- March
Posted By : Adrian Chan
Influence on Twitter

I have little time this morning to write the post this topic requires, but it’s been on my mind for a couple weeks and rather than wait, I’d like to put a few things down on the page. The issue of influence on twitter is not as simple as it might first appear. Twitter is becoming an increasingly differentiated social space, as we’d expect, but its lack of social structure and its limited forms of interaction engender a lot of social distortions. If it is the case that a social field exists on twitter, and I think it’s clear that ones does exist, then social differentiation is an anthropological and social necessity. People want to know who they are and where they stand vis a vis the rest of the audience.

Influence has been the most popular way of differentiating people on twitter. In its simplest form, influence is just popularity, as measured by total number of followers. But we know that doesn’t really say much. It’s a kind of “circulation” model, akin perhaps to circ numbers in the print world, or viewers and listeners in tv and radio. Not very sophisticated, we have to admit, and far less than what interactive media can offer. That said, influence as popularity is, well, a popular metric.

So let’s try to break this down a little.

For starters, a few observations. Influence is not absolute, but is relative. Power is not absolute, it is relative. And sociologically speaking, influence can be understood as a form of power. There are different kinds of influence.

Some types of influence:
Social influence, as rank or position within a community: the community may be defined as everyone, or a subset of everyone. Some influence metrics for twitter, such as Twitalyzer, measure one’s influence against the entire twitter community. Others, like Klout, must lop off the celebrities (with 250k followers, they are outliers and skew results dramatically) and provide a more meaningful influence score. A simple rank of one’s position among all twitter users strikes me as meaningless. (Although it must be fun for those at the top.)

Local influence, as rank within a social group or local community: It makes more sense to measure influence within a circumscribed group of people. The question then becomes “which group?” One answer might be: whatever group we can observe and measure. But let’s not limit ourselves to what we can observe and measure today. Let’s instead consider some different ways of qualifying influence theoretically:

Influence among followers, those a person follows and those who follow, differentiated by reciprocated and non-reciprocated follow relationships. (Note that without going into details, there is a high degree of distortion here, and dirt in the data, for some people deliberately ignore, say using Tweepler, and some ignore inattentively and coincidentally, and some ignore accidentally. A lot is missed on twitter, and not all people manage followers in the same way.

Influence among an affiliative group, the types of affiliation ranging from industry to social or cultural topic and context, profession, place, and so on. This is difficult to determine today, but over time can be extracted using hashtagged tweets, linguistic markers (words, topics, people names, event names, industry terms and so on), retweets (who is retweeted, what associative affiliation does it suggest), @replies (ditto), and twitter profiles. Scraping other social network profiles will help to qualify this. Here, Peoplebrowsr strkes me as a strong contender but one which, if it catches, will change user behavior also.

Influence as social capital, which might seem like social rank but involves more of the person’s actual and perceived social capital. What do I mean by this? There’s a well-known distortion in the follower influence/rank model on twitter: people increase their own influence by following the highest-ranked people on twitter. The crude version of following, in which a person follows popular people, probably involves associative social ranking.

By association oneself with a celebrity, influence travels back to the fan (in their minds, of course, but this doesn’t matter if a lot of people perceive it). It is an act of following that involves mimetic desire perhaps, and is both strategic and indirect: the person who follows john mayer, with nearly 400,000 followers, is possibly benefiting from John Mayer’s huge social rank.

As I understand it, social capital is more than simple popularity, is more than numerical or quantitative popularity, but is a qualified popularity. John Mayer, who follows 19 people, cannot have conversations with his followers. His fans are listening to him, but he’s talking at them, not with them. Social capital seems to me a measure of influence in conversational media, and thus involves the expenditure of social capital through conversation and social interactions. It is more than perceived value of capital (John Mayer). it is transactional capital. Because the currency of social media is attention, social capital involves a unique kind of influence: it increases as it is spent, and spending it costs nothing. Social capital involves relationships, and increases with reciprocity. It is engaged and participatory, both of which suit the online medium more than they do the broadcast medium.

One question to come out of social capital as influence is whether or not it is transferable. Does it belong to the person, or can it be extended to what the person talks about? Can it be transferred from a person to a role or position? Do people with lots of social capital transfer it to companies or brands they might represent? Because so much of social capital is built on authenticity and integrity, I think these are as-yet unanswered questions.


