The Uberinfluencer and Bottom-Up Networks

Malcolm Gladwell
ImageMalcolm Gladwell by the Business Makers

“Influencer theory” as it is discussed today on the social web is a pop myth. The various schools of thought lack the substantive analytical scientific study and proof to be considered legitimate or factual. Yet they are used by tens of thousands of marketers to strategically position themselves, their companies or clients online.

Auburn Professor Robert French said in a comment on Friday’s round up of these theories on SmartBrief and pictured below, “Frankly, all of the efforts in social media that I have seen aimed at defining influence and influencers are tools that seek to (a) drive sales of a company/interest, (b) try to elevate a company/interest into some form of ‘thought leader’ or (c) pseudo-research disguised as marketing promotion — maybe even (d) sell a book… You can’t build a theory on anecdotal evidence.”

The State of Influencer Theory

State of Influence Theory Infographic by Jess3, based on discussions of influence on Welcome to the Fifth Estate.

When exploring the infographic in depth notice the alignment of top-down versus bottom-up influence concepts, as well as the basic offsetting of these two theory families, with Gladwell and Watts taking opposite sides. The Gladwellian top-down theories include One Percenters, Trust Agents, leaderboards, and to some extent Free Agents (to be fair, Kanter and Fine’s theory sits within their larger theory of networks). The Watts bottom-up family of theories include the Magic Middle and Groundswell.

Indeed, when you look at the two popular schools of thought, the worse for the wear is the more popular top-down school, first propagated by Malcolm Gladwell. This school holds that there are “uberinfluencers,” a few types of people that can move entire networks and organizations toward action. Unfortunately, as time has moved along (from left to right), these theories have become less and less data centric, and as French says, built on conjecture.

Every top-down single theory is positioned within a marketing book with the exception of leaderboard technologies, which are selling their services. In essence, these ideas are increasingly popularized by bloggers and companies who have directly benefited from their school of influence becoming popular. In the case of the most popular theory – Trust Agents – we see a reinvention of the one percenter theory with a healthy dose of Edelman’s Peer Trust study in it.

The idea of an uberinfluencer is reminiscent of the romantic era of cigarette advertising, when the Leo Burnett Agency creation the Marlboro Man touched off an icon of male sex appeal. Unfortunately, he was more likely alone in the desert dying from lung cancer.

Similarly, the uberinfluencer theory is very popular, and dramatacizes the impact of one person in the midst. America loves the idea of a strong man/woman leading the pack. But as time has shown, these theories do not apply universally to contagious events online and in real life. It was quite amazing to see Malcolm Gladwell’s proclamation that the Revolution Will Not Be Tweeted last fall, only to have his words thrown in his face three months later with the Middle East uprisings.

There are some elements of truth to the top-down theories. Content creators or voices are important, and some become important due to their ability to generate peer trust (originally depicted as Connectors in Gladwell’s book), both of which can help set off word-of-moth events. When, where and how has become less and less specific as these theories have evolved.

The Contrarian Bottom-Up School

Duncan Watts delivering keynote address at SES New York
Image: Duncan Watts by the SESConferenceSeries

The bottom-up group, started by Watts in 2003 in direct response to Gladwell’s theories, has a different view, believing in a more networked effect to contagious events. Influence lies in smaller groups, more individuals and ideas/movements spread through weak ties. As an idea becomes popular in small groups and spreads, content creators and conversationalists notice it, and like their friends, help spread it across their extended relationships. This is the networked effect of influence.

All three of the bottom-up theories have the same weak point, that they are introduced within marketing books and, in the case of Sifry, a technology service measuring and selling blog influence. However, all three used analysis of data they collected to form their theories. At the same time, they have not had their results verified or replicated by third parties.

Watts, however, did a great deal of data analysis after his book came out while he was with Yahoo. He sparked an epic criticism of Gladwell in Fast Company, “Is the Tipping Point Toast,” and continued to revalidate his theories:

“A rare bunch of cool people just don’t have that power. And when you test the way marketers say the world works, it falls apart. There’s no there there.” And then, “‘All they’ll ever say,'” Watts insists, is that a) there are people who are more influential than others, and b) they are disproportionately important in getting a trend going.”

Well said, and true, at least in this marketer’s experience. The Magic Middle of all the theories has had the most weight with blog contagions, but as Watts so well makes clear in his work, contagious events are very, very unpredictable, and who will start them is an unreliable guess at best. There is more to it than simply marketing. There must be a need, and a societal readiness.

