Five reasons why measuring influence is elusive

Rebecca Denison is an outstanding, bright young professional and one of my favorite Twitter friends. She also happens to take exception with some of my musings on Klout and influence.  You know … I STILL think she’s bright! In the spirit of balanced debate I encouraged her to write a guest post.  Here we go …

Recently Mark wrote about the ins and outs of Klout, complete with some great insights from one of the co-founders.  It was a great look at the score and all of the work that goes into it.

The factors of influence were discussed, and we learned that Klout now accounts for 50 of these so-called factors. While I respect what Klout is trying to do, I must vehemently disagree that their score measures influence.

Klout measures tweets, retweets, popularity and the like. Klout measures numbers. True influence is much more than numbers. And true influence is felt both online and offline, and an automated digital algorithm will never be able to measure this accurately (at least not with data available today).

I’ve included what I believe to be just five factors of true influence below. What else would you include? This list will likely never be exhaustive.

Factors of True Influence

1. Actions Inspired: A common way to understand influence is to consider actions inspired. There is something to be said for changing minds, but I’ve always felt true influence leads to action. For example, when visiting a new brunch place in my neighborhood, I tweeted asking whether I should order the banana toffee French toast or eggs Benedict. A friend I often look to for restaurant recommendations in Chicago suggested the French toast, so guess what I ordered? That’s influence.

Think about measuring this. Yes, score like Klout consider retweets, which are an inspired action (I’ve influenced you to share  something), but they can never account for so many other actions because they cannot see them or track them. If you tweet and recommend French toast, and I order it, that’s incredible influence, but it would go unnoticed.

2. Context/Topic:  Influence is incredibly contextual or topical. I will never be influential to anyone about being an 18 – 24 year old male. I like to think that I’m influential to some people about things I’m knowledgeable and passionate, though. I’m the first to admit I do not have an overwhelming following (or even an impressive one), but I pull some weight in certain communities. Take the following tweet, for example:

I may not be an expert on rocket science or SEO or sheep, but when it comes to UNC (and UNC basketball), Nick finds me influential.

I know, I know. Algorithms attempt to measure this. But would Klout know that the above tweet indicates I know a lot about UNC basketball or just UNC? Klout cannot read the context and understand that McCants was an incredible player back when I was a senior in high school just waiting for my turn to go to Chapel Hill.

3. Situation/Location:  As much as context can be important, situation and location can be just as critical. If you ask me for recommendations for brunch locations (I LOVE brunch), but you live in NYC, unfortunately you’re out of luck. I don’t have much experience with brunch in NYC. Ask me about brunch in Chicago, and specifically in Lakeview or Lincoln Park, and I’m practically an expert.

Measuring this is not always easy since we don’t always volunteer our location information. And even when we do, sometimes we lie. I know a certain Canadian who uses 90210 as his zip code and was thus asked by Klout to be an US ambassador for Spotify. A quick peek at his Twitter profile would reveal his true location, but an algorithm can’t do this.

4. Timing: This relates to context and situation. Influence can be seasonal. While my friend Nick (above) seems to find me influential about UNC basketball, I can tell you far more people agree with him during the college basketball season. I’m really not all that influential about NCAA basketball this time of year, but I’m more influential about summer activities in Chicago (something which I don’t have influence over in December).

Trying to account for seasonal influence becomes tricky. Should I be influential about UNC basketball? How quickly does my influence fade? What if I participate in a weekly Twitter chat, but suddenly stop. Should my influence as a member of that community fade more or less quickly than as a UNC basketball fanatic?

5. Reach/Popularity:  It matters. I’d love to pretend that it doesn’t matter how many followers you have, but reach or popularity will always matter to some degree. If I do have expertise on certain topics, I will have the opportunity to influence more people if I have more people following me on Twitter or connected to me on Facebook. Please don’t believe that having one million followers is the solution to anything, but know that if you are an expert, having more followers who care and are influenced by you is better than having fewer.

