Showing results for 
Show  only  | Search instead for 
Did you mean: 

The 6 Factors of Social Media Influence: Influence Analytics 1

Lithium Alumni (Retired)

michaelwu.jpg Michael Wu, Ph.D. is Lithium's Principal Scientist of Analytics, digging into the complex dynamics of social interaction and online communities.


He's a regular blogger on the Lithosphere and previously wrote in the Analytic Science blog.


You can follow him on Twitter at mich8elwu.



This post kicks off a multi-post miniseries on the topic of influencers: how to find them, engage them, and collaborate with them in word-of-mouth (WOM) marketing programs.


Influence marketing today is in a state of experimentation that scientists call the pre-paradigm phase or exploratory phase. During this phase, everyone is trying different approaches based on experience. There are incomplete theories about why some approaches work and others fail, but there is no underlying fundamental principle that explains everything. My approach in this series is to see if we can gain a deeper understanding by analyzing the process of influence from a data analytics perspective, using a simplified model of social media influence.


A Simplified Model of Social Media Influence:

Influence involves two entities, which I will refer to as influencer and target.


1. The influencer's power to influence depends on two factors:

a. Credibility: The influencer's expertise in a specific domain of knowledge.

Please note: There is no such thing as a universal influencer, because no one can possibly be influential in all domains. The best that anyone can hope for is an influencer in a specific domain of knowledge

b. Bandwidth: The influencer's ability to transmit his expert knowledge through a social media channel.

Please note: Active influencers in one channel may not even be present on another channel. So influencers are not only specific to a domain of knowledge, they are specific to social media channels

2. The target's likelihood to be influenced by a specific influencer depends on four factors:

a. Relevance (the right information): How closely the target's information needs coincide with the influencer's expertise. If the information provided by the influencer is not relevant, then it is just spam to the target and will be ignored.

b. Timing (the right time): The ability of the influencer to deliver his expert knowledge to the target at the time when the target needed it. There is only a small time window along the decision trajectory when the target can be influenced. Outside this golden window, even relevant content will be treated as spam because there is no temporal relevance.

c. Alignment (the right place): The amount of channel overlap between the target and the influencer. If the target is on a different social media channel, then the influencer's information either take too long or never reach the target.

d. Confidence (the right person): How much the target trusts the influencer with respect to his information needs. Even if the influencer is credible, the target must have confidence in him. Without trust, any information from the influencer will be downgraded by the target.

chain_links2_resize.jpgThis model is very general, and it is intended to be applicable to any social media channel. However, it is by no means complete. I just like to use the principle of Occam's razor and start with a simple model that is consistent with the data out there and see how much it explains. We can always add to the model if it proves to be insufficient. As Albert Einstein once said, "Everything should be made as simple as possible, but not simpler."


Please note that a lot of attention has been focused on influencers, but very little has focused on their targets. Although it is easier to work with the influencers, we must not forget that it is the targets that we want ultimately. I hope this simple model will help you think about social influence from a more balanced perspective, so that even when we are looking for the influencers and working with them, we still have the targets in mind.


Now that you know the basics of how social media influence works, it should not be difficult to diagnose the success or failure of a social media campaign, at least from a data analytics perspective. As shown in the photo above, any broken link between the influencer and the target is enough to break the chain and stall the whole influence process. Next time, I will show you how to take the first step of WOM/influencer marketing: find the influencers.



Related Blogs


Not applicable

This makes good sense to me, emphasizing from the target's point of view that influencers need to have credibility and to deliver relevant information at the right time and place, for the target/s to take action. What you've written about influencers also accords what what I've read recently, that it's not necessarily the size of the influencer's network, but their passion which is key (which I'm guessing may relate to their perceived credibility). Bearing these points in mind I've asked in a recent blog as to whether community managers should themselves be regarded as key influencers (& therefore a potential revenue driver), despite the risks involved in turning them into "glorified marketeers" if not done correctly.

Lithium Alumni (Retired)

Hello Stuart,


Thanks for dropping by and leaving a comment to reaffirm the validity of my model.


I definitely agree that a user's passion is very important, and their passion probably result in a higher bandwidth in some social media channels. This could which may increase their visibility and therefore increase their perceived credibility, which is what I am calling "confidence" in this article. But this may not be the true credibility of the influencer.


The issue of "turning community members into a glorified maketeer" is a touchy one and must be handled carefully. This is a topic that I will cover in a later post of this blog miniseries. So stay tuned!


Not applicable

Hi Michael,


Thanks, I like the distinction between passion/confidence & overall credibility.


I look forward to your post on the delicate subject of using community managers to influence members around their purchase of products & services.







Ni hao Mike, (thanks to Nick Jr.   lol)

What a great start...!

I'm a new fan ... but I'm here for the duration of the miniseries (at least) ...  (grin)

* * *

Of the introductory 4 factors of the targets, I agree that confidence in the influencer is very significant ... part of that confidence may also stem from similar experiences in the lives of both. I believe given two influencers with similar advice to a single target, one of the discriminators the target might use is how similar each one of the influencers' experiences mirror the target's ... we tend to trust those who have lived in the same part of the country or had similar rites-of-passage experiences, no?

