One of the areas of my interest has always been statistical characterization of systems.
In the early part of my career, I used to be interested in the number of clock cycles taken for a particular foreground process to complete, page fetches needed etc. because I used to work in assembly language on a cross-assembler (page fetch was done as a macro). This led me into a lot of number crunching. I did this for my own pleasure because these numbers did not matter in my organization, which was an old-fashioned manufacturing unit that also made telephone exchanges amongst a variety of heavy engineering output.
Subsequently, when working with the GSM operator, I noted with great pleasure that my job included traffic engineering the public network’s switching systems. It gave me great pleasure to look at all the numbers, identify central tendencies, apply central limit theorem etc. and come up with predictions. One particular piece of work that I did with a friend from the Marketing team, a number crazy IIM grad, for predicting hours of least traffic at various cell sites to carry out planned maintenance, is still part of my fondest memories.
Alas! The world has changed…
Things have gone viral and peer-to-peer. So, there is no single point anymore where we could peer (pun intended) into the nature of the systems.
Let me illustrate with a point…
When Steve Jobs died, I started recording a few statistics. As a person who once started a page and posted a new entry in Wikipedia, I was interested in what’s going on at the Steve Jobs page. So, I got to the page-history and actually counted close to 600 edits within the 24 hours of the news breaking out – an extraordinary jump in the number of edits on that page!
Contributing to Wikipedia is voluntary work. What motivates these contributors? Is it a competitive advantage, one-upmanship or the flaunting peacock syndrome? Whatever it is, it forces people to contribute voluntarily!
Let us take another example…
Let us assume that you notice a guy walking along the path ahead of you. He keeps talking out loudly, apparently to nobody, every now and then. Upon closer observation, you realize that he is speaking aloud, apparently to himself, a variety of things that are about his work, his special diet, his exercise regimentation, his relationships and some more inane chatter. What is worse is that you realize that a few people walking in the opposite direction, totally unrelated, repeat what this person has been talking about for whatever reason. Looks like a funny visual, isn’t it?
And yet, on Twitter, we see a similar phenomenon! I have done some analysis of Twitter usage based on the sample of my connections. The average number of times non-celebrity tweeple, in my sample, tweet (as I understand the words are) was quite close to that of the celebrities! Looks like the common man has no less a need to announce about himself to the world as that of a celebrity!
There is definitely a projection of self and a promotion of self-interests through this process (quite unlike the outward altruism exhibited on Wikipedia). Besides the question of what induces this, once again, we notice the voluntary nature of contribution of this information!
The most voluntary group, if superlatives could ever be used in this context, is that of subscribers in an open communications market!
How do we influence the subscriber to act in a particular way in which we want them to? The obvious answer would be to deploy systems for “Business Intelligence”.
However, considering that we already live in a low-margin world, the cost considerations involved in such solutions quite often outweigh the benefits that we can predict or the returns take unusually long terms to materialize. One solution: Service Providers should look at is the ability to create and use a Cloud Based Business Intelligence solution that they can utilize and, offer as a service to their enterprise customers as well.
Most importantly, what is the source data on which we can develop such intelligence? Other than the usage patterns and a few other characteristics, exploiting which would be considered a breach of privacy rights(!), is there anything else that we could use?
Indeed! If the Wiki and Twitter models are analyzed, we realize that we live in a world where we need not spend a lot of money and worry on how we get intelligence for our BI systems – the customers themselves are willing to state the same!
All that a Service Provider has to do is one of two things – connect your BI systems to your customer’s Social-Networks and/or create and sustain a social-network context in your product offerings.
One word of caution though – in my analysis of social networks – I have noticed a tendency for a conditioned projection; an avatarization of the individual. This can cause errors in systematic inferences drawn about customer choices. The good (or bad) news is that, as we go into the future, the next set of subscribers comprises persons who have been involved in online multi-media role-playing games from a younger age – there is likely to be lesser divergence between real-life and cyber-life personality traits (which is a totally different subject).