My biggest issue with measuring reach or popularity is that for the folks with astronomical followings (ahem, Mark ;)), this factor can often skew their influence score. I don’t care how many followers he has, I will never consider Mark a reliable expert on being a teenage girl. Period. Because these numbers can be so high (and because they vary so much), this factor is really quite tricky to fairly factor into any influence measurement.

Some of these factors overlap a bit, but these are the five that I refer to most often when defining influence. What factors are most important to you? How do you define influence?

Rebecca  Denison is a social media analyst at Digitas in Chicago who is passionate about all things measurement and all things UNC.

All posts

  • Hi Mark

    Very good question (as always). Wikipedia defines “Social influence occurs when an individual’s thoughts, feelings or actions are affected by other people”.
    From the standpoint of a professional marketer on the local market it seems to be quite easy. If sales rise when he cried out, he has influence. The more he shouts, the more people are listening and the more are buying.
    So it seems to me, that influence has something to do with a kind of mass appeal or at least appeal in the special target group.
    Nevertheless we all know how difficult it is to change people’s behaviors. You might change their way of thinking, but it’s a long way to make them change their patterns (in live, in business, in family etc.).
    So influence for me means, bringing people to different ways of thinking, maybe to different thoughts and eventually to different actions. If that occurs, I do have real influence. Like Barak Obama, when he won the election. He not only made people believe that “they can”, he made it happen. His influence could be measured by the number of votes he received (I know, this is not 100% correct, because you do not vote for the candidate but for the electors).
    Back to social media: If you have to sell something, you can track the numbers, like how many of “the Tao of Twitter” is sold.
    If you want to know if your book has been read or not only read, but changed peoples handling of twitter… I would say: No way to measure it correctly…
    But you can definitely measure the intensity of discussions you encouraged: The number of RTs and mentions, the comments on your blog etc. But is this real influence? Does it change people behaviors?
    Kind regards from Germany
    Who is trying to influence the German banking world a bit with his Bank Blog (

  • Rebecca,
    Interesting post and great insights! I can’t say I disagree with you on any of your points, but the problem then becomes how to MEASURE true influence. 

    I believe that all Klout metrics are at least somewhat important, and also believe that if I were developing the next iteration of the Klout algorithm I would clearly find some way to make some measures more “real” such as what you have inferred about the “topics” of influence that are derived from individual words in your tweets and how Klout identifies you as an influencer within those topic areas. I do try to go on the site every so often and DELETE those topics that do not pertain to the real meaning behind my words, hoping this helps the algorithm eventually realize the errors of its ways. 

    So, although your points are good, I am curious as to what you suggest as alternative measurements? Or do you think we just shouldn’t try to measure someone’s true social media influence? 

    How can I measure the influence I have on children’s healthcare (CONTEXT)?
    For myself and my organization, it is helpful to know how many readers access my blog–I would prefer to count comments, but I don’t get many–so I use Google analytics for that purpose. Additionally, I DO  have people in the region (such as newspaper reporters and local TV personalities) commenting on the blog via email or by word-of-mouth, so I know they read it, and consider it useful to themselves but there is no way to measure that influence directly. Do you know of any other metrics that might help?

    Ultimately what matters is how many parents choose to obtain their children’s health care from my doctors, and how many physicians in the region feel comfortable referring their patients our way. That is the ACTION you are talking about. And we, of course measure these parameters, but I know of NO WAY to tie those behaviors directly to my activity within social media. 

    As a whole I believe we cannot and should not segregate social media influence from overall influence. And in the end, it may not matter whether one’s influence comes from activity on Twitter, on a blog, via Facebook or LinkedIn, or through print media, radio, TV, or academic circles. Social media just gives us ONE MORE WAY to reach our target audience, and its impact is clearly growing. If you ONLY have social media Klout, as currently measured, you may be completely full of hot air. Eventually those who seek you out for honest communication and “genuine helpfulness” (Mark’s term, I believe) will realize that and your true influence will drop, whether or not Klout measures it.

