Archive for December, 2006
Rob: did you hear?
Tony: hear what?
Rob: The NFL decided they wanted the equivalent of the yellow line to help the referee’s determine whether the team got the first down or not. They announced a contest: million dollar prize to the winner.
Tony: Its about time they fixed all those mistakes.
Rob: should we try?
Tony: yup … million dollars buys a lot of beers dude!
Rob: any requirements?
Tony: the solution is expected to: (1) tell us in real-time that the line had been crossed; (2) also it should only get tripped when the person carrying the ball crossed it and in the right direction; (3) not effect the normal play.
Rob: any ideas?
Tony: not really … but I remember hearing that our IT uses a product called Tripwire. Let me go ask them what it does … seems like something we could use
Rob: wow man … we could ski al year long
[ … next day …]
Rob: whazz up man … did they let you in?
Tony: yes man … but I got very confused there. They use this thing called tripwire … but it doesn’t trip. Someone could cross it a million times but it trip’s only when the manager decides to go run it.
Rob: that wouldn’t work.
Tony: nope. It will be like having chains … you got to go check each time.
Rob: I can’t imagine using chains on all the machines they have, no wonder they look beat like our referee’s with everyone shouting at them
Tony: It gets worse. This Tripwire they use doesn’t even tell you who crossed it. Or it kind of does, it tells you the last person who crossed it.
Rob: but we would want the first person wouldn’t we?
Tony: yes we would and also not just one person … but every person who crossed the line with the ball.
Rob: so this Tripwire stuff can’t tell you who or when they crossed … what do they use it for?
Tony: I don’t know man …
Rob: maybe your guys are not upto snuff on this stuff … I have a friend who works at this big company and he does something to do with SOX. I remember he was up one night and mentioned tripwire … maybe they already have something like that in baseball … I will check him out.
Tony: later dude
Rob: cya
…to be continued …
December 19th, 2006
I was at an ITSMF meeting recently and the speaker did a show of hands:
- How many people had deployed a CMDB? One hand went up.
- How many folks had actually seen a live CMDB? Only one hand went up.
- How many folks had projects to evaluate a CMDB? Entire room put up their hand.
- How many folks think that these projects would lead to implementaion? About 4 hands went up.
One of the things which came out of the meeting was people were struggling with what to put in a CMDB. Most people felt that ITIL was synonymous in most organizations with either service desk, incident management or change management. But how that translated into a CMDB was not clear.
December 15th, 2006
1. Outsourcing will reduce cost of my product release
At some point or another in the life of a startup, the question gets asked: why dont you outsource your development to India (or China)? And ofcourse everyone on the board talks about how this would save money and reduce the burn rate.
These days my advice to most companies is don’t do it. For most companies the length of the development cycle more than doubles when they outsource, because of the communication overhead and the distance eliminating hallway conversations and face to face interaction. So even if you reduce your engineering burn to 1/3 (which is roughly what it works out to be), the total cost of the product is increase in cycle time * burn = only 2/3rd of the original cost.
In addition it also slows you down, which coupled with the fact that you are spending some money on other functions also, may cause the total cost over a product cycle to achieve the same results be higher than without outsourcing.
2. We will do joint projects
One of the most common pitfalls is people say lets get a few people at a outsourcing vendor to be part of this project. This ican be a fatal mistake. Defining the boundary between the team in India and US is a very key part of ensuring success. The boundary should be drawn to minimize communication and dependepency: crossing the 12 hour time difference introduces inefficiencies. The only cases where it seems to work is where the Project Manager is really good and extremely detail oriented.
3. Microsoft, IBM are moving thousands of jobs, it should work for my company also
The dynamics of outsourcing are changing rapidly. I heard recently there was a rally in Bangalore to protest against jobs moving to China from Bangalore. The reality is that costs in India have risen substantially (as happened in Israel a decade ago). Real-estate today costs more in Bangalore than Silicon Valley, so does electricity. Good people cost about 60,000 USD. — which is not much different than North Carolina.
For companies like Microsoft and IBM it makes a lot of sense as their cost of employee in the US is about 200,000 – 300,000 USD per year, because there is usally a high overhead in addition to salaries in big companies.
But for a startup, where the overhead is minimal the equation has changed dramatically over the last few years.
4. It is just an engineering thing
There are companies with 50-100 people in India, where the CEO has never been to India. Even if these people are employees of some outsourcer, this simply boggles my mind. If you are going to outsource especially in software, these people are part of your future. They need to see the executive team often and require communication about the company as much as the folks in the US. If you are not prepared to travel as a CEO don’t outsource.
5. Let’s try it for six months
Outsourcing is a long term investment. If you go in with a six month philosophy it will become a self-fulfilling failure. It is the wrong thing to do, when you are out of cash to extend your runway. It is a long term thing, which requires investment of time and energy to succeed.