Back to the Fast Company article: “If society is ready to embrace a trend, almost anyone can start one–and if it isn’t, then almost no one can,” Watts concludes.

The less popular, but much more data centric bottom-up theorists hold that social network influence is not the domain of reach. They go further to show the strength of relationships matter more, that relationships and roles (a la Li and Bernoff’s Technographic profile) can change by the time or situation, and that more contagious events start below the uberinfluencer level, then rise up as events unfold.

Chasing Influence

As we can see, none of theories have nailed how ideas spread through communities. Only Watts’s work has yet to have a major hole punched in it, other than to say it is frustratingly murky and undefined. It seems that influence has yet to be defined in the social web era, much less in the 100 years of prior sociological study.

What is clear is that influence is perceived as online power. As long as other people are willing to pay to promote their ideas, products and services, influence theory will continue to evolve (or devolve, depending on theorist). But for now, despite evolutions and wrinkles, it still appears to be a battle between Gladwellian uberinfluencers and Wattsian groundswells.

As to the myth of the uberinfluencer, it has as much likelihood of producing a contagious event as Barack Obama’s odds of resolving the debt crisis without the help of U.S. Senators and Congressmen in both parties. Yet, as long as we live in a time where popularity and attention are valued, this myth will remain strong. Welcome to pop culture.

What do you think of the State of Influencer Theory on the web?

27 thoughts on “The Uberinfluencer and Bottom-Up Networks

  1. Lots of agencies have a good deal of time (and their credibility) invested in influencer theory. It’s going to take a while to get them to shift their strategies to reflect a more data-driven approach.

    • That is a good point. Really the entire PR market is built off of the uberinfluencer theory or a version of it.  Won’t go away anytime soon.

      • This is a really interesting piece, we come from a background of social marketing mixed with behavioral science and heavily data driven , that said co creating services and interventions and using a wide tool set including design thinking, and systemsthinking for us has ben the way forward. I agree agency’s have to wise up. I would love to share wider and deeper. do you facetime


  2. While I don’t agree with the premise (i.e.: “Influencer theory” as it is discussed today on the social web is a pop myth. The various schools of thought lack the substantive analytical scientific study and proof to be considered legitimate or factual) much of what you say here is accurate.

    Many/most of the top down approaches have been shown over and over in real world applications to fail miserably the majority of the time, at least in regards to social networking distribution models.  However, to say that about the ‘bottom up’ approaches would seem to miss the mark, at least in regards to the circles I run in.  But that’s not really critical to my point.

    It’s the *application* of influencer theory that in my opinion most often misses the mark instead of the theories themselves.  Was the right approach selected for the situation, and was its execution handled well?  Was the right study done initially to understand the psychographic and sociographic nature of the audience in regards to the desired response?  Was the desired response a reasonable one for reaching the objective?

    Manufacturing ‘contagion’ with any real degree of accuracy is the myth.  Dictating the wildfire distribution of ‘viral’ simply cannot be done when it comes to social (media) networks.  You can predict the varying degrees of (low) likelihood with a large +/- swing, but that’s about it.

    The issue I see in most cases is people trying to bucket the notion of ‘influence’ into a very narrow band.  Types of influence are widely varied and the applications for manipulating that greater still.  ‘Bottom Up’ or ‘Top Down’ is generally a misnomer except when applied to marketers and who they (often wrongly) focus their attentions on.  Actual distribution tends to not follow those distinctions because most influence is so highly contextual that a person with 5,000 followers can easily hold as much sway on a particular niche topic as someone with 500,000.  The Klout’s of the world have their place, the amount of attention and hope placed on them however, does not.

    All this said, there *is* a great deal of research out there if one cares to find it, understand it, and apply it properly.

    Cheers, and thanks for a good post.


    • While there is some truth to your comments, you are missing a factor.  The real secret to having any contagion is product marketing. The reason why Google+ took off is because they met a market need. Great product marketing builds things people need, and positions those products to successfully meet the gap. 

      Google+ is a second generation network that meets a gap not met by Twitter, and to some extent created by Facebook’s UX and approaches to privacy.  All the big and small influencers can talka bout it all day, and it wouldn’t take without this. That’s the difference between Quora and Google+. Of course, Apple is the best company at product marketing out there.