    So, at the end of this meandering diatribe, I am still left wondering what alternative METRICS are available to do what Klout attempts to do, but to do it better?

  • Wow! Thanks for such a thoughtful reply.

    I love the way you walk through different examples and types of influence. And you hit the nail on the head at the end, it’s just tricky to measure true influence in some ways. We may never know exactly how many people have shifted their strategy because of “The Tao of Twitter.” But we can measure how much it’s mentioned or how often something Mark tweets is shared. It’s a start, but it doesn’t get to the end goal (driving sales and success). Ugh! Frustrating.

  • Rebecca
    Sorry for not reading the intro paragraph. I thought it was Mark as usual…
    By your answer I get aware of the fact, it’s you  😉
    So you see one more point of true influence…
    Kind Regards

  • Thank you for such a great reply, Alice! I hope Mark sees this…this may be his next guest post. 😉

    I agree that most of the metrics Klout tracks are important and have a place when measuring social media. You need to track your retweets and full reach and what not to determine if your audience is growing or changing (depending on your goals of course).

    We should definitely try to measure social media influence, but I think it will take a lot of data integration and manual analysis until a better tool comes along. And it all has to tie back to your goals. What are you trying to get your readers/followers/fans to do? If your goal was to drive RTs, then Klout would be handy. If you want to eventually drive online sales, track conversions. If folks from Twitter are more likely to buy vs those coming from search (more random), that shows you have some influence on Twitter.

    There may not be a way to tie the end goal/action (parents and physicians choosing or recommending your doctors), but you can ask! If you give new patients a survey to ask how they heard about you (and ask referring doctors why they chose you), you may be able to see a direct tie. An indirect method may be to listen to conversations in social media (if available) and see how many people recommend your doctors or say they have chosen you because of your efforts.

    At the end of the day, as you say, social media influence may not be HUGE. You have influence in other ways, as you noted.

    Your metrics should always tie back to your goals. And it may take some manual analysis. When I started on Twitter/blogging, my goal was to become a go-to resource for measurement. My metrics have been things like how many guest posts I do or am I asked to do, how many times someone refers a friend to me, how many times I’m asked a question about measurement and how many times someone calls me a resources or guru or ninja or any iteration.

    I hope that helps. The real answer is there is no one way to do it right.

  • Not a problem!

    Exactly! Now I’m influencing you instead of Mark! I won’t let it go to my head. 😉

  • Rebecca,

    Great post. I particularly liked your example regarding timing and how
    your “influence” may wax and wane seasonally. I hadn’t really considered
    it having an impact on influence as much as spikes in activity around a
    specific topic.  You’re right, any tool (even Klout) cannot (as of yet)
    blend offline and online influence. We’re getting close in some cases,
    where social and more traditional data is correlated for insights.

    I don’t think we’ll rely solely on one tool to guide our understanding
    of our audience. I think Klout provides valuable insights and is
    important to monitor for trending and overall changes within a social
    media network.

    Thanks again!


  • I wish we could blend online and offline influence, but I don’t think the technology is there just yet! But we are getting really close in some places, and I’m excited to watch as this unfolds.

    I agree with you! We should not rely on one single tool or score, EVER. We should always consult multiple metrics and variables, and Klout can provide some great metrics. It is not the end-all, though.

  • Anonymous

    I’m sure I’d appreciate Klout more if my Klout score were higher! That being said, I especially appreciate your point that “influence” needs to take into account actions offline, something that Klout can’t truly measure. The thing I keep reminding myself is that Klout measures ONLINE influence. I could spend a lot of time bulking up my Klout score, but that may not allow me to affect the change I actually want to make.

    Or maybe I’m just bitter. 🙂

  • Wouldn’t we all care a bit more if we were as influential as Justin Bieber according to Klout?? 😉

    Connecting online and offline is probably the hardest piece of this puzzle and one we’re not likely to solve easily. So much of what we do online drives offline behavior. We know it, but we’re still struggling to prove it! Oy!

