SEO Blog - Internet marketing news and views  

Twitter and the Agents of influence

Written by Terry Van Horne   
Wednesday, 07 October 2009 14:10

(by Terry VanHorne)

Relationship marketing and Social Media Success

First off let me say that I have been around the web for a long time and can tell you that whenever something new comes along we get stories like; Loosing to the social web visualized. Sorry, but with that story, I felt more then just a little smoke being blown up a hole I'd rather not experience.

Yes Social Media is taking off but to use the sites they did in that post and not mention that the greatest recession since the great depression didn't play a role in the dramatic traffic decreases is to see a world no bigger than your Social Media cubicle. A few of the sites listed like the newspapers are likely very affected.

Don't screw up the data

I doubt people are choosing to buy from the Dell Facebook page instead of the Dell website. To the author... I strongly urge you to pull your head out of your a$$, take a look around the world beyond yourFacebook page and see the financial turmoil that may be just as responsible and quite likely more responsible for the down trend on those graphs. Also will you take it back in December when those store graphs are reversed?

Social Media people you haven't discovered one new technique that hasn't been around Internet Marketing for years! Twitter is just improved IRC if you really think about it.


The net, and many of the services like search and stores that these Social Media twits say Social Networks will replace is unfathomable. Sounds like the guys back in the day saying in 4 years we'll all be surfing the net on our phone... 9 years later and it's still a hope of AT&T and Apple and not really that much closer in NA. Believe it or not there are people who bought smart phones for reasons other than to keep their Facebook page updated.




The data

Thanks to Dave for providing the onsite data and help in monitoring the traffic from the first post. I think Dave was just reassuring me because I was very hesitant to post; How I became a Social Convert. The subsequent article Measuring Social Media Success discussed some of the problems with; Measuring Social Media Success.


Searching Twitter for "social media convert" without quotes revealed the following using unique URI's as "source": 152 Total Results

  1. Top 3 Individuals accounted for 82 results
  2. Top 4 Individuals accounted for 90 results
  3. 62 Individuals accounted for the remainder of results.


Those numbers and a little research into some evolving ad platforms started me on the path to realizing that analyzing my Tweet activity and in particular the "influence" of the Agent over followers would be important.



Assumption and Principles of Response Marketing:


Some will wonder why I used the URI to identify sources. Simply there were very few @ or RT in the Tweets. The URI was by far the best identifier for source of the "influence". To determine the source for a metric is tough in an environment where user input plays such a large role in the data being used for measurement of the activity.

IMO, that may be why the RT is being re-invented by the development team. This may well be a piece of the monetization puzzle. The recent capital injection was likely not provided without some indication of how Twitter w/could be monetized.


Only 23 of the tweets, (15%) reviewed included the author in a RT, @ or Via. This is not a good indicator for me as the author. That said, when I checked my shortener I found 44 clicks. By far one "influencer" received the majority of @reply included in the Tweet and there was no use of Via or other indicator of other sources.

Worth noting is that the "influencer" with the highest reuse of the URI, did not receive a single '@' or 'Via' citation (i.e. without using the URI as source the person with the most followers would have been missed entirely were only @ and Via used as means to determining sources in a study).



Determining sources of influence in relationship marketing


The importance of determining the sources of influence in relationship marketing is high as you attempt to leverage your relationship with the influencers. The more influence the person has the more leverage you get in the community from the relationship.


Of the four with the most signature URIs the Followers and Followed were:

  1. 6220 Following ~ 5470 Followers - 38 RTs ~2 accounts
  2. 2660 Following ~ 3508 Followers - 22 RTs ~2 accounts
  3. 7,233 Following ~ 7,343 Followers - 22 RTs
  4. 123 Following ~ 135 Followers - 8 RTs


There are also archives and social bookmarking sites like Tweetmeme, Delicious etc that could account for the lag/evenup. This indicates that click analysis **could be** useful in determining influence especially if the RT level is high over a short time period after the influencer tweets. This would indicate the degree of influence on the Follower by the Agent of Influence. Otherwise there would be equal pageviews and RTs of the content.

An interesting phenomena that Dave, (@TheGypsy), and I noticed watching the page analytics is that the the RTs were in excess of actual pageviews, however, by the end of the day the numbers had pretty much "evened out" as a RT/pageview ratio. Which indicates that in this case the RT is given without actually reviewing the content.


This would indicate that the followers of the three individuals largely responsible for the bulk of the retweets, first RT'ed the content and then later read it or the buzz resulted in curious clicks and the article was never read by those who RTed. Either way, and interesting phenomenon.


