Social Influence Measurement: The Good, the Bad and the Ugly

I was inspired to do another post on social influence measurement after seeing klout get trashed on a few sites.  I understand why people argue against it, but I think it’s helpful to look at it from all sides.  The post ran long so I’m running in a series: The Good, the Bad and the Ugly. Starting with…

The Good

Before we get into it, let’s set up what we mean by “social influence”. I’m using the definition Forrester analyst Zack N. Hofer-Shall suggested:

“The power individuals have over your brand based on the size and quality of their networks.”

Of course social influence vendors don’t publish their algorithms, but klout does provide clues on how it derives its score:

[important]klout score is comprised of: True Reach, Amplification, Network Influence[/important]

  • True Reach = the number of people someone influences
  • Amplification = how much you influence people
  • Network Influence = the influence of the people in your True Reach

Intuitively, this equation feels right. Doesn’t it?

In general, measurement is good! Metrics provide transparency and allow for meaningful, quantitative analysis and discussion. Yes, numbers can be used to distort the truth, but I don’t believe the social

Believe it or not, Blondie (Clint's character) was "The Good"

influence vendors have an agenda other than to build a sustainable business model. (They aren’t trying to get anyone elected.) As long as the algorithms are applied evenly across social media, they provide useful data and insights.



Erica Ayotte (@inthekisser), a Social Media Manager at Constant Contact, pointed out that even if Marketers can agree on who is influential, what do you do with them after you’ve identified them? Well, klout has an answer for that too. Social influence measurement is about more than keeping score. For example, klout has a perks program that allows marketers to reach influencers and offer them exclusive packages with the intent of supporting a Word Of Mouth campaign.

[important]See a list of current perks and claim those you qualify for <here>[/important]

Sure there is plenty not to like about these metrics, but social measurement vendors (klout, PeerIndex, Empire Avenue, etc) are still new and they continue to improve their services. For example, klout, who gets the lions share of attention, keeps adding services and features. klout uses its algorithm to measure influence on 12 social media networks. It currently has a new feature called “Topics” in beta. Topics provides users the ability to see the top ten influencers on a given topic.

Okay, social influence measure not perfect, but as Voltaire said “the perfect is the enemy of the good.”

The Bad

Let’s keep it simple, here are 4 things that are bad about measuring social influence:

  1. A high influence score doesn’t necessarily mean this person is actually influential. A high klout score would indicate an individual has a larger audience of influential people that engage with the content being shared. But does this measured influence cause action? Does a high score mean that person can help a brand move more product? Not necessarily. There are other factors that come into place like product/service accessibility, location, suitability and personal taste.
  2. Klout lacks context – you may have a klout score of 70, but that doesn’t mean you know jack about sewing, mountain bikes, grilling, scrap booking, etc. This leaves the scoring system pretty empty – you need other data to provide context be it (demographics, behaviors, etc.) However, klout is in beta with Topics. If it works the way one would hope, this will be a huge improvement in the system since it proves context. I may have a score of 70 on the Boston Bruins, but a score of 1 on dating Gisele Bunchen. For what it’s worth, PeerIndex also provides some context by measuring influence on eight industries.
  3. Some people fault social measurement scoring because it can’t t measure offline influence. I think that’s a specious argument. If you’re already tracking offline influence good for you. Consider their digital score additive to your assessment. But admittedly, using the definition offered, the solutions available do not measuring online influence. That’s bad because marketers can become overly focused on a digital score and totally miss the point of social influence.
  4. Vacation – if I go quiet for a week, my score drops ~5 points. Did I suddenly become less important?

I think this list may be too short, let me know why you think social influence measurement is bad in the comments.

The Ugly

I’m calling two things “ugly”:

  1. What if influence doesn’t matter?
  2. Data privacy

Influence doesn’t matter?

