Buzz made me write a blogpost after 3 years.

Dear friend @prolificd has written a well thought out post from the user's perspective about Buzz, specially its mobile version. I wanted to comment on that but what started out as a comment took on a life of its own, so here's a post on my thoughts on Buzz:

A discussion on Buzz brings me to Buzz mobile which, based on my friends' experience and a few others I have read about, is significantly different from Buzz on a browser. Google is the Internet. In the 70s Indira Gandhi's's supporters used to say Indira is India, India is Indira. Schmidt and co isn't that blatant. (Notice that I did not mention Page and Brin because Google as a commercial entity is trudging ahead of Google's technical brain, though the later does play a significant role, as you'll see.) They just want take the Internet to you, knowing fully well that while on the net your first stop would be Google. The increasing use of the mobile web platform, and they have Apple to thank for kick starting something that Palm and RIM couldn't manage with all their decade of worth of experience, would naturally turn them towards the mobile platform. As with Android on the mobile and Chrome OS on netbooks they are trying to raise the bar and set certain standards and thereby control the experience early into the market's growth so that they can influence you more. Here too, they are making great improvements in merging algorithmically curated data and UGC to provide an experience that would bring you back. All social applications leverage the network effect, where the n-th person joining the network affects the n+1th person's experience in the network. In other words as the network grows so does the utility of it, not like a traditional app/website where each person, whenever they start using it share the same experience. So they have provided a platform and its utility will increase as more people contribute. @prolificd lays down the potential scenario well enough in his post.

Google's well known Page Rank system works on the so called link economy where, at a very basic level, the no of incoming links indicate the quality of the page (mathematically, it increase the probability of the user arriving on that page while navigating without any bias). Page Rank, which incidentally is named after Larry Page and is not derived from a 'web page', is 'so 1998'. Important information generated at real time and thus not being linked to at the moment the search is taking place or not being available due to the physical limits of how often crawlers can run obviously require a different approach. Companies working on intranet search have been grappling with this problem because there documents on the network have no special incentive to link to others or being linked by others. One of the key factors in this is deriving the relevance of the document to a searcher from information other than that actually present in the document. Say, I search for a term 'passport processing' on the intranet. The search engine of course brings up documents that contain those two words, but how does it know which one is more relevant? Without too many links it looks at accompanying data like tags(provided by the author), the dept from which the document originates(Immigration dept would rank higher), the no of times the document has been used by previous searchers and so on. With the advent of social tools, collaborative filtering has also been applied. Rating, crowdsourced tagging, comments on the accompanying page, the relative closeness of the profiles (corporate directory data as well as social footprint) of the users who worked and interacted on the document and other such marginal aspects of the issue are combined to find relevance. This is commonly known as social search, the operative word here being relevance of people and data. Buzz is Google's way of collecting proper social data of its users.

As has been pointed out at many places, a gradual shift is taking place in how we are searching and navigating the web. A whole lot of people are going by recommendations from friends and people they trust. Facebook and Twitter are slowly becoming their landing pages and their social filter for things that they like. How does Google get a piece of this action? Traversing the public stream of tweets gives it access to the raw data but without the connections between people, it isn't of much help. While twitter surfaces that data, Facebook doesn't give out much. Moreover, fetching such data on a per person basis during search is prohibitively bandwidth intensive and slow in terms of the response time that Google strives for. The sparse connectivity information that these networks provide would lead to relevance matches between 50000 thousand people at the same level, ie there are that many people who are algorithmically similar to you and that's that. How do you filter further? You need more information and you need it quick. A potential solution: use the social connections that Google sits upon aka the most tightly integrated online social network anyone has, their inbox. Obviously, the whole most emailed means closest in terms of affinity is a big bad decision but I'll come to that later. The thing though is, getting you to filter those connections and taking your permission to use that data. Just because Google has you data does not mean they can use it the way they like. They are fuzzy about it and frankly speaking, without some of that data on you they wouldn't really be able to return the excellent results they do. However, explicitly using the data that you store with them requires permission and that's what Buzz does. It makes you choose or ignore people and their posts and provides a veritable trove of information to create added relevance for you. It turns you into your own social filter. This ties in to Google's social search initiative from a few months back. Facebook sits on data far richer than Google in terms of relevance and personalisation. Their initial privacy policy and commitment towards not using that data for tracking and advertising has actually lead to people providing them with better data than they do to Google. They however cannot use this data while Google who has the technology to do so does not have such rich data. Bring the two together and you have the vision behind Buzz. The future of search is not merely ten blue links but recommendation. Google knows it better than most.

With Wave Google flopped big time. A possible reason that even the early adopter crowd abandoned it is that it required a separate window kept open and adding of new contacts. Having learnt from that, Google placed Buzz in the inbox so that the one window that you keep open throughout the day has Buzz in it. This, I believe, is also part of the reason that a lot of non-techie and otherwise tech averse crowd is taking to Buzz at such high numbers, such as elderly parents, people who didn't like Wave because of its complexity and others who don't tweet or post on Facebook often. Seeding everyone's network with an initial no of people to follow was Google's way of bootstrapping this and they screwed up royally, possibly because they did not go through the usual round of testing with people who do not work in the Googleplex. The major buzz around Buzz will of course lead to this being forgotten soon enough. The sceptic in me would like to point out though that had Google used an advanced algorithm to 'read' your messages and pick users based on that, there would have been a commotion of another kind. As I said, for the mom and pop types, not providing a seed network would be the same as not providing the Buzz at all.

I end on a note that has been repeated often, this is a beta product and interface changes and user requested features will come and hopefully make things easier. So lets see where that goes.