My Conversation with Jeff Bezos

Tim O'Reilly, March 2, 2000

My February 29 column about Amazon's patents on oreilly.com, and the overwhelming support from all of you, certainly got Amazon's attention. I got a polite but short reply when I sent my first, private letter back in January. Yesterday, I got a call from Jeff. We talked for over an hour and a half.

This is my report on that call, and what progress we made.

I want to start out by saying that one of my first goals in the conversation was not to lambaste Jeff or his company, but to listen to him. I'd given him a public eyeful of my opinions, and I felt I owed him a chance to respond. So the first thing I'm going to report on is what he says about why Amazon did what they did, and what kind of sense those arguments made to me. Then I'm going to talk about what I suggested in response, and where we ended up.

Note that comments I ascribe to Jeff or myself are not literal quotes for the most part, and simply represent my attempt to capture the gist of our argument. Note also that I've taken some liberties with the order of the conversation, to make the overall narrative clearer. We had many sallies back and forth that I've smoothed into a single thread.

To cut the suspense short, while I don't think that we're all the way to a happy ending, we had an extremely productive conversation. We each made some important points that resonated with the other, and while our central disagreement remains unresolved, I think that we've laid out the issues in a way that will lead to fruitful further discussion. I'm hopeful that you will also regard this conversation as progress, and stay with us as we try to take it all the way through to a conclusion.

First off, Jeff wanted to explain why he thought 1-click was original enough to patent. It has nothing to do with the implementation, which he admits is fairly trivial to duplicate, but with the reframing of the problem. At the time he came up with 1-click shopping, everyone was locked in to the shopping cart metaphor, because that's what you do in the real world. You pick up an item and take it to the counter to buy it. On the Web, he realized, something very different was possible: all you had to do was point to an article, and it was yours.

Phrased that way, rather than in the convoluted language of the patent, I could see that Amazon was not just cynically abusing the patent system. Jeff could in fact be claiming in good faith that he had made a seemingly small but nonetheless significant innovation. I don't know if Jeff is right that Amazon was the absolute first to do such a thing--and I'd love to hear from any of our readers who believe that they implemented such a system before Amazon--but I could see that he was sincere.

Jeff argued further that you can't just attribute this patent, as many do, to patent office incompetence. He pointed out that Barnes & Noble had a chance to present prior art in court, and that after a review of all the evidence B&N was able to dig up, the judge granted a preliminary injunction. Jeff feels that this is fairly strong evidence (coupled with the press coverage when the 1-Click feature was introduced) that the feature was an Amazon innovation.

What's more, Jeff went on, small inventions can often seem extremely obvious in retrospect. The patent literature is full of this kind of thing. The significance of an invention isn't how hard it is to copy, but how it reframes the problem in a new way.

This may be true, I replied. But it is hard to believe that if Amazon hadn't introduced 1-click ordering (if they were indeed first to do it), that someone else wouldn't have done this. It beggars the imagination that this is so significant and unexpected an innovation that others should be prevented from using it. Patents are meant to promote the common welfare, the idea being that certain ideas won't be developed without government providing a degree of protection. Is this an idea that required that kind of protection to be developed?

Jeff countered that Amazon has made countless other innovations in Web commerce that it didn't patent, and that have been widely copied. For one example, he pointed to the way that they publish sales rankings. He actually wishes they had patented a few more of these things.

Why?

Amazon started patenting because they realized that they would be under attack by players who might well one day be able to put them out of business.

Jeff tried to reframe the problem from Amazon's point of view, and the analogy he used was arresting: "We don't want to be another Netscape."

Jeff: While we may be large for an internet company, we're small in the retailing world. WalMart is 70 times the size of amazon.com, and there is no reason to believe they are hesitant to use their market power. B&N isn't doing any innovation at all on the Web--all they do is copy Amazon feature for feature, sometimes down to the exact wording. Is that right? he asked.

How would you feel if Amazon ended up being acquired by WalMart as Netscape was acquired by AOL? Don't you wonder what might have happened if Netscape had filed a few patents?

Amazon isn't the bully here. In fact, we are standing up to a bully. These are companies that have been extremely aggressive in stamping out competition in the bricks and mortar world, and they are bringing their entire arsenal to bear on Amazon.