      Also, I willn ote your objections are the same as Watts, and have been expressed by him very clearly. As to the amount of research, I do agree with Robert’s critiques that they are ot open, and if they want to be taken seriously and demonstrate true market leadership they need to be.

      • To paraphrase, ‘no amount of marketing or efforts to shine a turd will make it anything but a turd’. And to that I say “ummm, yes” :)

        In regards to open research, I guess the question is ‘be taken seriously by whom?’ and if we’re calling research the same thing.  It’s a pretty small group that works at the top levels of this space, and virtually all of the enterprises that are serious about leveraging it know who they are and I hesitate to say take them very seriously.  When I say research I mean actual research, such as mapping datasets, overlaying behavioral analysis, etc. to understand how communications flow within social networks.  There’s quite a bit of that in my opinion.  It sounds like you’re talking more about attempted applications of that research towards marketing objectives, effectively…case studies.

        I’m not a huge fan of them, but that doesn’t meant that I don’t attempt to share what I can.  I talk with agencies and consultants every week who call with questions on this topic and are trying to do something very specific.  The vast bulk of those conversations are free and done simply because someone has a problem in an area that I find interesting and that I can quickly contribute some value on.  I’ve also genericized what I can, when I can, to get around NDA’s and be able to put something out into the public that might be helpful to others. (e.g. )  So yeah, the mature snide comment of yours that added nothing of value to a conversation about a blog post that I found very good was disappointing.  If you want to discuss a difference in philosophy with me then I’d prefer you just do that.  There’s really no need for the sandbox mentality.



  3. Great points, all, Geoff.  The idea of influencers (popularity, really – not influence) may be easier to sell (or pitch) than a data driven approach.  

    Funny, isn’t it.  And we (PR people) often talk about clients wanting hard #s to justify commitment to a campaign and proving ROI.  Well, then why don’t we actually give that to them?

    Even the concept of credibility (and how audiences ascribe it to sources) is not well researched in PR.  It is – way too often – pop culture today, IMO. Thank you for the thoughtful pieces you’ve shared on this, and other, topics.

  4. Have to disagree, Matt.  If your claim that there is “a great deal of research out there” relates to PR and marketing campaigns and applications, I’d love to see it.  So much (too much) is held as proprietary and not released.   How about all the authors of books cited by Geoff in his previous post on this topic?  I’d love to see them (including Forrester) release *full* methodologies and data sets for heuristic replications of their studies.  Good luck finding those citations, Matt … and those are the ones that really matter.

    • Research…is research.  There’s plenty of it out there.  The application of that research on the other hand (a la case study) is typically so case dependent (as it should be) that it holds limited value to outside parties, but holds a great deal of value to the one who spent the time crafting it.   And because of the nature of most of these objectives you’d have to release far too much inside information about a companies customer/prospect profiling.  And to be completely up front, rarely do consumers want to be viewed as ‘something to be manipulated’ which is how a lot of this would come across, so companies are hesitant to allow it to see the light of the day even if it was something I wanted to do.

      So I wouldn’t go looking for practitioners to release that information to anyone outside of clients under NDA anytime soon.  And that includes me.  While I applaud those who can find ways to maximize their revenues while giving away their “full methodologies, data sets, and heuristic replications or their studies” I am not one of them (nor are any others that I know outside of academia).



  5. So, as suspected, “influencer marketing” is more theory than fact with even something as trackable as the web unable to provide the necessary evidence that there is untapped network of people ready to make whatever 2011’s version Hush Puppies a sensational out of left field trend.

    It’s entirely predictable that in a world of fractured media consumption, social media marketers will say there is a needle in every haystack, but be unable to tell you exactly which one it is.

    It’s safe to say there are influencers in every crowd and many of them are online and maybe even ready and willing to crow about you in online social networks. Perhaps marketers should worry less about trying to find influencers and be more focused on letting influencers find them, but being ready them when they do.

    • Attraction not promotion is a very compelling argument in communities.  Unfortunately, Twitter went the other way, but I believe that Facebook and G+ hold promise.  Thoughtful comment, Dave!

  6. If we’re talking about marketing, I think it’s important to remember a couple of things. First, most people probably do not know who any of the people mentioned in this post are. They have no idea (and don’t care to even begin to explore) what “social media influence” is. Second, people who are buying products and services may be influenced by past experiences, their mom, their dad, their best friend, or the über influencer in school (that really cool kid who was the first one to get a car).

