In a recent Fast Company article, Jeff Bezos described his idea of "a decentralized, disentangled company where small groups can innovate and test their visions independently of everyone else. He came up with the notion of the 'two-pizza team': If you can't feed a team with two pizzas, it's too large. That limits a task force to five to seven people, depending on their appetites." Bezos led off the Web 2.0 Conference with a look at what those two-pizza teams have accomplished and a guess at what the Web 2.0 might look like.
Bezos began with a look back at what he called Web 0.0 by looking at the original Amazon.com home page. It is hard to look at this first web page and imagine what drove it to the current incarnation. The page was completely static HTML authored entirely by Bezos. The search box, which sits on every Amazon page, was not on the original front page--it was a click away.
Contrast this vision with the current Amazon.com homepage, which Bezos describes as representing Web 1.0 because "Humans create the content but machines place it." He notes that the whole site can be customized for a logged-in customer. Web 2.0 looks a little different for human viewers but dramatically different for networked devices. Bezos asserts that "Web 2.0 is about making the Internet useful for computers." He points to his company's web services page for an example of where we are heading.
Making the Web more useful for machines requires that you consider how a platform may be built around your offerings. Bezos predicts that you will see a lot of APIs opened up over time. Amazon.com started this two years ago and already has sixty-five thousand developers who have access to everything from customer reviews to product images. He says that, so far, they do not charge for web services, and instead, they reward people with commissions to encourage developers because the third-party applications built on this platform help Amazon.com sell more products.
Sites that he demonstrated include musicplasma, a music visual search engine; ScoutPal.com, which allows you to scan a bar code and find what the book mostly sells for on Amazon.com; and A9, Amazon's new search which is built on their own web services as well as those from Google and others.
Bezos explained that one of the driving reasons for Amazon.com's opening up their APIs to others is that "there are so many more in the world than there could ever be working for Amazon.com." Data that underlies an application like musicplasma is real time and changing and is expensive to collect. The third-party vendor benefits from having this data available to them, and Amazon.com benefits from having developers working on applications that they could not justify as an in-house priority.
In cases such as musicplasma, the business model is clear: "You click on links and you are taken to the Amazon link where you are encouraged to buy the album." Exposing more of the data and functionality of the Amazon sites is an ongoing process. Bezos wants to find the "useful guts," but still needs to think through how to charge for services and how it should all work.
The payoff from Web 2.0 is the amount of productivity that will be generated from these applications. Bezos concluded his presentation with the note that "even though it's Web 2.0, it's still day one and the snooze alarm hasn't gone off."
Tim O'Reilly returned to the stage to lead and moderate the question and answer portion. To paraphrase Apple, O'Reilly said, "web services allow the rip, mix, burn of the web sites." O'Reilly asked Bezos at what point a consumer of web services might be infringing on the provider of the service. Bezos agreed that this is an important question, and that part of the answer is to figure out how companies such as Amazon should charge for the web services they are offering.
He offered Alexa as an example of a web service with a lot of value. Even though it is currently free while it is in beta, users are already being told that there will be a charge when it comes out of beta. Bezos cautioned that "giving away your new assets for free is not going to go very far."
In addition, having a transaction tied to a web service attaches a concrete metric to that service. For example, O'Reilly asked about recent changes to search, which seemed to result in less effective searching. Bezos was able to counter that by saying that comparing the sales generated from two methods of search gives an objective way to compare two search algorithms. He added, "When we look at our search algorithms, we aren't guessing; we know."
The opening session concluded with a question for Bezos about what kind of data could be available that isn't yet. The answer fell into two categories: the data that isn't in digital form and the data that is locked up in a proprietary database. He argued that the Amazon feature, "Search inside the book," begins to address data that is locked up on paper. He said that the other problem is a business model issue. For example, there are low-circulation highly specialized newsletters that have to charge a significant amount to cover their costs. These publications have to protect their data, so some data may necessarily remain unavailable.
Bezos returned to the business value of allowing people to sample at the point of sale and search inside the book. He noted that the simple approach is that they are allowing people to sample at the point of sale. He told the audience to look at physical bookstores at which you are allowed to sample books. Imagine the sales at a bookstore where every book was shrink wrapped shut.
Daniel H. Steinberg is the editor for the new series of Mac Developer titles for the Pragmatic Programmers. He writes feature articles for Apple's ADC web site and is a regular contributor to Mac Devcenter. He has presented at Apple's Worldwide Developer Conference, MacWorld, MacHack and other Mac developer conferences.
See more Web 2.0 Conference Coverage.
Copyright © 2009 O'Reilly Media, Inc.