Sumit Sharma

Archive for February 2010

I missed out on a Serendipitous encounter yesterday

leave a comment »

Lets take this idea of influencing serendipity to a whole new level…to the physical plain enabled by physical location awareness and entities having the ability to have a DNA that qualifies whatever the representation is into a discrete set of values…and in doing so we’d be able to influence serendipity to a whole new level.

Here is an example of what I mean:
Yesterday, I was walking down 3rd street in San Francisco, and came across an advertisement on a moving bus for a Kathak + Tap dancing duet. I’d have loved to have known about that – but I didn’t. So then the question that arose was, perhaps I’m not plugged in enough? Perhaps I need to subscribe to certain newsletters or organizations? But then, I was thinking, how many do I subscribe to? Who, when, what and where do I go to, look towards and sign up for? Perhaps the answer lies in the premise of enabling serendipity on the physical plain(using location awareness etc), using my virtual (i.e. Facebook profile, Ticketmaster activity etc) and physical past activities (concerts attended etc) – so think about this scenario:

The ad on the bus has the following characteristics:
1. It has a well categorized description, using standardized descriptors (I’ve already written about the need for standardized metadata – this is where the semantic web comes in)
2. Its got a geo-sensor enabling it to be discovered if it slips my eye.

Here are my characteristics:
1. I have a user profile, be it on FB, Google, Amazon, Yahoo or all the above meshed into one. From this, the aggregator knows my interests…
2. I am on my mobile phone which is also location aware

Here is the role of the aggregator, or the serendipity producing engine (I’ve already written about aggregators and their importance and click here for that):
A bus comes near me and just by chance, it happens to have an ad, which according to the aggregator, I’d be interested in seeing. So, after recognizing I’m in the vicinity, it sends me a non intrusive email, text or whatever I specify in my preferences, and notifies me of this event. This eliminates the need for me to always worry about self selecting myself into a segment of customers that want to know the latest and greatest of something specific. Sometimes I like to be surprised…And, this reduces the work for aggregators to a certain extent because instead of always being on the look out for ads, content and information for passing on to me (in the background that is), it only needs to do real time serendipity..not online matching, but real time matching. Speed Dating.

More…this provides a great platform for people to get themselves out there, advertise products and services and infuence chance encounters with potential un-suspecting customers. Taking this even further, because there is no end to the potential of this stuff…but what if it hooked us up news/content based on certain similarities with other people, not necessary in our social graph, but people who frequent the same places as us, or those that share our bus route or drive a similar road everyday, or perhaps happen to download the same songs we do…this would all be invisible to everyone so privacy wouldn’t be an issue. The aggregator is merely a broker of information and how it does it is a black box. The list goes on. How cool would that be?


Written by Sumit

February 21, 2010 at 11:57 pm

Why Comcast should partner with Facebook, Google, Digg, Yahoo….

leave a comment »

All the data and information online is being sliced and diced in order to provide a better user experience for us: our user behavior and characteristics are analyzed by Digg, StumbleUpon, Aadvark, Geodelic, Yahoo, Google and Facebook to provide so called ‘relevant’ search results, recommendations and targeted advertisements. I’ve already written about the importance for a framework for organizing the vast web of data online to make this a feasible goal.

However, this movement begets the question of why are we only restricting such ‘relevance-production’ to online activity? Isn’t Comcast also a delivery mechanism, or aggregator, of entertainment to us? Only difference being that its on the idiot box instead of a laptop. So based on that, I do believe that Comcast and friends are in prime position to get ahead of the curve of expanding this idea of providing greater relevance to users and consumers of data and information.
The disruptions this will have to the business model of the media and entertainment industry are huge! The delivery and distribution models as we know it today, where we subscribe to a set number of TV channels for a fixed price is actually an out-dated business model that doesn’t serve the consumer, nor does it serve cable companies in the sense that much revenue is left on the table because of their lack of flexibility in contract structures with the Studios, production houses and speciality cable channels. If we were to treat the information available on TV networks as a cloud of information, similar to the cloud of information on the internet, my argument that the cable channel delivery model is outdated might begin to become clearer to you.

I see a couple of variables that come into play, which separate TV content from the internet:
1. Flexibility of accessing data and information. This is to describe the degree of walled areas of information for which special access or permission is required in order to access it. For example, in the internet this is less of a restriction as it is for TV if you were, say, a subscriber of only a basic cable package and hence wouldn’t be able to watch HBO
2. Layers of available information. How much information can we access about the information we’re reading if we want to go deeper? On the internet we can click on links, or go to wikipedia. On the TV, there is no interface to find out that pitchers stats from last season, or his college. TV content delivery is a one-way affair.

