Social Media metrics; the problems beneath the surface.

To begin, explaining the predicament I see within Social Metrics is difficult. I believe the best platform to start with would be Twitter, not because Twitter is the most used Network or most relatable. But because the numbers are so easily available, my point becomes more relatable. It can get a little more complicated when you start to mess around with your own metrics to present the data given in a more understandable fashion. Social networking metrics are some of the most deceiving statistics on a ‘personal’ scale and on an ‘industrial’ scale. What I mean by personal are things like: Number of followers, number of Tweets, average number of tweets per day etc. Industrial relating to Networks as a whole, using figures from: Number of users on a network; active, fake, duplicates or the ‘reach’ a company perceives to have with it’s customers.


“Just because you can measure everything doesn’t mean that you should” – W. Edwards Deming


Tim Cook

When I say Statistics are deceiving I mean people misunderstand what the numbers within networking represent, in the sense that the numbers are not exactly as they seem to be. Take Tim Cook, CEO of Apple for example; he is regarded as the number one person wanted at a conference to discuss technology. He has 745k followers, yet as of the beginning of October 2014 he hadn’t tweeted more than 100 times. This is just a simple example of how, although (without a shadow of doubt) Tim Cook knows what he’s talking about when it comes to Technology, on Twitter he is nowhere near as influential he can ‘potentially’ be to his followers. This being said he may have inspired the world with each and every one of the one hundred tweets sent out. But in the grand scheme of things, we can make the judgement call that his account is more inactive rather than a man of few yet incredible words.

Matt MullenwegThen on the other hand look at someone like Matt Mulenweg, Founder of WordPress. A little less mainstream by Tim Cook’s standards, reflected by his follower count being less than ten percent of Cook’s. But what isn’t as clear as a High Follower count and a low Tweeting count is the inverse. If you overlook the fact that somebodies ‘Tweet:Follower’ count is high; 6:1 in this case, you can highly underestimate how much of an asset the profile could be. Obviously this isn’t always the case; an 18-year-old girl could be in the same situation, so it takes some light research into the tweets to get the full picture. But generally speaking when you find yourself looking at CEOs of huge companies the more valuable account is going to be the one where their area of expertise is most discussed.


“Social media is one area of business where you don’t have to outspend your competitors in order to beat them” – Hal Stokes


100 Tech Speakers

Although this graphic looks relatively complicated, it’s not difficult to get your head around. I have designed it so that you can get an idea for the relationship between the number of followers between a group of people who are globally renowned for there distinct knowledge of their area (in this case technology), and where they lie on the scale of how much they are desired to talk at large conferences. If there was a perfect relationship in the way we’d expect, we’d see a line of best fit passing from the bottom left of the graph to the top right. This would show the more followers a person has; the more influential they are. If you look at the top 50 or so you can faintly see the weak positive relationship where this line would be. But the correlation is very weak and with examples like: Jeff Weiner, Peter Thiel and Mark Andreesen showing the exact opposite that we’d expect at that end of the spectrum. There is very little evidence to conclude positively that there is a relationship.

Where would we go from here to fix the identified problem?

In terms of Twitter, I believe the best place to start is not to overlook the ‘bog standard’ statistics, but use them in a fashion that is going to benefit the analysis of the account showing the characteristics of it’s specifics. From here I think we need to categorise the different elements of what we would like to break down the account into;

Reach

This would look into the account by who is actually influenced and affected by the content. So the variables that are going to be used are the ones that reflect a strong interest towards the tweets or account. Being followed isn’t enough, what does it take to get noticed?

Follower interest:

No matter how active an account is or how many followers they have, when analysing the bare statistics you aren’t considering how many of those tweets actually get noticed. To find this we would need a variable which would take all of this into consideration; how many [tweets + favourites + replies] you get per tweet. This will give you a value of the amount of attention per tweet.

Follower Value:

This ties in with the next point of Influential Potential. How much is a follower worth on Twitter? This could be an entire post in itself, going into business economics about how you can exchange your followers for revenue. In the sense of reach though, it’s a question of how to increase the proportion of the highly valued followers. By highly valued I mean those that have an effect on the growth of your network.

Influential Potential

In an ideal world how successful can this account actually be? How many people can this person inspire if they were to be suited to each follower they had perfectly. All we want to know from this is the maximum possible influence an account can have so to understand it’s limits. The values I think would be most suited to finding this evaluation would be;

  • Follower count, (obviously)
  • Number of inactive accounts per thousand: Around 1,000,000,000 accounts and approximately 284,000,000 active monthly users, this can be a big issue in any data gathering scenario. Therefore there are approximately 716 inactive users per thousand. (Source- expandedramblings.com)
  • Line of work the subject is in has a huge influence. If the person is a frequent face in the media who’s opinion is valued and respected, they are clearly going to have a bigger audience that takes notice. Whereas somebody just as famous, but less opinionated would be a totally different scenario. Mixing two completely different subjects together and analysing them both is a huge ‘no-no’.

To summarise, this post was intended to give a brief insight into not only how difficult it is to obtain accurate statistics for Social Networks in general. But also to express that when viewing statistics to do with Networks, or any other area for that matter, to never take them at face value. It is so easy to mislead people without strictly lying.

Big Data Context

Advertisements

About Lewis Waite

This Blog is part of a project I am currently working on to essentially, analyse the Social Networks of today. And by the end predict where we might find ourself in the future.
This entry was posted in Data analysis, Social Media, Statistics and tagged , , , , . Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s