David picked up on this gem from Lloyd a few weeks ago and I have been meaning to post about it ever since. It gets to the heart of what I think people like Will are trying to do, but at the same time gives validity to the argument that there is in fact nothing that can be done to standardise metrics into social media â€“ a Google Page Rank of social media marketing effectiveness if you like.
To quote David/Lloyd:
1. There is no killer metric
2. Track anything possible to glean insight
3. Social media is not just about numbers
4. Itâ€™s all relative (focus on benchmarking and trends)
5. Measuring social media does not = ROI for social media
6. View monitoring social media as a social intelligence programme, involving the worldâ€™s biggest focus group.
The roots of my argument are points 3 and 6 – Social Media is not just about numbers – consider your social media activity as a tool to delivering your product/service to the world’s biggest focus group.
Ok, so you can measure the number of conversations, you can see how many comments have been made, unique visitors a blog may have had, item views etc., but they only tell you how busy something is – not what was said and the feelings and emotions and sentiment behind these numbers.
Only one thing in the social media space talks in numbers – computers.
Only one thing doesn’t have emotions and feelings – computers.
If we are to work to a metric, how on earth do we score an emotion? Do we mark simply positive and negative sentiment and if so, how do we teach a computer to understand that sentiment? Can it pick up sarcasm, sorrow and joy? I kind of know the answer to that one as I work with some guys whose toolset is simply awesome ;-) at understanding this, but which still has to be taught sentiment by human beings who in turn have to interpret a sentiment personally.
Do we give an arbitrary score of 10 to a positive comment and 1 to a negative one? How do we score a phrase of indeterminable sentiment such as â€œDave told me company X kept spamming him but the product is alright.â€ Here, the product seems to have a positive sentiment, but the company sentiment is a negative one. Maybe we score the comment as a neutral 5, but do you want even a good product from a company that spams you? Who makes that call â€“ can a machine?
But in arriving at a definitive metric for social media, it is essential to apply weightings to different social media – and again, I think this is nigh on impossible. Just consider your where you focus your time when performing outreach activity, if you do it 😉
Different social spaces and activities require different levels of involvement and one thing I ALWAYS find when planning a social media campaign is that no two are ever the same.
Weâ€™re all talking and sharing thoughts about the same topics, just not in the same places â€“ so how can we create a ranking based on the different locations a topic is being discussed?
For example, on what grounds do I assign a metric to conversations on flickr, that in some way means anything to another photo-sharing website such as photobucket (maybe Page Rank is a good start, let Google do the thinking on that one!).
Human participation can be the only way that we can understand the subtleties of the feelings of each site/community members and only the human can respond to those feelings appropriately.
Iâ€™d love to hear from people to gauge their opinion on all this, and I may even get my arse down to London to Measurement Camp next time 😉Read More