How about a situation where content is delivered, not based on what networks think should be done, but based on what users want. Perhaps, the so-called “network programming” is also an outdated idea that should be progressing with the time. How many times are we told which websites to surf…never! We have full freedom. The one notable thing missing from the internet is the ability for content providers to reach appropriate audiences in the right context, as well as the ability for serendipitous moments on the internet to occur freely.

Now, imagine the same phenomena on TV. For this to happen, perhaps TV shows need to have Genes mapped out so that they can be accurately and, with serendipitous relevance, distributed to the right people in an appropriate context. So, how can it provide us relevance…lets break it down really simply:
1. Know who your customer is
2. Know what they want and why. Can’t over-emphasize this second part enough of needing to know “why” enough – because this unlocks a major constrain in the relevance delivery function of aggregators
3. Know what you’re providing

Perhaps, the question rather should be: Should Comcast look towards Pandora?

Written by Sumit

February 14, 2010 at 11:41 pm

Why Google Buzz isn’t working for me

leave a comment »

A couple of days ago I can’t remember where I read/saw this but they compared Sergey/Larry to Mark Zuckerberg:
Sergey/Larry: when you meet them, there is an awkward moment and not sure what to say or do…
Mark: he’s more of a “whats up man” kind of guy, and will give you a hug…

Enough said.

In my last post about whether analytics should be treated as a science or an art, I argue that there is an aspect of creativity and actual understanding of human psychology that should go into the algorithms producing data before us…Take a look at my friend list on buzz…out of the 26 people that are following me on day 1 of its launch I have only 5 truly relevant people…that is people I consider good friends and whom I’ve interacted with in the last 6months at some level. Once again, using the silo’d data set of gmail correspondence, Google has, to its utter embarrassment, taken liberty to go at this alone with out Facebook, Twitter and without analyzing other activities I do.
Of course, I’m not saying that Google should be producing friends out of the blue to me (I for the record wouldn’t find that creepy but I know 99% of you would – in fact I think that would be a great experiment) – in fact don’t these dating websites do that as well? – (read: we have double standards but thats a point for another time) but at least they need to stop being so robotic about it. You can’t just use plain math to hook me up with people, Google!!!
FB isn’t perfect, but you need to learn from there…give a hug, smile maybe and loosen up (just like Mark).

To sum it up, John Battelle couldn’t have said it better in his blog about Buzz:

Buzz does not let you do pretty much anything at all with Facebook. Buzz tries to find your friends and connections through algorithms that watch what you do on Google services, then lets you add more, but through the lens of Gmail. There are two things wrong with this: Algorithms, and email. Facebook didn’t depend on either to create its initial value. Instead, it let humans pick other humans, which honestly, is what social is all about at its core.

Written by Sumit

February 11, 2010 at 6:42 am

Posted in Social Networking

What blog sites need to do…

leave a comment »

With a great deal of humility, I would like to offer some advice to WordPress (the operator of this blog site), and I hope they follow it….
First of all, just for a minute please put aside fears and doubts regarding privacy, security, practicalities of how to get there. Now let me set some context…There are hundreds of thousands of bloggers out there, many of whom I’d like to occasionally read some interesting points of view. The important assumption is that following one person’s blog and each and every one of their articles is like only wearing a single item of clothing all year round, boring and after a while it becomes stale…’it’ being your relationship with that blogger’s articles. That blogger may continue to churn out all sorts of cool articles but they aren’t always going to be what you want…yes there is that element of being starstruck and feeling the need to read each and every article written by Bill Gates but I don’t think thats an very good diversification strategy , at least from a personal relevance perspective…

Which brings me to the point of this post: Lets access those 100s of 1000s of bloggers, out of whom at least a handful will post something that will be appealing to you…Question is, how?!?
Here is my suggestion of how:
1. Get to know the reader. Yes, WordPress – get to know who I am, what I’m interested in, what my background is, what I’m writing, who I hang out with and all that.
2. Get to know your data. Again, WordPress…get to know what your writers are writing, categorize them in meaningful ways that can be easily federated with other blogs and articles in various news sites – you won’t regret this because in the long term they too will be able to make easily constructed API calls to you as well.

Thats it, you know your readers and you know your content. Match made in heaven. Here is what happens now that you are in this enviable situation:
I get presented with cool articles from people that I didnt know exist, based on the WordPress intelligent algorithms. Yeah right, if Google and FB still can’t get relevance right good luck…but at least you konw what direction to head in and the steps needed to get there…

Know me. Know what I write. Now repeat this for every other user and article.

