SEO Blog - Internet marketing news and views  

Targeting Social Media Influencers

Written by David Harry   
Friday, 21 March 2008 20:27

TIPIS - The Influential Persons Identification System

In an attempt to harness the power of Word of Mouth, or Viral Marketing – Microsoft has been looking at processes that seek to identify influencers within social networks to attain the greatest reach for marketing efforts. By reaching the core influencers in a social networks an advertiser;

“…creates the potential for exponential growth of the marketing message, as an individual passes on the message to his friends, those friends pass it on to their friends, and so on. Viral marking, so called because it spreads similar to a virus epidemic, exploits pre-existing social networks to produce exponential increases in brand awareness. Viral marketing harnesses the network effect of the Internet.”

To that end, let’s look at a recent Microsoft Patent titled; IDENTIFYING INFLUENTIAL PERSONS IN A SOCIAL NETWORK

Viral marketing has been identified as a relatively easy to implement, low cost and produces depth of reach and response rates for the message. With the proliferation of larger social networks on the internet there are more opportunities to be leveraged within these communities for marketers. It is surmised in this patent that by reaching the core influencers in these social networks one can optimize the message delivery.

As you might imagine the main advantage of such methodologies is that the advertiser can attain maximum reach with minimal cost. By ensuring the marketing message reaches top influencers the advertiser can obtain the greatest potential coverage in a shorter time frame. As you might imagine, this process makes for optimal targeting and delivery. 

Creating the targets

The goal of exercise is to identify influential persons within a social network. By identifying these users they can achieve optimal speed and reach for messages they wish to deliver. Once a list of ranked users is achieved it can be pruned by looking at the users – top down – and removing related nodes (friends) below that user until a more manageable list is produced. Once the process of removing friends of top influencers exhausts the totality of users in the network, a final list of top influencers is produced.

The initial ranking or seed list can be based on a variety of factors. One such methodology would be based on the number of total friends that hey have in the network. From there all of the users/friends in the top ranked user are removed from the list, then the same for the second ranked user and so on until the set limit of influential users is reached.

Further to that lists of influential users can be looked at for topical interests and demographic factors and also re-ranked accordingly. Concept being that you can now search among influential users in a social network and have a ranked list presented to you based on interests, topics and demographic factors (age, location etc..).

Creating the reports

For implementation an interested party can enter a variety of topical interest and demographic criteria to generate a ranked list of social media influencers. The report can be a list as well as being generated in a two-dimensional nodal link-graph. The (interactive) graph utilizes elements such as color coding and line thickness to show relational factors of the influencer in the center node to his connections throughout the network.

  1. An orange line to indicate a user-defined friend, meaning that the desired person defined another user in his social network as a friend;
  2. a green line to indicate a page view, meaning that a user in the desired person's social network has visited a web page or a blog of the desired person;
  3. a light blue line to indicate a blog comment, meaning that a user in the desired person's social network has commented on the blog;
  4. a yellow line to indicate an instant messenger (IM) chat has occurred between the desired person and a user in his social network;
  5. a dark blue line to indicate that a mixture of no less than two types of relationships as indicated by (a) through (d) have occurred between the desired person and a user in his social network.

This seems like a very interesting and sensible tool. I wonder if some other smart folks might take a stab at creating such a application? I know if I were more active in the world of social media marketing an application such as this would ROCK when it comes to tightly targeted viral campaigns.

All in all, a very interesting read.....

To all those having a holiday weekend – all the best from the Trail of the Fire Horse.



0 # T-Enterprise 2008-03-22 12:47
Excellent article - cheers!
Reply | Reply with quote | Quote
0 # Dave 2008-03-22 16:42
Xcellent comment - your welcome
Reply | Reply with quote | Quote
0 # Gab Goldenberg 2008-03-23 22:38
If you read The Influentials by Ed Keller and Jon Berry (Ed Berry and Jon Keller? can't remember...) they highlight people who are actively involved in their communities - especially politically - as being influentials. Their research is skewed towards influence in public policy, but I'd say your power users are the most likely to be influential. Just cuz someone has a lot of friends, DN mean their voice is particularly important.
Reply | Reply with quote | Quote
0 # Dave 2008-03-24 15:10
Hey G - well I am certain (and is implied in the patent) that you can start the seed list with any metric/demographic you choose. The example of the most friends is merely an example of the methodology. It could also be set on the persons with the most comments/conversations or a combination of both.

The 'friends' is merely an example. If you remember my Sphinn awards, there are many metrics one can employ or use in combinations to reach the percieved influencers on a given site. Depending on the nature of the activity on a site (or combination of) would likely change...
Reply | Reply with quote | Quote
0 # seo pixy 2008-03-24 16:05
Very good and interesting article, congratulations ! Viral marketing is definitely increasing its popularity and I think it will become huge in a while:-)
Reply | Reply with quote | Quote
0 # Malte Landwehr 2008-03-25 16:18
Soon we will all try to boost our "Social Media Influencer Rank (tm) Microsoft" at Facebook in order to be asked by advertisers to promote a product, website or Facebook applications.
Reply | Reply with quote | Quote
0 # Phoenix2life 2010-05-21 14:03
It is cool to read about "TIPIS". Though this blog post is little dated, I have first time seen clear algorithm of grading influential person in social arena. Wonderful to see neat flowchart to enhance understanding.

From point of view of SEO and Advertisement as well as social networking this algorithm makes statistical judgment. Iterative gauge of person's network grading the influential people in the network as well as their social engineering influence on their own followers. Interesting thought point.

Hope to see some prototype or online utility about this algorithm. May be it is out there. I just need to search it. :D)
Reply | Reply with quote | Quote

Add comment

Security code

Search the Site

SEO Training

Tools of the Trade


On Twitter

Follow me on Twitter

Site Designed by Verve Developments.