Point to Jeff. Amazon's failure would be a bad outcome. Regardless of what we may feel about the patentability of its ideas, or about software patents in general, or about B&N, Amazon has been an innovator and a leader, as Netscape was once an innovator and a leader. Our industry is impoverished by Netscape's fall from leadership, as it would be impoverished if Amazon were to fail and be acquired by a less visionary company.

Is Amazon really in danger of that failure? The growing impatience of stock market analysts with Amazon's lack of profitability makes that less unthinkable than it once was, and Jeff may be right to be paranoid. But if Amazon's success really hangs on such a slender thread, then they deserve to lose their place in the sun.

Jeff replied that there are ten thousand slender threads, and to win against entrenched competitors, you need as many threads as you can get. Patents are only one small part of the competitive arsenal.

However, we didn't really go deeply into the extent of Amazon's danger. Here's where we did go:

Collateral Damage

We both agreed that the roadblocks placed in the way of B&N by Amazon's patents are small. After all, what did it take B&N to work around the 1-click patent? They had to add a second click for the user to confirm the order. To Jeff, this suggests that this is not a fundamental patent that is going to freeze up e-commerce. To Tim, this suggests all the more that pursuing such patents is a bad idea. They don't do Amazon that much good, and they do the software industry a lot of harm.

Tim: With the Web we've had this incredibly fertile period marked by a great deal of sharing and consequent innovation, most of it by independent developers who've learned by looking at what others were doing, imitating it and then playing leapfrog. And it is these developers whose efforts are most harmed by the fear that they may be sued by a player like Amazon.

Jeff: We aren't going after those developers. There are lots of people using 1-click purchasing on their sites whom we aren't suing. We're just going after the big guys who are going after us, the guys who are not innovating themselves but just copying us and working to crush us.

Tim: Would you be willing to make some kind of public promise that you won't be going after other people about this?

Jeff: I'll think about it and talk with my lawyers to see if there's any way we could do that without harming our suit against B&N.

Hmmm. Would that be good enough? On the one hand, better half a loaf than none at all. I urge Jeff to do this, but I am worried that it won't be enough. Jeff suspects that there is likely to be some genuine disagreement at the end.

If it's at Amazon's discretion to decide who is a no-account imitator threatening Amazon, and so subject to any defensive tactics, and who is an independent player, aren't we still in the same zone of fear and uncertainty?

My biggest worry, I say, is that of bad example. When a good guy like Amazon, who has been an innovator, starts suing over software patents, that gives cover to someone like Jay Walker of Priceline, who runs a patent mill whose explicit business model is to make money from patent fees. Not to mention companies like IBM, who by some reports are the biggest thugs around in the patent space.

Yes, Jeff, you're right that you aren't the biggest offender, and that it seems unfair to single you out when there are so many others who are doing so much more damage. You are being made an example of precisely because you are a good guy and Amazon a great company, which does care about both customers and innovation, as you said to me in your original email response. That's why it's so important that you set a good example.

I'm mindful of a wonderful article called "The Tipping Point," by Malcolm Gladwell, that appeared several years ago in The New Yorker, which applied theories of epidemiology to social phenomena. (Gladwell has apparently just published a book of the same name, which I haven't read yet, but is presumably an expansion of the article.) Gladwell's basic concept is that an infection turns into an epidemic when the rate of infection is higher than the rate at which people die or recover. There is therefore a point (the tipping point) at which a relatively small intervention can have disproportionately large effects. This is why, for example, cleaning up graffiti or petty crime can turn a neighborhood around and reduce the incidence of more serious crimes.

I didn't actually mention the tipping point in my conversation with Jeff, but it certainly shaped my argument. The Web has been developed under a certain set of informal groundrules, where imitation was the sincerest form of flattery. Those rules are under attack, as companies decide the Web is "good enough" and now they are going to change the rules under which it developed, and stop others from copying them. We have to strike now, before exclusionary and proprietary approaches become accepted practice.

At the end of the day, a culture is ruled not just by its laws but by its social norms. The social norms of the Internet and of the Open Source community, which have proven so productive in the development of the Web, need to be recognized, honored, and upheld. The public relations cost of violating those norms needs to be high.