Moving on then, there is more to say about influence. There are influential practices. Strategies of accruing influence (macro level influence) and tactics for maintaining, spending, and exercising it (micro, or tweets and “conversations” themselves). Influence in conversation, for example, is an interesting one. Who do we trust? Who do we believe? And for what? I’d like to make a couple points about strategic influence before raising some hairy questions about tactics.

Nietzsche, the philosopher of power and force, taught us that there are two kinds of power: power projected and power attracted. It’s bi-directional. Similarly on twitter, there is influence projected and extended, and influence attracted. Jean Baudrillard got himself in trouble with feminism with a book called Seduction in which he described the projected force as male, and the seductive force as female. Without getting into a discourse on the politics of power, we can safely say that projective and seductive forces are at play in cultural influence and identity, and are similarly at play on twitter. Twitter is after all real people, and its lack of a clear social field only amplifies some of the strategies and tactics we use to differentiate ourselves from others.

There is plenty more to say on specific strategies and tactics, but I’m seriously running low on time and need to move on quickly to a final set of distinctions.

Personalities on twitter are distinct. Personality distinctions in social media explain differences in how people see themselves, how they see and relate to others, and in how they interact and engage in social activities. Personality types are not the distinctions made by market segmentation, but are psychologically-based insights into how people behave and why, grounded in interests and motives. While personalities can explain what’s really going on, however, they complicate the influence metrics that broadly apply to a population at large. That said, influence becomes much more interesting when we view it as something exercised and as a social distinction that motivates participation. Here generalized models of influence fall way short of explaining individual behaviors.

Without getting into details, for a separate post on “How we tweet” will be required, we can highlight a few personality distinctions relevant to qualifying different ways of exercising influence. I have grouped personality types into Self-oriented, Other-oriented, and Relational-oriented types. These correspond to monological, dialogical, and triangulating social behaviors. And to monadic, dyadic, and triadic social organization. Everything social can be built on isolates (one), dyads (pairs), and triads (three). The modes of talk, or in our case tweeting, would be Self-talk (telling, talking at); conversation (beginning with the other person, or replying); and triangulating (mediating a relationship between two other people).

(the following types are mentioned in my personality types presentation)

  • A pundit, for example, may exercise influence by building a reputation, following, and in largely talking at that audience
  • A buddy may exercise influence by being highly responsive, attentive, and reciprocating
  • A critic may exercise influence by aggregating and distributing content that belongs together (impersonally)
  • An expert may exercise influence by talking to a topic in which s/he is personally invested in demonstrating expertise
  • A harmonizer may exercise influence by becoming involved in small group and social runs on twitter, often event-related

and there are more….

The measurement of influence on twitter is fascinating, especially when we move from quantifiable influence to strategies, tactics, and practices of influence. At this point in twitter’s evolution as a social space, we’re going to want to map the social field and begin to locate and delineate social distinctions. It’s what we do when social spaces first appear and populate. But twitter is a highly individual social media tool, and an understanding of what we do with it, and how we each relate to its social organization, can really only be understood if we involve the psychology of behavior online. And that requires an understanding of how we see ourselves, others, and social activity overall, and will be expressed in the different ways that personal and social influence can be brought into play through the conversation and social action possibilities afforded by unstructured social tools and communities.

Twitter applications and extensions: a list

Note: This blog post belongs to a series on “status culture.” The posts examine status updates, facebook activity feeds, news feeds, twitter, microblogging, lifestreaming, and other social media applications and features belonging to conversation media. My approach will be user-centric as always, and tackle usability and social experience issues (human factors, interaction design, interface design) at the heart of social interaction design. But we will also use anthropology, sociology, psychology, communication and media theories. Perhaps even some film theory.
The converational trend in social networking sites and applications suggests that web 2.0 is rapidly developing into a social web that embraces talk (post IM, chat, and email) in front of new kinds of publics and peer groups. User generated content supplied to search engines is increasingly produced conversationally. Social media analytics tools provide PR and social media marketing with means to track and monitor conversations. Brands are interested in joining the conversation feeds, through influencers as well as their own twitter presence.
This changing landscape not only raises interesting issues for developers and applications (such as the many twitter third party apps), but for social practices emerging around them. So we will look also at design principles for conversation-based apps, cultural and social trends, marketing trends, and other examples of new forms of talk online.
These blog posts will vary in tenor, from quick reflections on experiences to more in-depth approaches to design methodology for conversational social media.