  • Mark Schaefer

    You crossed the line Denison. : )

  • Larry Levy


    Great insight. We agree that influence can be measured in a wide variety of ways and there are different approaches being taken to get a handle on who is defined as an influencer. At Appinions, our Influencer Exchange doesn’t give people scores. Instead, our topic-focused algorithm looks at who’s attracting the most attention and, as important, the people that these people are writing/talking about. It’s a different approach than Klout but one that we think provides an innovative look at influencers.

    Larry Levy

  • Anonymous

    Great post @rebeccadenison:disqus … when I first started noticing Klout, it said that I was influential on: Sarah Palin (who makes my blood boil), the Republican Party (ditto!), and guns (ummm. because I tweeted about the National RESTAURANT Association being in Chicago. That still makes me laugh.

    You’re right… measurement IS very elusive, and my approach is to view it per situation taking into account all those elements you outlined above. Thanks for an insightful piece.

  • Anonymous

    FYI guys, the link on Rebecca’s name at the bottom of the post goes to a different Rebecca Denison (an attorney).

    Great post @rebeccadenison:disqus … according to Klout a few months ago, I was influential on: Sarah Palin, the Republican Party, and the NRA. It was amusing 🙂

  • That is an interesting twist! I like including who is talking to/about whom. It adds credibility and relevance. If Michael Jordan were to talk about me having vast UNC knowledge, that should give me a bump in influence, right?

    How do you determine topics? Is it keyword based?

  • Ack! You’re right. My blog is!

  • Thanks! I’m glad you liked it. I had some pretty ridiculous topics, too. Most recently I was given Klout in “coffee.” I don’t drink coffee. I checked in at a Starbucks with my family recently, but that’s not really something I talk about.

    Your approach sounds perfect! It’s tricky and often time-consuming, but I bet you come away with a lot more meaningful data than a simple Klout score.


  • Klout shows how many unique commenters, likers, retweeters and mentioners I have had in the last 90 days. As this number grows it is a fair assumption that my “influence” is growing. Same could be said for unique visitors to my blog. 
    From all this good things come.

  • Thanks Rebecca

  • That’s definitely true! I love a lot of the metrics Klout provides because many of them can be rolled up to help show and define influence.

    But do you want to be compared on the same scale as Justin Bieber or me? We’re probably not influential (or striving to be influential, as the case may be) in the same topics or areas of interest? Does it really matter which of the three of us has a higher score if no one will ever come to all of us for the same question?

  • Klout is useless. Well, it has some good data, but people focus on that stupid number and get all warm and fuzzy when it grows. 

    I think businesses focus too much on numbers and measuring things. If you need a service like Klout to tell you it’s working then you have some problems. If you are truly engaging with your audience, you will know because THEY will tell you! 

    I want someone to tell me how checking in at Starbucks, or anywhere for that matter, makes me influential on anything. 

    Follower counts don’t matter, it’s how you interact with your followers that does, which Klout (or any service) can’t track. I would rather have 10 people that are true fans, than 1000 that aren’t.

    Everybody wants validation, but a klout score does not validate anything. Good and bad reviews, customer emails and DM’s will validate you and give you the information you need to grow your business. 

  • I couldn’t agree with you more. Klout does measure some handy metrics that I like to keep track of, but I pay little attention to the overall score.

    If you can get those 10 true fans to mobilize or change their minds every time you tweet, that’s worth far more in my book than having 1,000 fans who only kinda pay attention to you.
    Thanks for your thoughts, Phil! 🙂

  • I like how you call it true influence but it is interesting to note how we immediately (automatically) compare influence to klout, peer index and empire avenue. While I agree that these tools only give you a very vague depiction of influence, people still liken influence to such metrics. Ultimately I think the best metrics one could get are feedback, both negative or positive. 