I cited sustainable agriculture as a guiding metaphor. If you take from the soil and don't give anything back, crop yields eventually decline. If you look at the computer industry (and in fact most industries that rely on proprietary roadblocks to innovation), they resemble slash-and-burn agriculture. You strip the soil of its nutrients and move on. A fertile new field is developed, usually by people on the fringes of commercial activity, who are building for the love of it. There is a rich soil of ideas, the shared heritage of all practitioners in the field. Innovation comes naturally, since anyone can build on what has gone before. Talent is naturally rewarded.

As commercial activity moves in, the field is fenced off. Innovations are no longer contributed back but are hoarded within corporations. Imitation is still possible, but requires re-implementation from scratch, which raises the barrier to entry. Companies start to see ideas as scarce resources that need to be protected (and yes, they are getting scarcer, though no one seems to ask why), and eventually they start suing each other to keep others from copying what they've done.

We saw this in the early days of the PC industry. The open hardware platform allowed small developers to enter the market, and there was a fury of innovation until some of the players got big enough to forget the rules of the game as they'd played it when they were young. Instead, they started playing by the rules of older, stagnant industries. Apple sued Microsoft over ideas that they themselves had copied from Xerox PARC; Lotus sued Borland; Microsoft didn't sue, but they developed an awesome array of tools to keep others from copying what they were doing. It became a winner-takes-all game where financial and legal engineering took precedence over technical innovation.

The desktop software world had become a wasteland by the early 90's. If the Internet hadn't come along (a new fallow field, with rich soil laid down by twenty years of open exchange of ideas), Microsoft's victory would have turned hollow. Instead, they got a new wave of ideas, from an environment where ideas are grown, not just consumed. In short, hoarding of ideas may lead to short term advantage for your company, but long term, it's bad for the environment...an environment that you yourself depend on.

So, yes, maybe it's far fetched that something as small as the Amazon 1-click patent could be the beginning of the end. But it's a sign that the rules are changing. And the long term consequences will be bad for all of us.

Somewhere in the course of this line of discussion, Jeff's comments turned from a defense of the right to patent to increasing signs of agreement. He still holds to his need to use any weapon he has--including patents--to give Amazon the competitive advantage he believes it will need to survive. Still, the sustainable agriculture metaphor makes a lot of sense to him.

Yes, it's true, he noted, that when companies like Apple and Lotus turned to lawsuits rather than engineering to beat their competition, they were losing their edge, and it was the start of a long downward slide.

At this point, it started to feel like Jeff and I were brainstorming together about other ways than patents to solve his competitive problem.

So, we started to ask, isn't the right way for Amazon to proceed to just keep out-innovating those big, hostile companies? How can Amazon develop strategic advantages by what they give away rather than what they try to keep others from using? How could they become more active in giving back to the innovative developers who will keep giving back to them?

Amazon has already become the reference site of record for the publishing industry, the site that everyone uses to search for information about books. How about ways to make Amazon even more useful for research as well as commerce, with a public API for cooperating applications? Wouldn't that be the Internet way, to keep yourself at the center of the expanding universe of new possibilities, rather than trying to close the gate behind you?

I won't go into details of all of the possibilities we talked about. The conversation was exhilarating, and I really felt that Jeff is sincere about wanting to think about the problems I raised with his approach, but in the end it is actions that will speak.

Right now, the immediate actions that we agreed on were three:

  1. For Amazon.com to think hard about what kind of assurances they could give to independent developers about their safety from patent lawsuits.

  2. To continue our conversation about patents, openness, and software innovation.

  3. For me to write up our conversation so far and share it with the public, even in its inconclusive state. After all, one of the rules of the Internet, as articulated so brilliantly in The Cluetrain Manifesto, is that a market is a conversation. We don't have an answer yet, but we're talking.

I'd like to invite you into this conversation.

What do you think about Jeff's assertion that Amazon is under attack from ruthless competitors who might have the strength to bring them down? Even if this is true, is a patent lawsuit a legitimate way to defend against them?

What kind of public stand would you like to see Amazon take on patent issues?

How could Amazon best support further innovation on the Internet? How would you like to see them give back to the developer community?

Thanks for all your support, --Tim O'Reilly

Return to: Ask Tim Archive