  • Adrian, great, though lengthy and overfull, post. I like the idea that social capital gains interest as it is spent, as opposed to when it is saved. That said, I'm sure there are ways of spending social capital that actually deplete its value (thinking like forcing through something unpopular).

    Also, while it's still too early in Twitter's life to view it as such, the social interactions aren't about the platform (as Twitter is simply a communication utility) but around the ways people connect and utilize the platform (you touched on this with your PeopleBrowser comment).

    I expect that groups/communities/affiliations will be both self-organizing (a la #tcot) and aided greatly by applications and websites such as Tweetworks.

    The ultimate unlevering in “social media” or social communication will occur with the implementation/availability of robust sentiment and linguistic analysis tools that automatically evaluate, sort, allocate, group, distribute and flag items in the social data streams.

  • Marc,

    Thanks for the comment. All my posts are overfull. I have the bad habit of theorizing while writing…

    I like to think that social capital is not expended when spent because I like to think that it requires the participation of others to spend it. I see your point though. Perhaps we could say that perceived capital diminishes if it is misused, but still retain the idea that spending social capital requires conversation/social participation.

    Agree w/ you on social interactions and that's in fact the gist of social interaction design. I like to say that the less structure there is in the design/architecture, the more is handled by social practices. And yes, self-organization will be facilitated by the evolution of tools and extensions. They are the means of production, if you will, of social networks and online talk.

    In terms of unlevering social media, I'm suspicious of the degree to which we can get meta out of talk. And I'm sure you'd agree that much of the meaning in online conversation and relationships exceeds the capabilities of automation and codification. Is it not possible that as meta and mining do make twitter richer and more differentiated, twitter and its users will lever up what they do with it? That they will go hand in hand?


  • Wow, amazing post. Not near enough people reference Jean Baudrillard when talking about influence on Twitter 🙂

    My name is Joe and I am one of the founders of Klout (thanks for the mention!). We look at influence not as a rank but as an aggregate. We measure influence between each node on the social graph. So of the people who follow me, there is a range of influence spanning between my close friends who share affinity groups and interact with my content to the random spam follower who will never see a single tweet I create. Figuring out how big of an audience a person has is the (relatively) easy part. To understand a person's influence you have to be able to analyze how deeply the connect with that audience. You can think of it as wide and shallow or narrow and deep in a lot of instances.

    The next challenge is figuring out what topics connect people. For companies looking to leverage influencers it's critical to understand what topics a person is specifically influential about. It's not as easy as saying “this person tweets about baseball 12% of the time”. You need to be able to look across the person's social graph and see if the people viewing those tweets even care about baseball.

    We are placing our bets right now that influence at the local level (and we consider affiliate groups here) is going to be the sweet spots for companies. Node optimization and understanding where you should plant the seeds of your message so it has the best chance to migrate across the web is a powerful weapon.

    Really looking forward to the rest of this series!

  • great — we're on the same page.

    I'd like to venture this thought. It's not mine, it's central to network and discourse analysis: influence is not a thing, but is a relation. It's a differential, a difference between nodes. So one can give it valence (say positive, negative), weighting, and directionality.

    One can contextualize this relational difference, say topically as you and marc both suggest, or as I suggest also by personality (don't pundits attract socializers, in a mutually beneficial and symbiotic way?).

    And one can situate influence within communication practices, within communities or tool-based social networks… this allows us to distinguish influence held by a person in one community, but perhaps not in another. And by style — a person is primarily talkative, or is primarily photo-uploading, etc.. .

    One can describe it as pro-jective or attractive (push or pull?)

    And one can observe it in communicative acts, some being more effective than others *in general* or *for a person* *within a culture/community/tool-based network.

    At the end of the day this is all hypothetical, and whatever works for targeting purposes (what's observable, measurable, and what leads to success metrics) is whatever will be good enough to a social media marketer, PR, advertiser, etc.

    The grail, to me, is in the motivational psychology of this — which is not required by marketing but which would be highly effective. For at the end of the day, it's about leveraging influence — and to do that one has to be able to engage with a user in the mode and style suitable to that user, and in a manner that embraces his/her interests and extends his/her competencies at social interaction online…

  • ps — the series on status culture started here, 6 wks ago:

    Status culture: facebook, twitter, and what they mean

    and all posts since can be seen at:

  • As Twitter grows in importance as a branding, customer relations, and marketing tool,
    influence is becoming a key component to determining

Leave a Reply