What is influence in social media?

[Elias Bizannes]

Nick Hac posted an entry today on Australia's top influencer's per Twitter.

http://www.shiftedpixels.com.au/blog/2008/10/australias-top-50-twitter-influencers.html?ext-ref=comm-sub-email

This is not the first time I've seen an effort like this - whether it be about online influence, Australians, etc - but what struck me was Nick's formula of ranking people.

TwitterRank = (Following / 10) + (Followers) + (Updates/100)

This is not a trivial thing, and I think it merits discussion. After all, "word of mouth" or WOM is becoming a dominant view on new advertising on the internetz. Being able to identify your customers on their influence can literally translate into dollars. So whilst Twitter is but one minor, niche services in the general world, it's a good case study to build on.

First of all, I don't see how people you "follow" constitutes influence. After all, isn't influence the power you have to change other peoples actions, without using force of any kind? Number of followers make perfect sense for the same reason, but shouldn't this be quantified by who exactly follows you?

Secondly, how should updates get recognised? If you update too much, you lose followers - so that's a real metric of the loss of influence you have based on updates (and so needs no adjusting). But if you update too little, you may also lose followers...either way, number of updates doesn't reflect anything about your influence as it's end-value is determined by the net followers you have.

Thirdly, what about engagement? In my eyes, if someone is glued to their seat when you speak - you are influential. It's no surprise that those who say the least tend to hold the most influence in the world - because they recognise less is more. In forums, I have seen self-proclaimed leaders of industry groups crap on, and after garbling their verbal diarrhoea, I've simply attribute lower value of them in my eyes. Or rather, next time I read about what they write, I'm less likely to read it, at least word for word.

So if someone re-tweets you, is that a fair measure because it shows they take in your updates? What about the amount of people that reply to you - is that engagement? If people get SMS'd when you tweet, despite the limitation to do it on everyone (cost, sanity, etc), surely that's got to show for something?

Fourth, does time come into this as a dimension? If I use Twitter at 2am in the morning, the American's see it, but not the Aussies. Therefore my influence in constrained to the people in America. If I update once a month, I also have less interaction with my followers, and therefore, can lose influence in the fact I don't have a relationship with them - as you can probably reflect in your own life, intimacy is a big factor in any relationship - the less frequent you speak to a friend, the most out of touch you become.

I've done a bit of work at my firm on determining how knowledge gets transfered in a big company, using McKinsey research. Influencers tend to be the best networkers - not just the amount of nodes connected, but the strength of the connected between notes.

Anyway, just some thoughts. It's an interesting subject, and despite ego pandering, is something we should understand and reliably be able to measure because it translates into real impacts on business.

--
Elias Bizannes
http://liako.biz

[Alex North]

On Mon, Oct 20, 2008 at 1:20 PM, Elias Bizannes <elias.b...@gmail.com> wrote:

TwitterRank = (Following / 10) + (Followers) + (Updates/100)

This is not a trivial thing, and I think it merits discussion. After all, "word of mouth" or WOM is becoming a dominant view on new advertising on the internetz. Being able to identify your customers on their influence can literally translate into dollars. So whilst Twitter is but one minor, niche services in the general world, it's a good case study to build on.

First of all, I don't see how people you "follow" constitutes influence. After all, isn't influence the power you have to change other peoples actions, without using force of any kind? Number of followers make perfect sense for the same reason, but shouldn't this be quantified by who exactly follows you?

Perhaps a formula that took into account the influence of people who follow you would be an improvement, much like Page Rank for web pages.

[Phil Morle]

Elias

I think your point about using the time of a tweet to express which
geography is being influenced is extremely important on live platforms
such as Twitter.

I also think that an influence algorithm needs to include @reply data
as this suggests the extent to which others have read and felt
compelled to respond.

Cheers

Phil

Elias Bizannes wrote:
> Nick Hac posted an entry today on Australia's top influencer's per Twitter.
> http://www.shiftedpixels.com.au/blog/2008/10/australias-top-50-twitter-influencers.html?ext-ref=comm-sub-email
>
> This is not the first time I've seen an effort like this - whether it be
> about online influence, Australians, etc - but what struck me was Nick's
> formula of ranking people.
>

> *TwitterRank *= (*Following */ 10) + (*Followers*) + (*Updates*/100)