October 2000

Date: Thu, 05 Oct 2000 08:53:07 -0700
From: "Tim O'Reilly" <>
Organization: O'Reilly & Associates, Inc.
To: Computer Book Publishing <>
Subject: Re: bookster

David Rogelberg wrote:

What will happen is that content will become the ultimate branding builder. Publishers will then have to figure out how to leverage their brands to sell services and products.

For example, if I were Tim O'Reilly, I'd probably buy or start an OpenSource consulting/development company, and give my books away. He's got the brand, the credibility, and could really play in a much bigger game than publishing.

We're lucky, then, that you aren't me, because if you were, O'Reilly would be out of business fairly soon. (We are playing in a bigger game than publishing, by the way--see our conference business, our online publishing business (,,, our spinout of high tech startups (from GNN, sold to AOL in 1995, to LikeMinds, sold to Macromedia last year, to Collab.Net, now a well funded startup on track to an IPO in the next couple of years)--but our book business continues to show tremendous growth. We have more revenue from books than Red Hat (also a great brand) has from all the open source services they've been able to muster with IPO-fueled turbocharging. In fact, in this model, "services" means something much closer to what an ISP does, to what people like LinuxCare or even Red Hat have been exploring. The evolution of internet services from free peering (e.g. usenet over uucp) to pay by the month services demonstrates how people will end up paying for "content dial tone". See my extended discussion of this point.

I'm with Harry Helms. While we are training people to expect content to be free, we're also training them to expect that they'll get it on a cable-tv style model. What is an ISP, after all, but a monthly service for pay for software that is given away for free, and services that almost all users of that software used to get by swapping among themselves?

I predict that we will see a lot of content aggregation into paid services that will either be offered by new players, or will be offered by a variety of old players who already have billing relationships with customers.

At O'Reilly, we're following our own advice. Our Safari ebooks service, which will be available later this year, will provide "O'Reilly content dial tone" for a monthly subscription fee. Yes, people will be able to pirate the material if they try hard enough, but we're betting that enough of them would rather pay us and get reliable access.

I also want to repeat my old adage that the math doesn't support peer to peer as a standalone operation. Millions of products trying to reach tens or hundreds of millions of consumers means billions of possible interconnects. Peer to peer has application in some very important niches (especially in the formation of small, ad hoc peer groups, whether of people, programs, or devices) but it won't take over the world just because of scaling. Web links are pure peer to peer--anyone can link to anyone else--but within 5 or 6 years, we've seen the emergence of major intermediaries to help people find things. Web publishers now have "sales people" who work to get web pages listed in Yahoo and search engines, much as we have sales people who try to get our books carried by large retailers. For some other thoughts on this topic of disintermediation and reintermediation, see$8.

In this kind of context, what peer to peer does is push centralization "up the stack." That is, where we used to centralize files for download, now we all host our own web pages, and pass around links. But then we find that we need to centralize those links. The difference between and napster is simply this: one provides a directory of data, the other provides a directory of metadata. (For more info on the work we've been doing to explore p2p, see

A further point about Napster: at bottom, it is a parasitic technology, and can't stand alone. Why? Because it relies on the music industry to market its metadata, its "name space." How do you know a song exists? Sure, there are opportunities for referrals by friends, and so on. Word of mouth is definitely important, but as any publisher knows, it isn't the whole enchilada.

Further, the namespace that napster searches is intuitive and well known. Only two pieces of metadata are required by the user: artist and song. (Napster adds a few more: the location, the file size, and the bandwidth available, but the user doesn't have to know these things.) Now imagine something more complex: even in the music world, classical music doesn't fit in the namespace. I want the version of Tosca by Puccini with Maria Callas and Guiseppe di Stefano, not the one with Renata Tebaldi and Guiseppe di Stefano. Now apply that to books. Now tell me how you're going to search a napsterized online content space for ebooks? You need services that help people organize the name space (i.e. publishers and distributors, of some flavor), and you end up paying for those things in one way or another.

The idea that all content will be free is a myth that is spread by people who are trying to sell book security services, and popularized by people like the RIAA. I predict that the impact of ebooks is that content will be available in more forms and formats, and that we'll find ways to monetize many, if not all of them. Money is about exchange of value, and if you provide value, I guarantee you that ultimately, you can figure out how to get people to pay for it, because if you don't, you stop producing it.

Others have pointed out that tv didn't kill movies, VCRs didn't kill movies, etc. etc. I heard a great quote yesterday from Jon Potter, executive director of the Digital Media Association, at a conference we were both speaking at in Washington. He said "You have to look at for the interests of the content creators and the content consumers. If you do that, everything in between will eventually sort itself out." (This was advice to Washington policy makers that I heartily endorse.)

A final point: it has been possible to download pirated Microsoft software for years, but as we know, that hasn't stopped Microsoft from making an ungodly amount of money. The industry originally tried copy protection, but then they came up with a better answer: accept some level of piracy, but strengthen the social norms (and the legal consequences) against illegal copying. In this regard, I find a statement I heard from Jonathan Schull of Softlock, which worked with Stephen King on Riding the Bullet, to be instructive. Softlock provides encryption software that encourages "superdistribution" (i.e. people passing along their copies, which can then be unlocked by the recipient upon payment.) Nonetheless, almost all people got their copies from central download sites. Softlock did a poll to ask why. "We thought it was wrong to pass it along" was the most common reason given, despite all the efforts Softlock and King had made to encourage passalong.

In short, this is a complex issue, but one that will sort itself out. No need for panic or for pontifications about the end of the world as we know it.