December 13th, 2006
Perspective from some work I did at Cornell (with S. Keshav) in the late 90’s …. YouTube has changed the video paradigm completely. But most video’s today are short clips, as the clips become longer …. I believe that the evolution of YouTube will cater to a model of “micro-viewership”:
… Out of the nearly 5.6 billion people in the world, 1.2 billion own television sets, 800 million have phone lines, and less than 200 million have access to the Internet. One way broadcast of video, therefore has a large and well established audience. However, TV broadcast is economical only when each channel is viewed by millions of people. In contrast, with the Internet, we believe that channels viewed by hundreds or even a few thousand geographically distributed viewers are economically viable. Moreover, the Internet can support tens of thousands of simultaneous channels and other multimedia content associated with video. We call the one way delivery of video content on the Internet “Internet TV”.
…
Video on demand requires quality of service guarantees as it attempts to provide VCR functionality over the network. Supporting operations like fast-forward or rewind requires the applications to rebuild the video playout buffer. This requires bandwidth and delay guarantees from the network.
[ ... which is difficult for the Internet to provide ... ]
In view of the above arguments it seems unlikely that video on demand will succeed on the Internet. On the other hand broadcast television without VCR controls is a one way transmission from the TV station to the viewers, and is more tolerant to delays in the work[1]. One-way transmission leverages the strengths of the Internet architecture and is a more realistic model to send video on the Internet.
Figure 1.1 Premium paid for the timeliness of video content
But the television network already does this reliably and efficiently. In order to be successful Internet TV should either reduce the cost substantially or offer value-added services. Internet TV has two major advantages over broadcast television. First, the pricing model of the television industry makes it difficult to broadcast programs to small viewer populations. Second, Internet TV is able to integrate other media with video enabling a large class of applications.
But the television network already does this reliably and efficiently. In order to be successful Internet TV should either reduce the cost substantially or offer value-added services. Internet TV has two major advantages over broadcast television. First, the pricing model of the television industry makes it difficult to broadcast programs to small viewer populations. Second, Internet TV is able to integrate other media with video enabling a large class of applications.Figure 1.1 shows that people are more willing to pay a premium, or be more tolerant about video quality, for watching movies in theatres or for watching events live. Although live events with large viewer populations can be broadcast efficiently on the television network, it is uneconomical to do the same for small viewer populations. Television channels have three main sources of revenue: (1) subscription fees, (2) advertising, and (3) pay per view. For channels with comparatively small viewer populations the bulk of the earnings come from the subscription fees. Only channels with 80 to 90% market penetration earn significant advertising revenue. It seems unlikely that it would be ever economical to broadcast to an audience of 1000 to 2000 viewers.
On the contrary an audience of 2000 viewers is considered large for the Internet. The Internet will have a distinct advantage over the television network in broadcasting live events to these small viewer populations. We also believe that there will be a large number of events of interest to only small audiences. The Internet already transcends national and international boundaries and united people with similar tastes and interests across geographical boundaries. Several events can therefore be broadcast live to these small communities.
Further, video from these events can be integrated with other media to build powerful applications. For example, the live broadcast of a chess championship may be accompanied with a chat-session with the leading chess players of the world. Internet TV allows video to be treated as a data type to build larger applications.
To summarize we believe that Internet TV will have the most impact to broadcast live events, video in conjunction with other media types, to a small geographically distributed audience. This thesis addresses the problems encountered in making Internet TV feasible. This chapter presents an overview of the various problems.
[1] As a existence proof note that the radio and the television networks already have different amount of delay. The viewer receives the same event being broadcast on both the networks at different times.
But the television network already does this reliably and efficiently. In order to be successful Internet TV should either reduce the cost substantially or offer value-added services. Internet TV has two major advantages over broadcast television. First, the pricing model of the television industry makes it difficult to broadcast programs to small viewer populations. Second, Internet TV is able to integrate other media with video enabling a large class of applications.
But the television network already does this reliably and efficiently. In order to be successful Internet TV should either reduce the cost substantially or offer value-added services. Internet TV has two major advantages over broadcast television. First, the pricing model of the television industry makes it difficult to broadcast programs to small viewer populations. Second, Internet TV is able to integrate other media with video enabling a large class of applications. But the television network already does this reliably and efficiently. In order to be successful Internet TV should either reduce the cost substantially or offer value-added services. Internet TV has two major advantages over broadcast television. First, the pricing model of the television industry makes it difficult to broadcast programs to small viewer populations. Second, Internet TV is able to integrate other media with video enabling a large class of applications.
December 10th, 2006
The Demo Conference (http://www.demo.com) is a great place to launch a product in the US. Vijay Anand and others have launched Proto (http://www.proto.in) with a similar concept for India. This is a fanastic thing for the Indian startup eco-system.
The first conference is to be held Jan 20-21, 2007.
December 9th, 2006
This is fantastic video for enterpreneurs:
http://video.google.com/videoplay?docid=-6909078385965257294
he has also written some interesting books:
http://www.sethgodin.com/sg/bio.html
December 5th, 2006
I arrived in the US as a graduate student in the Computer Science department at Stanford in 1993. And as part of being a PhD student, had to do my share of being a teaching assistant. Everyone at Stanford at that time was talking about companies … I would rarely find people talking about their grades. Actually I just assumed that this was normal in the US.