Written by Sumit

February 10, 2010 at 11:15 pm

Is analytics an art or a science?

leave a comment »

I’ve already written about how Facebook could do a much better job at providing me relevant information and as I increasingly ignore whatever ads that Facebook presents to me on my homepage I ask the question of Facebook, and other large hoarders of data (they know who you are…): whats the point of protecting your data? Wouldn’t it be that much more helpful if analytics were applied, not just to Facebook’s silo’d data, but to data for example? Perhaps, the pressures to produce revenue are clouding the minds of the product managers and marketers and inhibiting a purer, more deliberate effort, at providing relevance. Right now, this wave of advertising gains through silo’d data is well and high, but it is my humble prediction that unless more intelligent analytical decisions are made, this (dare I say) bubble of advertisers’ excitement of Facebook’s internal analytics applied on their own internal protected data will give way to that next new start-ups innovative way of applying analytics to a much broader set of data. I’m being dramatic I know Facebook won’t drop the ball, but you hopefully get my point….

Sometimes I wonder what is needed more: Statisticians or Psychologists? In my opinion, I think its an art, and so psychological analysis will entail how humans behave, why and probe deeper into our behavior which will be the foundation for providing …relevant material to us, and of course to influence serendipity. So hopefully all these incumbent search engines, content aggregators and friends realize that one day, people are going to get tired and frustrated with the imperfect results in terms of relevance of data and information and begin to share APIs and their protected valuable data which I think will continue to lose value and hold back innovation. So, in essence the emergence of search engines and content aggregators may be the start of a sort of bubble (how big this bubble gets remains to be seen) – my prediction is that once this bubble bursts we’ll see headless chickens running around (those that didn’t put a lot of effort into creating relevance through using creative analytics and appropriate metadata filters) and the strong survivors will be those that have gotten their analytics right by putting in artistic and creative effort  and in trying to aggregate as much data as possible from as many APIs as possible.

So to all those that feel that Facebook and Google have already won the game, I strongly disagree…yes they’ve got volume and hence plenty of traction, but sources of disruption never care about how big the incumbents are…lets see…

Written by Sumit

February 8, 2010 at 5:18 am

Some thoughts on Digg

leave a comment »

I’d like to offer my perspective on Digg so far since I’ve started using it only recently and will admit that I’m not as active on it as I think I’d like to be. I think one of the reasons is because there are too many choices for me to choose from in terms of sharing what I read or come across and to some extent, I’ve invested so much in some of the choices that I don’t feel like changing unless Digg offers some higher value to me… I’m already a Facebook user and am also fairly regular on twitter so I’d probably post what I like onto Facebook or tweet them – I know there is the interface into Facebook via FB connect but what would truly incentive someone to do that when they can really just “Faceook” something? Facebook, through its Facebook connect platform, has put itself into a position to where it has one up on the Diggs and Stumbleupons of this world because on top its knowledge of what we like or prefer, it has an extremely well established social network. Digg isn’t going to compete with Facebook as far as social networks are concerned, and so to increase user ship I think Digg will need to overcome the following challenges:

1. Increase penetration and omnipresence across the internet: Lack of a uniform, mainstream population that are Digg customers…if Digg truly wants to become the de facto aggregation tool, known for providing trusted relevant content to us, it  ideally needs to increase its presence and breadth across the web because it needs to accurately analyze our overall web activity, not just activity/article/webpage that happens to have an associated Digg link which appears to be where the majority of Diggs come from. This entails for the Digg icon to be available almost wherever we browse and whatever we do. And if it can’t be everywhere, then at least Google, Facebook, Delicious, Yahoo, Aadvark, Fark and StumbleUpon and many other such services will need to share their APIs AND data with Digg so that it can have access to the vast amounts of data concerning user activity and related information spanning across these entities in general.  What Digg needs isn’t only just API sharing, but data sharing – to clarify, API sharing allows Digg to build an interface into Facebook but thats helping Facebook more than Digg since Facebook gets more data at Digg’s expense. Rather, data of Facebook users’ activity needs to be shared as well.

