Few (random) thoughts on Social media- Using C.L.A.W principle

Note:
C.L.A.W - stands for Crude Look at whole. It's a term coined by physicist Murray Gell-Mann. 
Watch this video to learn more


The current stage of social media tells us what, how, where, but fails to answer Why? It also fails (at least partly) to combine WWH to answer why?

I believe that social media is the fabric necessary to make search sensible. It's the layer that underlies semantic web. Consider how search works. I want to search for the latest android phone available in Bangalore. I key in the words & I see a list of links that lead me to a page where I might find the information I am looking for. Else I go back to the results and open a new page. Basically the pages are ranked by relevance calculated by keywords and importance calculated by referrals. However when we put the social layer on top of search we might see results which are in this form:

1. Two of my friends bought a new android phone from shop x in Bangalore and this is what their social stream (twitter, FB) says about the phone.
   
2. Or they are saying the latest offering is no good wait for another model that's going to be launched in December
   

Either way it helps me make a decision, based on what my friends recommend, rather than giving me loads of information that confuses me. But, closer look exposes a problem. What if half of my friends recommend the phone and half don't. How will I make a decision? I think this is where expert review comes in. There will always be experts in a field & their view will hold more points than an average Joe's. So my recommendation engine should tag experts to the product that I am buying. So if I am looking for the latest android phone I will be shown that even though only half of your friends like it the expert has given it a rating 4 out of 5. This same process can be used to shortlist experts. Experts can be the average of ratings given by an established media agency let's say New York Times and the people within the entire network who recommend. If I find out later that the phone was actually good the system remembers this and encourages me to rate the expert.

I believe that collective knowledge system is necessary but not a sufficient decision support engine. As the amount of information rises, it will become more essential to move towards a system that understands us and helps us choose. A system that segments the information thrown at us as soon as the day starts.

Another important point is how will the traditional advertising driven reward system change when most of us start using technology that understands us, not only by our connections & choices but by constantly learning about us. Advertising bets on the assumption that we want the stuff we are looking at or would want something we are shown. So targeted advertising bets on our ability to buy something that is relevant to our current environment. Environment here is defined as what is in front of us , what is being shown to us by the medium via which we are experiencing the world at that particular instant. But this is a loose model. As it doesn't represent whether I am even in the mood to buy or even look at that portion of the screen where the ad is being displayed. There are other errors with this model. For e.g. if i am reading the news of a car accident, the last thing I want to see in front of me is a car advertisement. Hence my behavior against the medium is not solely controlled by my choice to interact with it & so it is a vague indicator of my choice.

There are two parameters that add relevance to the connection.

• Memes
  
• My offline environment
  

Memes can be defined in context as, the pattern of behavior that is trending. For e.g. trending topics in Twitter. What this means is that more people are interacting with a particular kind of cultural unit, thought, object or idea. This clearly indicates that the unit, thought or idea is relevant and can be used as a basis for a promotional activity, which has a greater likelihood of succeeding. A major problem with this idea is that our systems aren't sophisticated enough to utilize these trends and create a targeted advertising campaign around it before the meme dies down. The system also needs to sense the meme that is likely to remain hot for a duration at least as long as to create maximum sales impact.

So this model uses targeted advertising derived from:

1. Impulse shopping tendency- people tend to buy it as long as it's hot
   
2. My Social fabric- likes, interests, recommendations, exposure
   

An interesting question that comes forward is what effect does my location have on the choice that my social medium is encouraging me to make?

From the medium's perspective, how should I trend memes so that the user is encouraged to make a move? How relevant is it as per his current environment (defined above).Should this meme represent my ecosystem or sense the web as per user's social profile? Is the user more likely to buy the latest issue of Mr. Rushdie's book if the user is near a book store and an important event just happened in Mr. Rushdie's life? As this won't necessarily translate into similar behavior for user's connections, if the user finally buys the book and likes it, should the system be rewarded/promoted for this connection? An analogy will be Devesh, 5 of your friends have tried friend finder. Give it a try? In this case , Devesh 5 of your friends have made likeable purchases using our product finder , would you like to give it a try?

I think the next step is to envision a complete transaction and contrast it with the 360 degree process of attract, engage, sell, retain & use for promotion. Also need to put some maths into this process.