Posted in: Uncategorized

The Problem with’s Web Services — Why Amazon?

I was reading Baris Karadogan’s response piece to Venturebeat’s criticism of Amazon’s strategy to provide “muck” for start-ups. The analogy he provided of the shift to fabless semiconductors is interesting, but I have a few other comments that are more in line with the original Venturebeat piece.

Why Amazon?
Perhaps I can’t be objective because I work at Google. I am willing to accept that idea. However, in seeing S3 and EC2, I am left with the constant question of “Why Amazon” whenever I read about these services. They are clearly useful and valuable to start-ups. But why should Amazon be in the business of providing them?

Amazon has been investing heavily in capital expenditures (to the tune of $200-250 million per year for the last two years) and its Amazon Prime shipping discount program and it isn’t clear exactly what the company has gotten out of all of this investment. Year over year revenue growth appears to be slowing (based on the most recent 10-Q filings I was able to pull down from the web) and could actually dip below 20% year over year quarterly growth in 2007. The stock continues to drift and investors are beginning to ask the company some tough questions about where all of this money is going and how it’s helping the company grow. It’s only natural for investors to ask these questions. That is the purely financial “why” part of the equation — why should Amazon be investing in these new initiatives when there are some key questions about how fast the core business is growing and there is a tenuous link between the perceived value of S3 and EC2 to their core online shopping business.

The cynical view might also suggest that one way to get more value for capital investments that haven’t exactly paid their own way would be to increase the number of users on the system, thereby generating new revenues and hopefully getting some economies of scale and the usage volume grows. Opening these kinds of services up to 3rd parties could be as defensive as offensive for Amazon.

The other companies who are notably pouring a lot of money into datacenter investments, Microsoft and Google, are coming at it from very different positions. Google is growing like mad and has the free cash flow to finance this kind of investment. For its part, Microsoft has a ton of cash on the balance sheet and can afford to invest to catch up on the Internet. So, from a financial standpoint, these companies can afford to make those investments.

More importantly, there is some perceived logic to Microsoft and Google investing in this kind of infrastructure. Both companies provide hosted services and have shown the ability to monetize hosted services on the Internet. What does Amazon get out of having SmugMug run on its infrastructure? Will it generate more business for Amazon’s core e-commerce offering? Or is the belief that these businesses will be profitable and meaningful on their own?

The problem facing Jeff Bezos and the team at Amazon is to help investors and analysts understand what these services mean for Amazon. Are they designed to feed the core e-commerce business? I suspect the answer to that question is no. More importantly does this signify a shift in Amazon’s strategy away from core e-commerce and into a new model for the company? I have no clue — if that is the case I am sure Bezos and company are keeping it close to the vest. Without some explanation, it kind of feels like an odd departure for the company, no matter how well-done and useful these services might be.

I have one other side comment. In the current environment, I am inclined to believe that web 2.0 applications experiencing rapid growth should be able to find people to fund their scalability problem. Look, if you have a web application that is growing like crazy (MySpace, Facebook, etc.) and you are facing real scalability problems, there is plenty of money out there to finance your datacenter expansion. I have a hard time believing that Digg, Meebo, or whoever would not be able to get capital to expand, and on friendly terms, if they were to go to the investment community and let folks know that they need money to keep up with stratospheric growth. What does not exist, and probably should not exist, is tons of money to finance speculative technology purchases. This is what got us in trouble last time — companies spending millions of dollars to buy datacenter capacity and bandwidth before the 1st user even showed up.

***Update: There is a very interesting post on the SmugMug CEO’s website talking about the hard savings he estimates they are getting from using S3. You can check it out here.

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to Top