There we are! We have started on Social Media and run for some time. We know the ropes, we have met the right people and made the right noises. We are now ready to see where we stand and how effective we have been. And that is where we need to throw the dictionary away. Yes we do! There are many people who take the dictionary seriously and end up measuring the wrong the thing at the right place. So, away with it! And then start measuring!
Whenever I see clients or prospects trying to measure or analyse what they have been doing on SM and try to make course corrections to their campaign, I also see them going blindly looking at mention counts. Mention counts are the number of times that you find a dictionary match for your handle. They look at the number and go all teary eyed and proclaim that their campaigns are resounding successes.
They are forgetting the key factor – sentiment! And that is an unforgivable blunder! This means that you measured the wrong thing and are going to invest more into the wrong thing! Is that sensible at all? No!!
I should not get excited whenever I hear someone say “Shakthi”, for all I know, they might be saying “Shakthi is an ass!” and that is definitely not something that I would count as good. If they say “Shakthi is a cool chap”, it may be cool, but if what they are fully saying is “Shakthi is a cool chap when he is not drunk” then it is not all that cool.
See what I am saying! Throw the dictionary away and start looking at sentiment. If a lot of people think that I am drunk and offensive all the time, I need to fix that and not go around proclaiming that so many people talk about me.
This analogy holds good for brands and handles on social media as well. The need is to take a whole look at the conversation before understanding and analyzing what the sentiment is. So what we need are the right tools. And what are the right tools?
NLP is one important and critical tool, the natural language is a funny thing, and sarcasm is the toughest sentiment to measure online. Sometimes people actually say things that are good but mean the opposite. So the analytic tools should have a strong NLP core engine that understands these nuances.
Look at the following two sentences
“Woman, without her, man is useless!”
“Woman, without her man, is useless!”
Here all the words are the same and are in the exact same order. But the comma changes everything, and these two mean exactly the opposite of each other. Hence the next the requirement is to have a strong grammar engine that can deal with these in an effective way. Otherwise the analysis is useless
The third and probably the most important part of the analysis is provision for bias and extremes. Let us face it! Even if I am the king of the world, my brother-in-law will think that I am not a big deal. And even if I am not successful, my mom will think that I am a gem! These kind of biases exist for brands too and these have to be carefully removed from any sample that we derive for analysis and course correction.
So, go on. Throw the dictionary away and start analyzing Social Media! Cheers!