My partner Albert calculated early last year that it takes about 1/10th the hardware, software, bandwidth, storage, and other expenses to build a web service compared to what it took in the 99/2000 time period. That was just as services like Amazon Web Services, Google AppEngine, 10Gen, and other "cloud computing" platforms emerged as real options. It's gotten even less expensive now. As Albert pointed out in his cloud computing talk at Web2NYC, the first 5mm page views on Google AppEngine are free. It doesn't get less expensive than free.
It's a lot less expensive to build, deploy, and scale a web service than it used to be. Open source software makes it less costly and easier to build an app. Tools like Get Satisfaction make it cheaper and easier to provide support for a web app. Blogging and Twittering makes it cheaper and easier to get the word out about your web app.
There's been plenty written about how all of this threatens the venture capital model. A two person team can go to Y Combinator, get a little bit of capital, work for three months, launch a web service, and be in business. It's not just Y Combinator, there's TechStars in Boulder, Seedcamp in London, and a bunch of other such programs.
I don't think this model threatens venture capital at all. I explained why in this post back in December 2006. We've embraced the "less expensive" model in a big way with investments like Delicious, Tumblr, Disqus (Y Combinator), Zemanta (SeedCamp), 10gen, Adaptive Blue, Etsy, Outside.in, and Pinch Media. That's about half of our current portfolio and in each case, our first investment was small, the team was small (less than 10), and the monthly burn was well below $100,000 per month. In some cases, the burn was (and still is) well below $50,000 per month.
On Tuesday we had our Union Square Ventures portfolio together for our annual portfolio summit. During that meeting I pointed out that a number of our portfolio companies have figured out how to build web services that reach 10mm to 20mm unique visitors per month with a total team of less than five people. Tumblr is probably the best example of this. David and Marco are still the only developers at Tumblr. They have a customer service person and several part time members of the team. Tumblr powered blogs reach about 20mm unique visitors per month. There are about 500,000 Tumblr users who collectively generate 150,000 new posts per day. Tumblr is a big web service with a tiny team. It is incredibly capital efficient.
My question to the rest of our portfolio was simple. What can the rest of you learn from this approach? How can you get more capital efficient? And I'll ask the same thing of all of you who read this blog. Can we harness this massive reduction in capital requirements to figure out how to survive and thrive in the coming downturn?
Om Malik reported yesterday on his blog that Sequoia brought it's portfolio together this week and gave them a talk about cutting costs and preparing for a downturn. Ron Conway sent an email to his portfolio suggesting to them that they cut expenses so that they can survive another three months. I've written recently with my advice, not just to our portfolio, but to everyone who reads this blog.
I received an email this week from the CEO of a company who I have known for a long time. He and his senior team made some adjustments this week. They let a few people go, closed all of their open hiring slots, cut off low ROI marketing programs, froze salaries, and made several other adjustments. All in all they cut between
Post new comment