Day 2 notes from Web 2.0 Summit in San Francisco, CA:
[my analysis and notes are in these square brackets.]
A Conversation with Jeff Bezos, Amazon
- S3 and EC2 are storage and processing. They should not be very interesting. So why are people so excited?
- Because the time from concept to delivery has been collapsed. We removed server hosting, contract negotitation, server infrastructure, etc.
- We have removed the “muck”; You come up with an idea, you wade through the muck, and then eventually you get to working on the fun parts.
- And then when you come up with your NEXT idea you have to go through the muck all over again.
- up to 70% of your time, energy, and dollars for web-scale apps goes into undifferentiated, heavy lifting and “muck”
- We want to swap the 70/30 to 30/70 (undifferentiated muck / differentiated work that makes you stand out). That will free up 40% for more creativity and differentiation!
- Web scale computing should be elastic; fast; always on; rock solid; simple cost effective; pay per use.
- It is being used by Xerox Global Services (S3, Simple Queuing Service and E3), Second Life; PowerSet (Natural Language Search) – now using EC2 as their back-end infrastructure;
- Tagline: “We make muck…so you don’t have to.”
- “We have been a low-margin, high volume business with aggressive cost structures for 11 years. I feel that that is the best and most defensible model because it doesn’t have a soft underbelly that people can come after aggressively unless they’re as good at it as you. “
- [For a long time I have been trying to come to terms with conflicting approaches. I now see that Dan Pena‘s model (>80% margin or don’t do it) is great when you’re consolidating an industry and need to do cashflow lending to buy the companies you are acquiring. And given that his goal is to consolidate, aggregate, and then re-sell, that model makes sense. Bezos’ approach (which is also the Dell / Wal-mart approach) is viable when the intention is to build the business up and keep it. I have to admit that I hadn’t considered that operational excellence in minimizing margin could be a defensible strategy but it makes sense now.]
- Battelle: Why are you doing this? Bezos: We think this is a great business and we’re good at it. We have three businesses: consumer facing, seller facing business, developer facing business. The consumer/seller businesses drive our revenues. The developer facing business will one day have significant impact.
- [I’m surprised that Battelle would ask this question but as the emcee, it is his job to ask the obvious and/or painful questions. Maybe that’s why he’s asking it. It only makes sense that Amazon, having developed this core competency, would turn around and resell it. Jay Abraham has a long set of principles, one of which is, figure out what your company has become really good at from a practices, systems, and technologies perspective and turn around and resell THOSE in addition to your core business.]
- Fulfillment by Amazon is a very simple service. We have a global network of fulfillment services. You can ship something to us and we’ll receive it. We get it. We stow it. And then you can send more calls to us and when you do, we’ll pick those out and we’ll send those things to you. We’re giving pay by the drink, variable cost fulfillment to the marketplace. This allows developers to use us by writing software, to treat this 10 million square foot network of fulfillment centers as a peripheral device like a printer.
- Battelle: Sun tried to do grid computing before and failed. You are succeeding right out of the gate. Why? Bezos: We have a policy of not talking about other companies. But I’ll talk about S3. We think that there are three reasons it is succeeding: it is pay-per-use, it is self-service, and it is VERY simple for people to work with.
- [The shadow space of that statement is that Sun’s attempt was none of those things…which was true. It was heavy to implement, heavy to negotiate the contracts, and relative to S3 – very inflexible. S3 is to Sun’s grid as Google’s AdSense is to Doubleclick.]
- Battelle: what is your cost of power? Bezos: Our biggest costs are not power, servers, or people. The largest cost is the opportunity cost of not fully using our facilities. And the same goes for our customers. The rest of the world is building data centers that they’re only using on average 17% of the time. People are buying 747s and parking them 83% of the time! This doesn’t make any sense.
- [This goes to the heart of the Abraham principle above. Not only does this type of service mean that Amazon is running at near 100% asset efficiency (or more, since they’re making money on them?), but their customers can run at near 100% capacity as well, rather than at 17%. Higher asset efficiency = higher shareholder value.
- [Mike Cannon-Brookes and I were joking that it would be fun to build a company and take it from concept to bizplan to website to goods to online ordering to application development to online fulfillment to peer-based service and support….in 24 hours. Sounds ridiculous but with the number and type of services now available, it would be do-able.