The contrast became very dramatic when a few years later I ended up in Cornell and there was no talk about companies … but everyone talked about grades. Actually most of the undergrad population had friends who were in MIT and their constant effort was to show that they are better than their MIT friends ….
Since then several years have passed by and I had a similar experience in Bangalore. I was sitting in Bangalore club with some of my mentor-partner friends and listens to enterprenuers pitching their ideas. And the contrast was striking … in the valley most peoples ideas start with the PC, Storage…, now the web … in Bangalore most ideas start with phone, phone and phone. When people want to do a company, they think phone. The eco-system is “early-adopter” mentality, Bharti, Reliance and others are open to launching 3rd party applications on a revenue sharing basis.
Again the contrast was sharp and stark.
December 5th, 2006
We (the think-smarter crew) are going to introduce you to a novel technique called the “blank presentation method” which will enable you to identify 70% of what should be built with 10 blank sheets of paper. No kidding. The following parable brings out the basics of how this gets done.
A story is told of a tourist in Paris in a bar one night. He got talking to a young Parisian and they hit it off. As it got late, the tourist said “I must go now”. The Parisian replied, “Let me walk you home”. “You don’t even know where I’m staying”, said the tourist. “You don’t need to tell me, I’ll take you there”, said the Parisian. Intrigued, the tourist accepted the offer. The Parisian took the tourist’s elbow and led him out of the bar. He then proceeded to walk the tourist home by exactly the same route that the tourist had walked home the previous two nights. The bemused tourist asked. “How did you know where to lead me? Have you been following me?” “I didn’t lead you” replied the Parisian. “You led me”.
How did the Parisian take the tourist home? What he did was actually quite simple. After exiting the bar, he applied slight pressure on the tourist’s arm as if to turn left. The tourist’s home was to the right, however. The tourist reflexively reacted to the pressure negatively by pulling his arm slightly away, and the Parisian was able to correct his course and turn right instead. Similarly, at each intersection, he guessed which way to go, and was able to correct his course based on the tourist’s reaction to the pressure he applied on his elbow.
Thus, through a simple process of hypothesizing where to go, validating the hypothesis, and correcting his course, the Parisian was seemingly able to lead the tourist home.
Lead, listen, learn, iterate.
December 3rd, 2006
From: Rosen Sharma
Sent: Sunday, December 03, 2006 9:04 PM
To: ‘jboudreau@mercurynews.com’
Subject: RE: India VC 2.0
John,
Read your article about India VC. I want to thank you for increasing the awareness of what is happening in India. Please don’t take the following as a criticism, but more as an informed opinion.
Everyone who goes to Bangalore for the first time does the same thing: they go an meet the Silicon Valley bank officials, meet Sridhar Mitta (with TIE) and a couple of VCs. While for the first time visitor this gives them some perspective but it is a one-sided point of view.
I think almost all US-connected firms with India funds are completely screwed up. Their structure is determined by the needs of the partners in the fund rather than the reality of the environment
Most funds in the US are built on 2% management fee and a 20% carry on the profits. So if a fund is $100M the management fee will be 2M and the carry is on the profits earned. (NEA is one exception to this – it is cost based – which allows them to raise very large funds.) On an average a US general partner in a fund makes $400K-$1M in salary per year. Thus a $100M fund can have 3-4 partners and the rest for expenses. Most US funds for India have the same or similar economics.
…
So we have funds which have USD 100M-150M under management. Such a fund can support 3 GPs. A fund is typically invested over 4 years in the US (there are several around which have a 3 yr investment cycle). If we assume the same for India, this means that 25M a year is being invested. You would invest 5M/company that’s 5 companies a year among 3 GPs which is reasonable. But consider that you can invest 2M/company, that’s 12 companies a year … that’s a bit much.
Now let me ask the key question: what is invested in each company is determined by what the company needs or what the VC needs to invest to make their model work? Why are the US-connected Indian VC’s more keen in investing in companies which can gobble up 5-10M? Why are they investing in real-estate and hotel companies? Most companies in India need an initial investment of less than 1M.
This is one of the problems: the US-India funds are based on the needs of the individuals and the funds rather than the needs of the companies.
There are several other problems with the models which exist today. Most companies over their life time in the US end up using $25M+ of cash, and companies in the India are not going to be very different. The eco-system in India today can’t fund individual companies to that extent (round C, D, E). You have people like Warburg who will do later rounds, but they only do certain kinds of deals. So if I do round A and B in India there is no bridge to US for getting the later rounds done. Infact it is going to be a struggle. In addition, most of the people who would be GP in an India fund, would not be on equal footing with GP’s in the US (not from a compensation perspective), but from a relationship, clout perspective. So if you are a GP in India and your company needs more money, can you swing the partnership into funding a tough situation: unlikely.
My opinion is that it is not the entrepreneurs in India which need to be frugal, it is the VCs. They need to have an economic models which meet reality of the market, and a cost basis which makes it possible. We have a long way to go.
Rosen Sharma
–
President & CEO
Solidcore Systems
www.solidcore.com
December 3rd, 2006