A Fast Company article, Is the Tipping Point Toast? explores the research of Duncan Watt.  Watt ran computer simulations to determine how things (viruses, ideas) are spread. It’s not that influence doesn’t matter, it’s that influence doesn’t matter as much as we think it does. Watt’s said:

All the clever (and lucrative!) targeted viral campaigning may ultimately be less effective than good old mass marketing.

Here are Watt’s reasons:

  1. No one can clearly explain how an influencer actually influences. Is it one discussion? Several discussions? A mix of activities? Is an influencer so influential that they mobilize the influenced to

    Duncan Watt

    also be a strong brand advocate?

  2. Societal readiness is the determinate factor regarding whether or not an idea spreads. He uses the example of forest fires. There are thousands of forest fires a year, but only a few become roaring monsters. A specific mix of environmental factors are required for a hugeforest fire: dry woods, lack of rain, distant fire departments, etc. Intuitively, this makes sense. Cultural factors play a part in how “right” ideas are for a community.
  3. Marketers like the idea of influence because it exudes a measure of control. (and that control can be sold to clients.)

Data Privacy

I’m a cup is half-full type of guy coupled with some good old conspiracy theorist. When I learned of klout, it could only measure two or three social networks: twitter, LinkedIn and Facebook. Now that klout has scoring on twelve platforms it starts to feel less like influence management and more like data collection.  Let’s be real. Just as on Facebook, you’re not the customer, you’re the product.

[notice]You’re not the customer, you’re the product.[/notice]

I’ve reviewed klout’s Privacy Policy <here> but it deals with the information you share directly with klout, not the information you later give klout access to when you connect klout to social networks such as Facebook.  For example, when you connect klout to Flickr you see this warning:

"including private content"


It seems that I’m granting klout access to information that is kept private on Flickr. I asked klout Marketing Manager, Megan Berry (@meganberry) about this. (For the record, Megan has always been a real straight-shooter with me in my brief exchanges with her.)

Regarding the Flickr example, Megan replied:

On flickr: Your private photos will definitely stay private. We are
asking for the lowest level of permissions flickr allows which is
read-only access and does include private photos (we can’t change
that). However, we NEVER use your auth to display photos, or look at
them or share them. Instead, our algorithm looks at the response your
photos are getting — i.e. are people commenting, favoriting them etc.
We will never access or use your private information.

On privacy in general:

Klout collects data from your networks in order
to discover your influence. We do not share any of your private data
or email address with 3rd parties. We do not access your Twitter DMs,
private Facebook messages, private Flickr content or any other private

I want to believe Megan, I have a lot of respect for Joe and the whole team, yet unless I’m misunderstanding the privacy policy, what Megan shared is not in their stated klout privacy policy. Why not spell it out further?

My advice is for users to be aware of what they are agreeing to be connecting your social networks on Empire Avenue, PeerIndex and klout. Don’t take your privacy for granted. If your concerned, don’t link your accounts. Some measure (a little? a lot?) of the scoring comes from the public view of your web 2.0 life. Perhaps that’s enough for you.

Feature photo credit Rusus Gefangenen


    • Allen Wong on September 30, 2011 at 12:59 pm

    I think with influence, as with poker, the best answer is that “it depends”.  Depends on the situation you are looking at to discuss influence.  An example is someone like Britney Spears.  I would think her Klout score (if she even has one) should factor in music somehow.  But I wouldn’t consider her influential at all since I don’t think she is a great musician.  

    Another example would be if I mentioned to followers that I went to a hotel that had bedbugs.  My Klout isn’t that high, but the context of the message is more influential to people since most would instantly avoid such a hotel.

  1. I would definitely give you +K in poker along with my ante. I would bet that you’re way more influential at poker (or photography or a host of other things) than Britney. I’ve argued that what the vendors lack in their calculations is “context”. But I believe that is on the way. The “science” of measuring influence is still more of an art form, but I think we’re still in the early days.

    thanks for your comments.

Leave a Reply

Your email address will not be published.