Web 2.0 – Tim O’Reilly and Jeff Bezos
(not a transcript!)
10/5/04
Tim:
Apple – rip, mix, burn
Rss and web services allow us to do this with the web.
At what point does this trip over business models?
Jeff:
If there’s real value, businesses will be able to find a way to charge for them and to pay for them. We will charge for Alexa when it comes out of beta, but we don’t know how much yet. We need to explore that. There will be business models and they will evolve with the web. People will always be able to figure out how to make money.
Web services is a special case because there’s a built in business model.
Tim:
Ebay and Amazon lucked out because there’s a built in business model for them. Google has not, which is why news.google.com is still in beta. We first met over the one-click patent issue, which is one unique aspect of what could be exposed as a web service.
Jeff:
The shopping cart is available now. We’ve thought about one-click.
Tim:
You have a digital identity system with millions of customers in a trust relationship.
Jeff:
How do you charge for that? We have to experiment, but I don’t see a reason not to expose more things over time.
We find the useful guts in Amazon and expose them to see what other people do to find useful ways to use it that surprise us.
Tim:
Amazon has continued to innovate. We could describe you as a change junkie. You guys have nailed e-commerce and yet worked hard at becoming more than that with everything we’ve talked about. Not a single other site invites users in as much and that has paid off. You were early to embrace Amazon as a platform, so what’s the secret ingredient?
Jeff:
Companies need to find unique aspects which users can enjoy and figure out how to make that into a business model. The community aspect with customer reviews and list mania influences the user experience. We have never had a business model with strong network effects. Customers are loyal until someone offers them a better service, so we must keep innovating. Architecture participation is the way for innovation. We invite customers to participate in the website. Anything, whether network effect or not, which creates improvements in customer experience may be a huge effort, but is worth it in the long run.
Tim:
Your search became harder with search inside the book because the results would show content inside of the book instead of keyword matches on a title.
Jeff:
We can measure accurately whether search algorithm changes create an improvement or not by testing simultaneous database results to see which generates more sales. Sales is a noise free indicator. But we don’t mind robots with credit cards either. We’re not guessing when we make changes.
Tim:
I’d love to see the data. We’ll get to audience questions soon. Ok the x-prize has been won. And lots of Internet entrepreneurs that made their fortunes young and are now doing kickass stuff. What makes this generation have a different bent on what to do with their money?
Jeff:
Oh God.
Tim:
New passion to change the world.
Jeff:
People follow their passions. I want to get up there.
Gary Wolfe – Wired:
Search inside the book is exciting because of the underlying data gathering function in scanning the data in. There’s an issue going further in the sense of data which is proprietary or not available. What type of data do you long for which could be the basis for web services to really work. What data do you long for that we can’t get to now?
Jeff:
The type of data which is not in digital form but is valuable and proprietary. There are 2 kinds of problems. Search inside the book addresses non-digital form. And the other is a business model issue, but it will get easier over time. Real editorial team and real effort are involved in creating this valuable content.
GW:
How can we get at that data?
Jeff:
SIB is allowing people to sample at the point of sale. At the point of sale is important because sampling increases sales. It’s an effective technique. Physical bookstores allow you to sample books. iTunes is an example of this working in the web world.
Tim:
There’s a third data problem, which is data which is either digital or not, but where the rights are ambiguous. In the early days people would play fast and loose if there was no real value. With music, at first it was ok, but not it’s not. People clamp down on value.
Q:
To improve the experience for the least common denominator, is html cross-browser better or something outside the browser?
Jeff:
I continue to see the browser getting better and better. In A9 you can drag and drop bookmarks into the bookmark button. You can drag all the components around. JavaScript makes this all possible. We’ll see more and more robust but lightweight apps on the web.
Tim:
An example of a plugged in non-browser experience is iTunes.
Q:
In 10 year chunks or more, capitalism aside, do you see the emergence of massive online universes? What’s the intersection for you?
Jeff:
I’m a big Gibson and Stephenson fan and I’m in favor of such universes and I think they’re likely to happen. You’re starting to see them in the SIMs and MPGS and how involving they are. People spend huge portions of their lives in these universities.
Tim:
In Hong Kong there are companies similar to sweat shops where people play EverQuest and see the identities on Ebay.
Jeff:
Maybe we’re in the universe now (Matrix).