Social Media influencers



The Power of Influencers


Another indication of influence is that the URI in the message, if it is RTed or highly influenced by the Followed Broadcaster the URI, was left intact and was in fact the way I tracked the RT's to specific user because often they were also not included in the retweet. In fact looking at the timeline you see the roughly 50% single URI were all mostly later in the day with the highest levels of activity almost directly after the three major "Agents of Influence" posted their tweets.

One of the Agents was Dave and I'm humbled the Dojo supported the article the way they did. The principle Agent was known and assumed to be either following Dave on Twitter or subscribed to his blog.

One metric I see discussed occasionally is the value of the RT to the authors/original tweeter. Sorry, but an astounding number of RT'ed messages removed any indication of source and author of the content. As I am a ... cough cough... author/writer wannabe this was a big downer.


Personally, I only remove authors/sources as a last resort and often will not add further comment to the RT in order to do so. Rremoving author/source is a little like taking ownership and for a "social media" topic that seems more than a little "unsocial" or... a sad reflection of the new internet mentality i.e. netiquette is dead... they took it out behind the barn and it never came back.

I came across that as I reviewed the mentions in the twitter search. One thing I have noticed is that the old methods of rolling up my sleeves and reviewing raw data provided significantly more "useable/actionable data" than the the Social Measurement tools. Currently Topsy is a favorite for measuring the overall buzz of old metrics such as clicks and the new click the Retweet.



Influence Measurement Tools


Our fav SOSG, CJ, (@missmcj ) recently Tweeted a great resource, and while reading that I found this fantastic discussion of an algorithm to measure influence or as I refer to them Influence Metrics (see TunkRank influence measurement in action). Watching the research on the human side of the "social phenomena" will be interesting. I don't believe it is just the volume of followers and your message is missed due to large numbers of tweets to review.

The human element or influence in a tweet is harder to measure. However, if you assume a few things about human nature like trust in the value of the information being shared, the degree of edits to original "elements of twitter" of tweet RT, @, via and URI. You could also assume by the time lapsed between broadcast and RT to be a measure of influence as people with a higher degree of influence over their followers will get a quicker response.

Another action of the follower that is possibly an indication of trust and or respect is the edit of the URI.


Twinfluence is calculated on the number of followers and doesn't really measure influence in any meaningful way. It is one way to measure influence but it is more about reach than influence. I want to measure and determine the "direct influence" of an Agent not reach. I'd prefer 1 follower who had a hundred followers and influenced 10 people to take an action over an Agent with 5000 who influenced 30.

Twinfluence misses the mark

As an example I have an account where I am developing a local stream with news, weather sports, local events etc. To that I've followed a number of the local tweeps looking for individuals with entertainment and information value.



The Flaw in Follower Counts


This has been an awesome experience as I have discovered an amazing community of internet developers, marketers and entertainer/bloggers. To say that until recently my experience locally has been abysmal, is an understatement. My experience has only been with dinosaurs and the uninformed.

I was recently honored by a #FollowFriday from a person with a very large following. The recommendation resulted in a small spike in new followers. However, when I first started using Twitter a tweet from Dave and Rob Watts resulted in many more new followers, at that time I had half as many followers myself.


I offer that information as indication that the number of followers possibly provides a flawed measurement of influence, assuming, influence is determined to be the ability to sway opinion and more trust-ability to influence actions by followers.



How do we Measure Influence?


Identifying "influence" is a part of my relationship marketing strategy. I assume that I'm going to bring visibility to that person, and in doing so leverage the activity so that the person with the most ability to influence is given the most attention (i.e. many years of marketing has taught me the ability to influence decisions is a skill that can be taught but is often just as much personal charisma as anything) Therefore teaching people how to influence can be a less effective method then just hiring people with character.You either have it or you don't.

The Influential's, for me, is the definitive paper on measuring Influence on Twitter. Followers are metrics of reach not a measurement of influence. Measurement of influence seems difficult if you are talking the whole Social Web, however, there are features of Twitter that are true indicators of Influence. The use of ReTweet, @reply and attribution in tweets provide metrics for a much better system to measure influence then the follower audience numbers etc. IMO, the reuse of URI is also a citation/attribution worth consideration.


My opinion on the methods for identifying sources is that using @ and Via citations as the only identifier is a flawed strategy due in part to the length of a tweet and the RT comment on the information being pointed to. This could be why the RT is being changed by the twitter team. Using the URI is also a show of respect for the source.



About the author; Terry is an old school SEO/IM geek that works out of International Website Builders and the founder of - You can also hook up with him on Twitter. It seems he's become a regular here on the Trail... so I wanted to, once more, give a great big THANKs to T for hanging out and sharing his experiences.


Be sure to check out the first part of this series; Measuring social media success


Search the Site

SEO Training

Tools of the Trade


On Twitter

Follow me on Twitter

Site Designed by Verve Developments.