Are Klout, Kred, PeerIndex, SocialBro and Others a Waste of Time? Possibly.
Social influence continues to be the subject of much debate, and it appears that there are more indices appearing. Klout is probably the most well known but has been controversial when algorithm changes have been made. The focus of debate has always surrounded the resultant Klout score, usually led by those whose scores have declined.
Late last year this prompted me to write a blog post called Social Media Measurement. The purpose of the article was to question whether there was any link between ‘influence’ and a true business benefit. My conclusion then was that the best way to measure the impact of social activity was to analyze the results using Google Analytics.
Since that article I have given little further thought to the value, or validity even, of an influence score. Klout’s algorithm has changed again, Kred seems to be gaining support, and PeerIndex, which I had always thought was reasonably stable as a measure, has also made changes to its algorithm.
It is PeerIndex that has prompted me to write this article as I recently made an interesting discovery, which suggests to me that ‘influence,’ as measured by PeerIndex, is potentially of far less value than I had imagined.
Here’s how I approached the problem. I have about 4,700 followers on Twitter, and I follow about 3,700. From time to time I do a little housekeeping, in an effort to ensure that those I follow, in particular, are entirely relevant to me.
On this occasion I decided to look at SocialBro to see if this could help me. The first thing I spotted was that there were 1,300 people who followed me, that I did not follow back. As such that doesn’t trouble me, but I thought it worthwhile examining this to see if there were some people amongst this number that I really should be following.
I quickly found 5 people who tweeted on relevant topics, and were active, so I followed them back. Just 1,295 to go! Needing to refine my process I started using the very clever filters in SocialBro, and made some interesting discoveries. I do not intend to bore you further with the process, but detail my key findings below:
- 45% had not tweeted for 6 months
- 28% had not tweeted for 1 year
- Nearly 5% had not tweeted for 2 years
- Of the 45% (586 people) who had not tweeted for 6 months: 30% had an PeerIndex influence score of 50 plus and 12% had a score of 65 plus
- Of the 28% (365 people) who had not tweeted for 1 year: 32% had an influence score of 50 plus and; 13% had a score of 65 plus
The PeerIndex score contains elements for Audience, Activity and Authority and, in some instances, these are all zero for a given person, and yet they merited a score of 65 plus on PeerIndex. Conversely I have found scores where a high Activity score had been allocated to someone who hadn’t tweeted for over a year.
I fully accept that Twitter is not the only social platform that is factored into PeerIndex’s score. As far as I can determine however, Twitter is the only social channel you can’t remove from PeerIndex, and I assume its use must therefore add to social influence.
I imagine that there is an entirely reasonable explanation for my findings, and I would be pleased to hear it. On the surface however what I have discovered appears totally bizarre, and suggests that there is little credibility in this (and possibly other) indices. My conclusion of last year was that there is a need to link social influence to measurable business benefit. I still believe this is correct in principle, but the measurement of social influence needs to be demonstrably better than this if it is to have any relevancy at all.
Peter Rees is the Founder of WSI Digital Marketing Solutions and can be followed on Twitter @WSIeBusiness.