together to provide a meaningful overall metric or score. I know there are a bunch of metrics underneath each of those categories, and I know I’m not giving them to you, but I think this easy categorization can help you start to get your head around influence online. Happy to try and answer specific metrics questions underlying this in the comments.