Wednesday, December 20, 2006

Social Networking in India

Orkut, Gazzag, Friendster, Myspace and a score of others! What is so attractive about social networking that we hear almost every week about a new social networking portal?
Well, there are many reasons. It would be repetitive to list all of them here. Though the most fundamental reason I can think of is that these sites are a gold mine for search portals, job portals and all such other portals that makes them so attractive for internet firms and VCs worldwide. People who otherwise fill personal information forms elsewhere with "not so correct" data seem to fill loads of "correct" data at these sites which can be mined to correctly serve targeted ADs.

Targeted ADs then translate into good revenues for firms who advertise online. This strategy is very cool for US and other nations where e-shopping is a big rage.
But what about India ? Here we do things differently. We hardly buy anything online. E-shopping is not such a huge business here. Why and how can a social networking start up target good and steady revenue in a market such as India. My solution is a little far stretched and may not be true at all. But then it never bad to listen to a new perspective.

When I see social networking portals such as Yaari and Minglebox coming, I see a clear lack of business idea. How are they going to make money in market where large internet user base doesn't translate directly into revenue. And also what different strategies are needed to tap this market. Make no mistake, user base of such portals is big and used by people who have money to spend. Every tom, dick and harry in my office and elsewhere has got an orkut account. Not only that these people log on to these portals several times a day.

First look into differences in Indian and US internet scenario. In India, despite a major telecom boom, internet speeds are slow to say the least. Most of the ISPs provide at max a 256K'b'ps connection. That is Kilo bits per second and hence at best can provide 32KBPS of net speed. (Two things this 256Kbps generally tops at around 150 Kbps and MTNL just annoucned a 2 Mbps connections...Things are bad but changing!). So what does this has to do with a social networking portal ? A lot! Bad net speed translates into a slow and heavy social networking portal into a beast which sometimes takes 1-2 mins to load a page....sometimes. This maybe a reason why people in India are still stuck with Orkut when elsewhere Myspace and other such rich social networking portals are becoming popular.

On the other hand telecom boom has made sure that almost everybody now has got a phone and people spend a lot of time and money on these little gadgets.

So the million dollar question which still remains unanswered is who does one go about making a Indian social networking portal? Here is what I suggest
1. Make it simple, keep the total loading time on a 128 Kbps connection less than 10 secs...preferably 3-5 secs. Thats a big task. 128-256Kbps would generally be around 80-170 Kbps. Thats 10-20KBps. That means pages should not be larger than 50-100KB (at max 200KBs). You need to let go a lot of images that you put into those menu bars, buttons etc. A text only page with 8-12 , 5-7 KB highly compressed images seems to fit the bill. Keep that CSS file as small and custom made as possible.

2. Server side need to be very quick. Java is a complete no -no. Php would invite wrath of hackers(though it would be fast) and would be difficult to maintain (coding wise). RoR is good and fast too...with less security issues and a carefully tweaked Apache server with a fast database (I am careful not to mention which one) would be good. Lisp is also a solution. Static pages is not a bad idea it seems(if security aspects are taken care off). What I mean is that user profile pages and home pages need to be a static html file. No database connection (or very less of it). Would definitely be faster that database driven dynamic pages.

3. Ajax is the key. A lot of users in India like Gmail. Just because it's swift and prompt. A lot of ajax coding would also make dynamic pages a lot faster than they usually are.

4. Link with major Indian ISPs in such a way that there are server which serves the user has a internet back bone of that particular ISP. This is rather doubtful starter but whats the harm in trying.

Business side of it
in the next post

Sunday, December 03, 2006

M$ and Novell

Not as bad people are projecting it. Give some credit to M$ for atleast taking a step in forward direction.

Google and Orkut

It's nice to see small changes appearing in Orkut but I still think there is a lot of room for improvement in orkut. For instance that Donut thing is really very irritating. Also the way they have integrated gtalk with orkut! They could have integrated the Gmail-talk into orkut and that would have been quiet something. But still...atleast a start. Scrap alert feature is quiet good though. But scrapbook page needs to be ajaxified...completely
What about a picasa and orkut linkage ?

On Future of Google

Google recently integrated most of their services. Google talk with orkut, google talk with gmail, docs (writely) with spreadsheets, docs and spreadsheets with google pages, Google groups new avatar in gmail format etc. A lot of stuff still needs to be done like photos with gmail and gtalk(maybe even docs). Then there will be search integration with spreadsheets, blogger with gmail (even gtalk with blogger...who knows!). And many more such small-big changes...
But what Google apps or for that matter other web apps lack is a server at the client end. I think we users think it's very unreliable to have a important document stored only at Google servers. I think there is a need for Google to use Google Desktop as a server at the client ends so that user may also work offline. This, in my view, requires a lot of innovations. Saying it easy but actually doing it may not be possible at all...due to concerns of size, memory requirements, reverse engineering concerns etc. But if done, it would be a definitive step towards a web based era of desktop computing and this I think is what google is about to do in near future.