What, essentially, is social? At the risk of simplification, it is the felicity with which day-to-day conversations – word of mouth, if you will – converts to digital or mobile text, video and voice transfers. Slice word of mouth and you will get all kinds of content – gossip, rumor, debate, discussions, family chatter, party talk, dinner table chatter, toilet graffiti, jokes and bathroom humor, doggerels and bawdy lyrics. Social is textualization, imagification and videography of all of this. The only difference is the dynamics of propagation and the scale of reach, which, on the surface, seems to be a different kettle of fish, but it really isn’t.
Which brings us back to the core question: How does one, first, map objectives and outcomes in using social media and how does one proceed to measure that? Off late, it has been amusing to see much of the ‘measurements’ getting propagated by agencies and clients desperately seeking to score ‘numbers’.
Here are the five biggest bull shitters of social media measurements:
a) Facebook ‘likes’ or ‘followers’: There is a whole industry of fraud that has been built around large numbers.
b) Twitter ‘trending’: Ok, no secret, it takes a bunch of 10 odd guys with a ‘follower’ base of about 5,000 to retweet to get a ‘trend’.
c) ‘Reach’: A hilarious one – This one is about the total number of people who apparently saw your tweet. So, if you happen to follow a person – among 400 others you follow – and that person happened to have a follower base of, say 1000, then his tweet has reached a 1000. Problem is, you aren’t really sitting and watching his every tweet. And neither are 990 of his followers. His real reach is possibly closer to 10.
d) Impressions: Which takes the above absurdity further. If he tweets on the same subject twice, the ‘impressions’ are 1000 x 2. In a recent situation, we were looking at about 2000 tweets and a wholly hysterical 36 million reach.
e) Tonality: Either someone is sitting and scoring out each piece of content or there is an algorithm that has marked the following post/tweet as positive or negative Churchill’s response to his arch enemy Clement Attlee being rushed to a hospital after he collapsed, “I hope it is nothing trivial”.
Run that through any algorithm and see how that one is scored.
So, how do you real measure Social? First, no single engine gives you everything. The best way is to develop the sense of how communication is flowing. Here are a few measures that, in combination, give an accurate sense of how the communication is moving… except that, remember, that unlike offline the flow is far more dynamic and unstable.
Let’s take a few measurement tools that help move towards an estimation of RoI.
The first is to see how much of the conversation that is being generated is clustering around the topics that are set out as key ecommerce drivers. For example, at a given point of time, if the clusters map is a bit too vacant while there are plenty of single-line threads; it simply means that the audience is reacting to noise and not to a message. For example, in this particular time frame, we see for Flipkart a very large number of singletons while the clusters don’t have a large business, product or promotion impress.
A second measure would be to look at the word clouds across the spectrum of predisposition.
For example, in the instance above, the word cloud on the positive side tends to show a huge clutter of similar ecommerce brands. Clearly, in the tussle for brand mind space, there doesn’t seem to be any clear winners.
One key measure for an ecommerce brand has to be both a dominant brand in the space but also creating unique properties that others find different to duplicate. Measuring how those sub-brands are trending would give direct connects to product movement. Too often, the mega brand stands head to head with another mega brand. While it may well drive traffic, there is just too little differentiation happening. It is in that sense, a bit of a Catch 22. You need a lot of noise to stay top of mind, at the same time, all the noise tends to knock off the differentiation. A good example is the telecom brands, which today reflect very little differentiation because they have spent just too much time building noise volume over specificity.
A third measure is to see where the conversations are emerging. The more they emerge from third parties or getting loaded in any one direction, the more the brand needs to pay attention to the conversations happening on those platforms.
When the mother lodes of the conversation are mostly around owned platforms (as it was for Flipkart), this reflects a situation where the community “talking up” the brand is getting limited. In a sense, tthis reflect the duality of Social – the more people talk on your platform, the more you are in charge of the conversation; on the other hand, the less are the platforms talking about you, the slower is your brand movement in the wider audiences. A balanced web is a good measure of how the ecommerce brand is working the bandwidth.
A similar measure comes from tracking which are the channels that are generating the maximum chatter for the brand and how much of it is original. Ecommerce brands should work out on each front how a particular form of chatter benefits – or not – its particular brand promotion or context.
Often, geographic spills may be of little use. This often happens when a social campaign ends up having a global appeal and becomes a trendy hashtag to an audience that is a little too spread out. Watching for demographics is also important. Too many men getting involved in, say, a lingerie promotion may not really be adding up to sales and may well point towards take-off points, which are counter-productive.
Finally, and most importantly, watch those volumes. However, do not mix them up with reach or impressions. The best measure of social is the actual – how many people spent those Nano seconds posting, reposting, sharing, liking or sharing. The rest is theoretical. Sharp departures in trend lines can be dangerous and need to be watched carefully.
In conclusion, the key to effectively measure social media ROI is to keep the measurement in real time as it operates more as a seismograph than a compiled report. Avoid compiling data – yesterday’s tweets plus today’s tweets equals’ zilch.
In case this does not answer the question of how to measure money value, the simplest way is to work backwards and cost out how much investment went in and chalk out what return it gave in terms of a positive acquisition.
Supriyo Gupta is CEO Torque Communications