Measuring Influence: 4 Learnings

Measurement isn't just bells and whistles

Measurement for its own sake is a waste of everyone’s time and money. It’s got to be in service of a strategy.

You might say that the foregoing statement is a canard; no one is beating down our doors asking us to just measure something, anything.  But there remain a feisty few, particularly on the social media side of the equation, who keep offering up horsepuckey in the guise of gold bullion.

Witness “4 Ways to Measure Social Media…,” a well-intentioned piece from last summer on Social Media Examiner. Author Nichole Kelly subheads the article with “exposure,” “engagement,” “influence” and “lead generation” — the “4 ways.”  Kelly’s on firm ground about exposure, pointing out the difficulty of a) getting good data and b) ensuring you’re counting only once, though equating reach to awareness is a colossal mistake.  Engagement,  too, is solid (if output-based), covering @replies, DMs, links clicked, comments and subscriptions. Good stuff.

Influence is listed third and lead generation fourth, showing exposure, engagement and influence as the top of the funnel leading to conversion.

The section on influence is underdone, and erroneously says tone (positive, negative, neutral) IS influence.  In fact, according to Yahoo!’s Duncan Watts, Winter Mason, and Jake Hofman, and the University of Michgan’s Etyan Bakshy, influence can’t be credibly determined from content analysis. Read all about it.

I heard Watts speak on this topic during the snowy last week of January at a meeting of the Institute for PR Commission on research, measurement and evaluation, of which I’m a member. Influence is a huge question, and Watts,’s work made me recall the somewhat hoary idea that understanding your specific audience (whether final audience or intermediary) is a lot more important than trying to calculate the exact number of impressions represented by friends of friends and retweet followers.

I pick on influence because it’s the biggest question in social media.  In fact, it’s been a big question in communication in general since the days of Lazarsfeld, Berelson and Gaudet and the two-step flow. Who are the “opinion leaders” and how do we calculate their effectiveness?

Here are four questions that hold promise when considering how to measure influence:

  1. Does the opinion leader “play” in the right sandbox for our intended audience/stakeholder?  Chris Brogan and Brian Solis have lots of followers, tribes that hang on their every tweet. Are their tribes our tribes?  They’ve got awesome scale by sheer numbers, but it’s anyone’s guess how involved they are or whether their followers in turn reach people we care about. We could get Brogan or Solis to talk about our service, product, leader or whatever, but to what end if their followers aren’t the right fit for us?
  2. Can we create a solid chain of links from the opinion leader’s actions to our desired actions?  If we’re working on building corporate reputation, retweets, Facebook “likes” and blog comments should have a relationship to opinions voiced by our final target audience. Simply passing along a leader’s statement (tweet, post, comment, etc.) shouldn’t be construed as adoption! Here’s where content analysis shows promise, especially in blogs and perhaps during Twitter chats. The opinion leader’s output should have some effect if he/she is truly influencing others. Note that this is a qualitative effort and suffers from lack of scale.
  3. Are we mistaking popularity for influence?  Celebrities routinely land atop the Twitter rankings, and there are brands on Facebook with upteen hundreds of thousands of “friends.” But having a lot of friends/followers just makes you popular. See #2 above.  We’ve long wondered about how to judge the effectiveness of influence in conventional relationships, but I don’t think many of us think the most popular student in high school was necessarily the most influential.
  4. Are we inappropriately drawing general conclusions from narrow findings?  Influence is personal and specific.  We make assumptions about readers of newspapers, TV viewers, etc., and have a body of research to back those assumptions up.  In social media, the appearance of influence may be mere output, or outtake at best. Outcomes outside of e-commerce are tough to come by, though clear objectives can solve this problem quickly.

The best measurement starts with research up front, which informs our strategy and objective-setting, followed by more research to determine effectiveness and progress toward objectives.  It’s not just tactical measurement designed to cover our butts or justify our budgets, especially when it’s trying to measure influence.


Tags: , , , , , , , , , , , ,

6 Responses to “Measuring Influence: 4 Learnings”

  1. Julie O'Neil says:


    Nice piece. I agree that influence is the key variable to figure out. And, the impact of time should also be considered when analyzing influence. A person may hear something online and act on it three or four months later. This element of time complicates measurement.

    Julie O’Neil

  2. Sean says:

    Thanks Julie — @kdpaine also has an excellent post on the subject. I just read it on Ragan…

  3. Rich Becker says:


    Marketing and public relations professionals are absolutely mistaking popularity for influence. The reason is simple enough.

    Influence attaches to ideas not people or media or institutions.

    Where the individual or institution plays a role is in its ability to convey expertise, authority, and reputation, which is established over the long haul. And the size of sphere around a person has very little to do with the intensity of the influence.

    I could make someone an “influencer” based on existing measures within about 30 to 90 days, less if the topic has an intense following. It isn’t even very difficult.


  4. Sean says:

    Hi Rich – sorry to be so tardy. Thanks for the comment. You’re absolutely right. I didn’t even get into how gaming the system is so endemic. There’s a lot of crap out there…

  5. @MetricsPundit says:

    Most of the software folks playing the measurement game have even less research foundation than the 22-year old Social Media Experts using it. And so long as Users can point to screenshots of their favorite free tool as proof to the boss, they will prevail. I hope the VC’s choke on their investments.

  6. Sean says:

    Hi there @mp — thanks for the comment. It made me laugh — I think the biggest question in PR right now is regarding the role of research. It’s being abused terribly in a lot of social media context, with outright sophistry and deep black boxes preventing scrutiny and scholarship that could help our profession immeasurably.

    Did you see that Twitter’s been valued at more than $10 billion? The VCs are SMOKING their investments…