    Although Social Media is an important part of the marketing world as a subject and as a tool, 2% is 2% when it comes to population use, right?

    Also, I think there needs to be a delineation between “influence” and maybe something like “cajoling.” The Fast Company “Influencer” project should have been called something else like the “Fast Company Begging Project.” I was not influenced to click links. I was begged and pleaded with to click links. The same holds true for how speakers are chosen for SXSW. You are not voted in because you’re influential. You’re voted in because you beg, plead, cry, and otherwise manipulate (in a nice or in a mean way) people to vote for you.

    To me, influence means that the sheer gravity of a person has an effect on your axis. I’m influenced by Mother Theresa because the sheer tangibility of what she did speaks to me, and it makes me want to be a better person. A truly influential person does not need to write a book about why they are influential. They don’t need a nickname. They don’t need to deflect criticism about why they may not be influential. They just are. Or are not, to quote the influential Yoda.

    Every comment should end with Yoda. Just sayin…

    • LOL, and the bucket of cold water lands.  You and @Mediaphyter would get along so well!  I can’t say I disagree with you. There are man legends in their own mind in this business,

      Yeah, the influencer project. So much to say, so little that would be deemed suitable for public consumption Oh well.


  7. It is my personal opinion in the same vein that beauty is in the eye of the beholder, so is influence in the areas of social media. The simple (but not limited-to) fact remains: social media influencers are great at influencing … other social media influencers. I often break down the ‘State of the Influencer theory’ into an example that leaves me both disinterested and upset (to a degree)…

    [Disclosure: I’m just picking names to use as an example, nothing more]

    Scoble could read a blog written by someone from Des Moines Iowa, that has 3 RSS subscribers that he happened to come across in a deep google search for terms. To which he then takes, puts the ‘scoble touch’ on the blog topic, and subsequently the ‘scoble army’ thinks he has a direct line to the mouth of God.

    I guess the point I am trying to make is I think the idea of Groundswell vs. Top-Down at its purest of points is something that truly is a crap-shoot. I’ve seen entirely too many data-sets that support either, but as we all know — any marketer can turn numbers into magic. Consumers buy into ‘influencers’ for the same reason that they purchase the US Today in the checkout line at the grocery store .. someone else’s life is better, someone else can tell the story better. The basis of consumerism is rooted in a follower mentality. I purchase because you tell me I should, or because the cool kids are, or because it helps me get the girl.

    Something I’ve always struggled with is a very simple question, at least to me: at what point is it the same people talking at each other about the same thing, but in reality the baseline for the opinions never change? 

    I often get the feeling that the first person to grab the mic and share a theory wins, and then more time is spent supporting the theory than testing it.

    Which begs the question: is validity measured in credible action-results or credible action-support?

    • I think we saw the influence bubble with Quora.  Unfortunately, action support wins on the social web; however, that has brought quite a bit of scorn and doubt upon it from the business community. And rightly so.  Good comment, and thoughts.  Thank you for analyzing it.

  8. Since most people spend their time reading opinions that support their own, it’s obvious that people will follow the “influence” of someone who does that. My readers already agree with me before I begin a post…or they don’t read me.

    • Not so sure I agree with that.  I follow lots of people I disagree with, and often get comments to the contrary here, and on my social networks. Just ask @Bethharte:twitter So, perhaps, perhaps not.

  9. The Scoble-Godin-Brogan (et al) effect has to be considered as well. At some point, no matter how smart, insightful or cool those guys are, when someone sees their posts 219 times per hour (thanks to RTs, shares, etc) and then sees tons of comments fly by, you can’t help but change the channel (eg: mute that post in G+ terms). Not sour grapes – they worked darned hard to get themselves heard at that level, but on G+ in particular the S-N gets particularly low as a result. If everyone quotes Seth, is everyone smart?:)

    • I’m not sure I agree. I have Scoble in G+, but not Godin or Brogan and find all of their signal to noise ratios to be relatively low. So I think this is a question of individual networks. I tend to gravitate away from the sycophant crowd as I find them to be as interesting, as well, Wonder Bread. 

  10. Pingback: What the Circles Illustrate About Influence | Geoff Livingston's Blog

  11. Pingback: Polarization, Fanboys and the Non-Middle Ground

Comments are closed.