2. Lack of matching of people and recommending what someone like me might like, lets not confuse this with using my social graph though … I’ve already shared why I think the notion of using our social graphs to gauge our interests is probably a misconception. If we were to focus onto matching commonalities between users, and maybe not just based on social networks and commonalities of Digg histories, but based on other characteristics – more thought needs to be put into this which brings me to my third point…

3. There is little or no ability to map out other variables that come into the physical and mental worlds…i.e. integrating time of day, location, micro events, macro events, state of our mind and body and thousands of other variables for accurately understanding context to name a few: retail, dining, music etc. The more the better. Of course all this depends on the prevalence of sensor networks and advancement of sensor oriented gadgets. My argument that humans don’t just exist in the online plain means that aggregators and search engines will eventually need  to bridge the physical and mental worlds. Basically, extracting as much context as possible, and recognizing that context doesn’t just come from analyzing online activity in terms of what we “Digg”, but should also encapsulate our activities in the physical world, and even more – what goes on in our mental state…are we happy, sad , frustrated, curious, stimulated while making a “Digg” ? Imagine if Digg has all this information available to correlate our Diggs with in order to better understand the context of why, where, when we Digg.

Written by Sumit

February 5, 2010 at 6:35 am

Lets combine our social and professional networks

leave a comment »

This note is about the need to consolidate professional and social networks.

In the theory of The Big Shift, John Hagel has spoken about the need for individuals to develop trust based relationships in order to foster a culture of collaboration and information exchange resulting in tinkering and productive friction. I agree – I’m more inclined to sharing ideas with someone I can trust and get along with, and this trust factor is enhanced by social commonalities.
In this regard, I do feel that it makes more sense for sites such as LinkedIn and Facebook to converge towards a common platform and eventually consolidate – it just makes perfect sense. The case will be made even stronger for the future generations and here is why: (Link to my LinkedIn sustainability article).

Ellen Levy noted during a Supernova09 panel, people will have 3 options in filtering their data: brute force, artificial intelligence or social graphs. I believe that the latter two options can’t be separated since social graphs, as we know them today, don’t contain enough relevant information. The networking services today are operating in silos and have very specific information and I think that all this information should ideally be merged before we can extract meaningful relevance.

Coming back to the fact that I separate my Facebook, Twitter and LinkedIn networks: right now they are all separate, and until and unless Facebook can share user data with LinkedIn and Twitter, I will need to migrate all 3 social sites to get meaningful information. Why do I need to go to all three sites for my information? Perhaps its because of different activities and interests at different ages leading us to becoming hooked on a certain platforms and before we know it we’re members of several social and professional network sites.

My activity on LinkedIn started off because I received a few invitations from professionals who at the time were early adopters of. As more people such as myself started joining, we were trying to start a professional network because it was the only place where we could. However, at the same time, we found ourselves joining Facebook for staying in touch with friends and family. Facebook then went onto evolving from a networking service to a entertainment platform so now we’re on Facebook to play games with each other in addition to the original reasons why we joined.
We then joined twitter because it became this powerful tool to stay in touch with the same friends we have on Facebook, however in a much more efficient manner. Another reason we joined Twitter was because it was a way to get insight into the lives of famous people and professional role models. It became a forum for learning and knowledge transfer – a new advertising platform for bloggers, thinkers and innovators to share their ideas. These are three very distinct reasons why we are members or these three social networks.

Now if there was a way for us to merge our LinkedIn networks with our social graphs, and do so in a way that fostered meaningful relationships between people with commonalities that extend beyond professional affiliations that would make a lot of sense.
A huge factor in getting to know people and develop networks is why people go to business school. Think of activities we pursue in business school as a giant Facebook community – we drink, party, eat and enjoy social and cultural events. We then use each others networks to grow our networks and explore areas of our career that we’re interested in. This in my mind, is an example of why Facebook and LinkedIn need to share information with each other. In order to estabilish that professional connection, or introduction, I need to get to know someone and have a beer with them or share a joke with them…Why are business deals made on the golf course, or in that luxury box during a ball game? Its that social and cultural bond.

Right now, which networking site has the upper hand, who controls the value? Is it Facebook with its social graph and a majority of humanity with its ever increasing membership base of 350M members?
 Even if there was a consolidation of networks, our social graphs will only be a single element of what we need to help filter information. There are so many other dimensions such as affiliating the context of location, time and our social, extra-curricular and professional preferences to our decision making algorithm. Social networks are definitely crucial however, there is an assumption that they are accurate reflections of our preferences which may not always be true, nor should they be treated as the primary set of reference points in filtering information to us. Limiting choices to us based on social graphs may limit serendipitous moments. Rather, we need to look at ourselves, our own behavior and figure out what we are like and using other data sets such as social graphs, we make filters that are much more richer in their logic and design.

If we start from Facebook social graphs, and use them to create LinkedIn relationships, wouldn’t that sound easier?

Written by Sumit

February 2, 2010 at 8:56 am

Posted in Social Networking