Month: March 2011

Gingerbread On The Samsung Galaxy GTi9000

I have been running the leaked early beta release of Gingerbread (Android 2.3.2, I9000XXJVE) on the Samsung Galaxy S for about a month now and I can tell you that it looks really really good. The build is very buggy, so I would advise anyone who is tempted to do the same to hold on till Samsung puts out an official stable release (it will also not play nice with your warranty should you chose to run it on your phone).

There is little by means of cosmetic changes that is noticeable to those who are not detail-oriented and it is largely consistent from the 2.2.1 days. The notifications pane gets a new coat of paint and new icons. The apps stack is a straight carryover from 2.2.x, though I guess they may change it by the time a stable release is out. Almost everything worked out of the box for me and installation was done using Odin, which is never a comfortable thing.

Before I flashed to this build, I was upgrading to the CM7 beta release for the SGS, but after I set up the phone it figured that it just would not get on the mobile network. CM7 was looking really good till that point. I will probably move to that once they put out a stable release.

The Gingerbread release by Samsung, though, fixes most of the bad issues of previous releases of Android on the Galaxy S. The boot process is really really fast, unlike previous releases where you had to wait forever to get the phone booted into an usable state. The fact that I have not rooted it or applied any lag-fixes on it should tell you how good it is. Neither Eclair or Froyo have run unrooted/non-lag fixed on my phone for long since the experience used to be that bad. If they can close out the nasty bugs, Samsung may have a winner on their hands with this.

Since it is a beta, as mentioned earlier, the OS is still buggy in places. It can freeze up and a few reboots are the norm in a day. For me, that is worth living with since none of the cooked ROMs have given the kind of battery life this buggy release has given me. I have tried it in various combinations with wifi, GPS, data and it has consistently outperformed anything else I have run on this phone.

It will be really interesting if the guys at Cyanogen can fix up CM7 for the Galaxy S in time for Samsung's official release of Gingerbread. In either case, it is going to be finally happy times for SGS owners – another couple of months and the phone, which is an excellent hardware package, should get the kind of software it truly deserves, making it a really good iPhone competitor at a significantly cheaper price point.

Filed under: Android, Industry, Mobile

What Ails Broadband In India

Every year we are presented with a rosy picture of how the internet will explode in India in terms of growth and usage. Even though this scenario has remained unchanged from the times of bubble 1.0, active users in India has never gone over 100 million by even the most liberal of estimates. In the years since 2005 this story changed to hang all its hopes on 3G as the one-stop-solution to all the penetration woes. While the 3G auctions may have made the government's books look a lot nicer, it has done little to further the cause of truly empowering India on the digital front. Meanwhile, with 3G pricing showing little recognition of market realities, the new saviour in town has switched to LTE. I won't hold my breath for that one to work either.

So what really ails broadband in India? I have looked previously at the product aspect of the issue, so I won't go over it again. In this post I will cover policy and pricing instead.

Policy: For a nation that can benefit hugely from a well-connected population, we have policy and governance that does little to foster it. The astronomical rates at which 3G auctions have gone will ensure that telcos have little inclination or incentive to price their 3G offerings at a reasonable rate. Even though there are plans that are as cheap as Rs 99 per month, they have such low usage caps that with any real-world usage, users will wind up paying many times more that amount.

We need to remember that the current environment is one where the bulk of customers are generating less than Rs 300 in ARPU. These are the people who make up the majority of the market and expecting them to pay even Rs 100 on a whim on data is being ridiculous. So you can imagine what a non-starter the severely capped/priced plans are going to be. For the larger market to use digital services liberally, the pricing needs to be much more realistic. Yes, the regular users are using the internet more these days, but we need more non-regular users to be online.

Internet in India has the potential to be transformational beyond belief. We have made positive strides in e-governance and digitizing parts of the government machinery, but the government needs to do more to provide the platform on which a vast majority of our can access information, education and services. The current levels at which the services are priced, they don't ensure either unfettered use or they wind up being unaffordable.

Pricing: There is considerable pressure on margins for telcos from voice. Since voice APRU is not exactly shooting through the roof and VAS having now gone past its glory days the squeeze is truly on in data ARPU. Even in 2009, the ARPU for data on Airtel was $13-$14, which even by today's standards is considerably higher than what most telcos earn on voice. In 2009, you could easily find more than a handful of plans priced under Rs 1000 per month.

