Sumit Sharma

Archive for April 2010

The relationship bw the Media Industry and the internet companies

leave a comment »

Metadata definitions of BigData, so that context can be derived, and subsequently enrich the quality of our experience on the internet is a form of a trajectory that companies need to take.
The current trajectory that Google, Facebook, Amazon et al are on is to move on a silo’d path towards owning the internet. O’Reilly’s premise that this should be treated with caution because it may restrict openness on the web is absolutely sport-on.
Opening up APIs isn’t enough to be participants in the internet of openness. In fact, it is self serving as far as these companies are concerned because they are enabling themselves to expand themselves – or – enable their core to expand out. Granted, their cores are very different from what you’ve been calling cores in the enterprise/industrial context, it still is a restricted or limited entity. Your argument that there is always going to be someone smarter outside your organization should be their operating premise. If that were the case, they’d share their data, and spur further oppurtunities for new ideas to come up in the form of analytics, recommendation and serendipity producing engines.
If cable companies were able to take a parallel to this, and open up their APIs and delivery mechanisms so that we could apply our own filters to content, based on metadata analysis, media delivery would be severely disrupted. This would turn the world of scheduled programming upside down, and instead of forcing content down our throats, it would enable us to pick what we want. Going back to the internet companies going down a trajectory of potentially restricting openness, there are two potential scenarios I see happening: I think they’ll prompt a response from the media companies to open up, or the media companies will prompt these internet companies to open up. I say this because, whoever makes that first move will expose the power of applying contextual relevance of data delivery and user experience, and prompt the wave of opening and sharing data. Perhaps, it is an opportunity for a shaping strategy for media companies or internet companies to consider, where the act is to open up data, the asset is the data, the platform is to formulate a federated metadata definition.


Written by Sumit

April 19, 2010 at 11:28 pm

Real time search to a whole new level

leave a comment »

Siri is a new company that can take in your voice command and spit out results to you in the hope that it will be relevant of course to your needs. I’d be curious to know what their sources of data are, and if its just cold directory data and coupling that with location and some other obvious pieces of data…
For starters in contextualizing users’ situations I’d try to create a function that incorporates at least some implications of location, time and micro events
…however if you were to think about it, they are missing a crucial piece of the puzzle, a social graph of some sort. I know I’ve argued about the over-hyping of social graphs before, and I still stick to my guns for that, however think about the importance of integrating popular ratings, what users possessing similar characteristics are doing and so on. Basically, tapping into LinkedIn, Facebook, Twitter to access and build some sort of a similarity graph and categorize which bucket a user falls into, and then imbibing APIs from platforms such as Yelp to incorporate ratings as well. So the function would be:
Siri_Recommendation = f(location, micro_events, time, Social_Graph), where Social_Graph= f(social graph, ratings)

Again, its all about sharing, opening up platforms and then getting chomped on by Google, Facebook and the like because all thats unique about Siri in this case is they (better) have a kick ass algorithm (the Siri_Recommendation) or their voice recognition algorithm is so sophisticated that it captures tones, pitch and can incorporate that into the background context of the search – long shot.
Bye bye Siri, if not now, then definately later.

Written by Sumit

April 10, 2010 at 11:26 pm

Check-In apps are over-rated…

leave a comment »

Check-In apps are over-rated, let me explain…
After SXSW this year, Check-In apps have really been gaining some momentum-click here for an awesome visual – and many people, including myself, are using Foursquare and GoWalla more than ever. Right now, they’re fun, even unleashing a little bit competitiveness by inducing check-in envy or by feeding our need to be a mayor of our local café. Now lets fast forward to a year, maybe two from now when location presence becomes quite mainstream in that there will be much more users than there is today, what will be special about these check-in apps then? In my mind, the question is what will they have to offer and what will make them stand out?

I believe a key value-add for these check-in apps will be the ability to make inferences from our check-in data, and provide some insight, or recommendations. How can this be made possible? By pursuing at least the following 3 steps at a minimum:

1. Federate data: It would be very interesting to see the analysis one could come up with if you combined data from various sources – Imagine the scenario where Blippy shares its API with both Foursquare and Facebook and vice versa, the product of this could provide some interesting insight into our behavior through insights gained by analyzing the layers of data about our purchasing behavior, social characteristics and where we hang out, thereby providing a richer context. It could influence serendipitous interactions with people, products, locations, ideas and God knows what else. It would make life exciting, and add value to our life. This democratization of data is something that not only will help alleviate the information asymmetry issues that still exist today, but it will spur further innovation through the various insights and conclusions that can be made possible when data is combined.

2. Enhance the mobile advertising platform: The impacts this will have on mobile advertising will be huge. We’re already seeing location centric advertising taking off, but if you add further layers of context to location centric advertising that will be even more valuable. Imagine a platform where advertising is provided by an algorithm that has layers and layers of contextual data, incorporating elements from various different plains.

3. Making us do a double take: We will all stop and think about what we’re doing, where and why. This can be a good thing. With this influx of data being produced by us comes a sense of responsibility in what we produce. This time, the responsibility has a direct impact on ourselves, because the data we create is what defines us. Every single data point becomes a part of our persona. Think about that for a minute. Data and time will become almost similar, in that they are what define us. Only difference is, with data you can turn back the clock so to speak – if you don’t like a piece of data the defines you, you can simply delete it. Eventually, we’re going to get tired of producing data and deleting it over and over again, and soon this will lead us towards not putting ourselves in a position where we will need to delete the data in the first place. This will fruition in the form of where we choose to be physically, who we hang out with, what websites we hit, books we read, products we buy etc. Its basically an extension of, “you are what you eat”.

So, long to short, Check-In apps are a medium to solving a part of the solution – that solution is comprised of several inputs as I’ve written about here. They are an input to what aggregators will need to make as the web a more valuable experience for us. An aggregator such as Facebook is in prime position for using check-in apps because through Facebook Connect, they are actually gathering data from many external sources as their external ecosystem continues to expand. Facebook already owns masses of data, however they are now getting the ability to contextualize their data in what were once unimaginable ways. They can add layers of our spending patterns, location distribution and restaurant menu choices to our data. In a way, I think making your app, such as blippy or foursquare available on Facebook is a short term gain, but a long term suicide. They gain by being on a platform that everyone is already a part off, thereby reducing the pains of adoption but my argument has always been, see a previous post about this, that at the end of the day, its not who owns the data, but what is done with the data – thats where the advantage will lie. Check-in is already getting commoditized – just look at GoWalla and Foursquare competing head to head, they’re already sharing APIs with Facebook and the like, and before we know it, it won’t be which app you’re using to check-in, or who is seeing you check-in, but what value is there in your check-in…and the check-in app will be a vehicle, a part of the puzzle, not the sexy app as it is treated today.

Written by Sumit

April 1, 2010 at 4:24 pm