* * *

This sort of analysis simply fascinates me ... as well as discovering all the significant nuances that are commonly overlooked that influence people to behave the way they do. . . and especially online.

Keep it up ... you have a new loyal fan...


Until that time...  Earl J.

Lithium Alumni (Retired)

Ni hao EarlJ


Thank you for the nice comment. I'm so glad to hear that my analysis resonated with you.


What you said is definitely true. At then end of this miniseries, we will have to deal with the confidence of the target. That is the last bottleneck that we must overcome in order to be successful in influencing the target. And depending on what data you have, you can potentially boost the confidence of your target, by making the target aware of the similarity between him and the influencer.


This is precisely what I meant when I said that even though we will be working with influencer's data, we must not forget the target. Even though we are finding influencers, we must also make sure that we find one that is similar to the target in some ways. At the end, there is a tradeoff among the 6 factors that we will need to optimize.


I look forward to hear your comments on my future poosts. Thanks again for dropping by.


Not applicable

Michael - 

I'm very much enjoying the first in your series - which focuses on the dialectic of Social Media.   I agree with all of your model's components.   For me, it raises two elements and a question.  


First, I think about translation to the medium.  In order for bandwidth and relevance to be effective, the communicator needs to be able to translate their concepts into the channel's available format.  A video or a 140 characters in Twitter have very different contextual carrying capacity


Second, the timing also depends on the channel.  Twitter only provides the last 5 days of data, but blog posts from 2007 are still driving traffic to my web site.   So again, the nature of the medium impacts the factors you've identified ... and different people's communication styles are suited to different media.  


Finally, my question is on metrics.  You mention analytics at the end, but some good examples of tools that measure such factors as Credibitlity, bandwidth, timing, influence, etc. would be very helpful. 


Looking forward to your other posts on this topic!




Lithium Alumni (Retired)

Hello Skip Shuda,


Thanks for the thoughtful and inquisitive comment. There is definitey differences in expressive power for different social media channel. That is why bandwidth is defined "The influencer's ability to transmit his expert knowledge through a social media channel" in the article and it depends on the particular channel, or what you called medium in your response. This is also why I included the alignment of channel as the 5th factor. Very good observation 🙂


As you mention, timing is also channel dependent. I would imagine that if you need to do analytics on twitter data, you probably find some ways to save and aggregate the tweets first. But if you do it directly from twitter search, then the temporal sensitivity (timing) would be very narrow. However, twitter follower/following data are persistent over a much longer time range, and that seem to be a metric that lot of people like to user to determine influence, where in fact they can only fine high bandwidth users that may be, not credible in the domain of interest, irrelevant and have poor timing sensitivity.


For your question, the next article in this miniseries is already out, and it explains the various types of metrics or data for determining a influencer's bandwidth and credibility.

Finding the Influencers: Influence Analytics 2


For timing and relevance, it is more difficult to simply have a metric because it depends on the target as well. That means we cannot derive a metric for the relevance or timing of an influencer. We can only have such metric for a influencer and a particular target in mind. However, in the following post (part 3) of this miniseries, there are data manipuations that we can perform to accomplish relevance and good timing. So stay tuned!


Not applicable

Hello, Mike -


Another relevant  facet of the influencer is the breadth of their knowledge and the breadth of their audience/targets.   A well known celebrity such as Stephen Fry doesn't have any particular credibility on specific topics, but is followed because he is so generally intelligent, humorous, and well spoken.   A number of TV news broadcasters carry an outsized ability to influence their viewers because of the nature of the broadcast channel.   Again, they're not packing credibility in any specific topic, but are given credence due to the frequency and nature of their communications to their audiences.


I suspect these factors can be fundamental characteristics of an influencer taxonomy as well.


On to the next two articles,



Lithium Alumni (Retired)

Hello Dave,


The breadth of knowledge and breadth of their audience are indeed important, but they are more suited for influence that are not social media (ie, offline influence). We can certainly put them in a social media context, but if that is the case, then I would argue that both of the factors you mentioned are subsummed in the 6 factors already.


1. Breadth of audience is essentially a social equity metric (see Finding the Influencers), and it is a measure of his bandwidth. Bandwidth as defined in this article would include his network, audience, reach, etc.


2. The breadth of knowledge, the fact that he is generally intelligent, humorous, well spoken, etc., essentially gives the target more confidence, which is also a factor that we will cover this week. In fact Your example of Stephen Fry is an example of a respected authority figure that the targets trust. This will be discusee in the last article of this miniseries.


The next 2 articles of this miniseries is already out. The last one will be posted sometime later this week.


Thanks again for the comment.



Not applicable

There are many social networking websites where we  can share our blog and get traffic for the blog. Now a days social networking websites are very famous for the business also.Here you mention some good information about social media influence..Thank you for sharing .