  • It really is interesting. It’s hard to mention influence in social media (especially on Twitter) without someone mentioning Klout or Peer Index. I will keep fighting until they’re no longer synonymous. Klout has a value on its own, but it is not influence!

    Feedback is always valuable, and it will often provide far more valuable insights for you than just an influence measure. That’s a great point!

  • Ohh.. just you wait sir, thinking of a clever RT for this.

  • Well said, all of it. One factor I’d add – which you allude to in your list – is motivation, why. Why did I RT a Mark Schaefer post, is it b/c I think it’s smart? Will it give ME influence w/ others, sharing a smart post? Did I just do it b/c I want him to RT one of mine, benefit from his influence? Why someone is tweeting, linking, sharing something about a particular topic at a particular time could have a million answers, somewhere within lies context, reason, motivation – what is really influencing them to take those actions. 

    See also the K+ stuff, the follow/unfollow game, etc. that have little to do w/ real influence. It suggest linking the FB account which for those like myself, is personal and has negative impact on my professional influence, ergo it’s not linked. It doesn’t (last time I check) look at my other online activities, areas of true engagement like blogs. I’ve mentioned it before that it measures online usage, the activity in whichever accounts one links to it; it doesn’t really cover online influence, much less true. There is room for improvement but it starts with what will be defined as influence; counting a simple RT isn’t it. FWIW.

  • You make a good point about measuring intent or motivation. I wonder how easy that would be to track? I approached the five factors with the idea that with today’s technology and some manual work, they could be measured fairly reasonably. Motivation would be far harder to track. But it definitely should be a consideration.
    Thanks for getting me thinking again!!

  • You’re very right, it’d be very difficult to track, measure, fit into an easy algorithm. People follow some big # folks that they really don’t read or engage with, it’s not about influence.. it’s about them wanting to pretty up their own scores. I think I’ve mentioned motivation in almost any comment or post on Klout b/c when I think of what influences me – online or off – it always circles back to why.

  • Should motivation affect your score or mine? If you tweet something, and I RT it for a reason…does the reason affect how my influence is changed or yours? I think that’s the part I’m unclear about.

  • I gotta figure out a better way to explain to but IMO it changes what the ‘influence’ really is. I could RT all your posts on Klout for example, which would give you points on Klout. But if I’m doing it b/c I want my own influence, want the pick up and RTs myself, then it really isn’t about your or whatever influence you wield. Or take sentiment.. if every one of my RTs leads w/ “this is wrong” then it’s kind of a false barometer of influence. IDK.. like I said, I need to think more on how to phrase it.

  • I think I see where you’re going. It’s definitely a tricky thing, and it does feel like it’s always shifting.

  • Pingback: Five reasons why measuring influence is elusive | Social Business Marketing |

  • Larry Levy

    Rebeca – topics are keyword and concept based. This means the user can use simple or complex Boolean to construct their topics or rely on our concept extraction to formulate the topics. I hope you get a chance to take it for a spin sometime soon or visit us at the up and coming Pivot conference where we’re on a panel with Brian Solis.

  • Sounds interesting. How do you determine and define concepts? Is there human analysis in there, too?

  • Pingback: Marketing, Public Relations, Social Media | 3Hats Communications()

  • Paul Wilmott

    Rebecca, Klout is fantastic at finding people who are great at creating, aggregating, and sharing online content.  A similar company, Meet Insight measures the level of trust (and therefore influence) a person has engendered  This works perfectly in the marketing of ‘trust based’ services.  5 years ago neither would have been used or popular. How the world is changing..

  • How does MeetInsight measure trust? What metrics or inputs are used?

The Marketing Companion Podcast

Why not tune into the world’s most entertaining marketing podcast that I co-host with Tom Webster.

View details

Let's plot a strategy together

Want to solve big marketing problems for a little bit of money? Sign up for an hour of Mark’s time and put your business on the fast-track.

View details


Send this to a friend