By Airtel's own admission its current popular plans are Rs 1299 and Rs 699 per month. By 2009 the company had completed work to make its network 100% ADSL2, which meant that any further increase in infrastructure cost would have been because of the upgrading their network to VDSL2 in select areas. The other cost component, backhaul, does not seem to be a major problem since they were easily allowing 100GB caps on lower end connections earlier, which they are now scaling down to 20GB – 50GB, depending on the plan you are on. Moreover, the network has inbuilt capacity to support IPTV, which requires about 8Mbit of sustained connectivity. So, the story of network capacity being a bottleneck does not hold too much of water.

Looking at the figures, it is easy to understand why the company is going hammer and tongs and after getting more margins out of users. It is easier to get a data user to shell out an extra Rs 500 per month than to get a voice user to do the same. This, though, is an awful thing since it is now leading to a situation where due a lack of transparency and bad pricing usage will go down in the longer term as users are forced to shell out more for less.

Looking at Wimax, which is the only realistic option that can get the internet to areas beyond the big cities, the situation is no better. You will rarely find plans here that cost less than Rs 1000 per month. In economic classes where the spends between all family members combined on voice is less than Rs 1000, how can you expect them to shell out Rs 1000 for data alone, especially when there is a lack of compelling applications to even get them started in the first place?

Conclusion: The big Indian digital story requires us to get most of the population hooked up to the internet. Right now, at even moderate usage levels, that will cost at least Rs 700 per family. This is unrealistic and the root cause of why the proclamations of 'the year of internet' in India has never come true. At this point in time neither the telcos are willing to help the situation alter its course by introducing pricing conclusive to growth and usage nor is the government doing anything to ensure that they have every incentive to do something like that.

Filed under: Industry, Internet

Why The Y Combinator Model Does Not Work In India

Y Combinator exists in an entirely different environment from what we have in India. For any healthy VC/seed/angel funded start-up ecosystem to exist, it requires two things as must-haves. The first is the existence of reasonably sized exits at regular intervals via the IPO route or through the M&A route. The second is the existence of large enough markets to support at least two to three profitable players in any domain.

YC exists in an environment where the two conditions are well satisfied, leading to a situation where there is a considerable premium on a quicker access to funds (for the start-ups) and good talent/concepts (for the investors) where there is often an excess in supply on both sides of the table. India, on the other hand, has a real shortage on both fronts – good start-ups and good investors.

The main problem is that the Indian market for digital goods and services is tiny. In a non-existent market, neither product finesse nor pricing can make much of a difference. There is barely enough size in the digital domain to sustain large profitable companies. At best we see companies that are small and profitable or one market leader who is struggling to grow in its domain or diversify out of it.

In such a situation you can't get into the early-stage/start-up investment scene armed only with effort on your side. For the fight in underdeveloped markets the right gun is money and you need plenty of it. If you walk into it with effort and connections as the only weapons you won't go anywhere. Sustaining a first-mover disadvantage is a costly exercise in greenfield areas. If you don't have the money to sustain your portfolio companies, the effort alone won't save it. It is a finite non-scaleable factor.

In trying to sustain the number of portfolio companies mostly through effort than funding, the incubators/seed/angel investors won't be able to give the companies the time/effort required to grown them. Thus they actually reduce the ability/probability of the portfolio companies to succeed.

At the current market size I don't see the possibility of any of the smaller-scale funds being able to do a great deal of good. Their possible exits are so small that I can't imagine too many LPs being happy with the rate of return the GPs can bring to the table. If you invest in the region of ~10 lakh per company, the domain that you are investing into is also likely to be very small. Since we don't have well developed markets, it is not possible to start small into a big product.

That said, the parties who can play big in this – the VCs – are also increasingly moving away from their traditional roles and behaving more like PE funds. This is understandable since there are not enough big success stories in the market that can come close to the kind of rate of returns required by the VCs, thus they wind up investing in established companies than early stage companies or start-ups.

This is, obviously, not a healthy state of affairs. In fact it is quite regressive and it does not bode well for the country as a whole. Everyone can see the untapped opportunity, but the overall direction we are headed in is making it extremely unlikely that we will ever tap into it.

Filed under: Business, Misc, Start-ups