Manager who were driving down a mountain in a van and shortly after leaving the top of the mountain, the brakes failed on the van. It started going faster and faster, getting more and more out of control. But luckily before it went too fast, the driver managed to crash into a tree and the three men got out the van shaking but unhurt and confronted their dilemma, How do we get down the mountain? So the Project Manager says, I know, I'll point a committee and the Hardware Engineer says, don't worry, I brought my Swiss army knife, I take the van apart and put it back together by morning. It will be as good as new. The Software Engineer, said Gentleman, Gentleman, don't you realist the first thing we do in a situation like this is we push the van back up to the top of the mountain and see if it does it again. Now of course, what's amusing about this story is that the Engineer is treating the auto crash as if it was a software crash. And everybody knows what you do when your software crashes. You press control delete and see if it does it again. So, of course, this illustrates the point that it's a good idea to do that, in the case of software its not such a good idea to that, in a case of an automobile cash. So bits are different than atoms. Its almost a coshay these days, they don't wear out, they are easy to copy, cost is to create and so on. And these differences have led some people to say that we need a new economics to understand this new economy of bits. But the pieces of our book that Stew just described, was we think that's not true. We think that many of the important aspects of the information economy were also present in the industrial economy, it's just that you have to know where to look. The emphasis is different because the facts that were rare in the industrial economy, things like increase of scale, network facts and so, are actually common place in the information economy. So our claim is that you don't have to rely on speculation or to understand the new economy. You can learn a lot from history, case studies, antro-mechanomic analysis. So our book, Information Rules is, as Stew told you, been quite successful. It was endorsed by Andy Grove, Chuck Bases, Scott Cooke and several other leading entrepreneurs in Silicon Valley, was Editors choice for book of the year, business book of the year in Amazon and right now, its in the top 100 on the overall list and in the top 10 on the business list. So I ask Stew, how long I should talk, I said, would 45 minutes be optimal. His reply was if 45 minutes would be optimal, that 30 would be even better. So what I would like to do is just summarize briefly some of the important themes in the book and I have chosen 5 main themes. One is differentiation of products and prices. This is his idea of personalization, customization and beginning, as we call it in the book. I want to talk a little bit about rights management, studying the terms and conditions for your electoral property and then I want to move on to talk about switching costs and lock in, its kind of a nurture that we get, we hear a lot about how we are entering a friction free economy, well that's true in some ways but as you will see, there are still a lot of effects that lead to a nurture and switching costs.

I do want to talk a bit about networks and positive feedback, one of the important aspects of the information economy is that it has a lot of networks in it. In many cases these networks exhibit the effect that economist called "Network Extranalities" were the value of joining that network, depends on the number of other users, the number of other people that are in that network and that leaves to some very interesting industrial dynamics. And finally, I am going to talk about standards and inter connection. Nowadays you don't just have to worry about your competition, you also have to worry about your complimenters, in a very real sense you cant compete if you aren't compatible, so a lot of business strategy has to relate to what your complimenters are doing as well as your competitors. In the book we discuss a 6 topic, a topic of inti trust and competition policy. My co-author was a chief accountant for the department of justice. We told our editor we needed to have a chapter on anti trust; she was really quite reluctant to go along with it. She says, well does that really belong in a book on business strategy. Our view is that if you follow the principles we outlined in the other chapters to be so successful, you would have to anti trust and she thought that was a convincing enough argument. So the first two topics are mostly to do with information, that is content and then the last three topics, really have to do with the technology side of things or the conveyer of that content.

So, I want to start now talking about differentiation of products and prices and I want to talk about this in the context of information goods. So information goods is a term that we use for referred to anything that can be digitized, sent over the internet, text, images, sound, database, movies, multi media, everything we see in these wonderful demonstrations just a few minutes ago. And those kinds of information goods have several important economic characteristics. One of the most important is that they have very high first copy costs and then very low incremental costs. So it's very costly to create that first copy of a book, movie or a digital object. It's very cheap to create additional copies. And now all the first copy costs involve high fixed costs, those costs are also, what we call sum cost. That means that they are not recoverable costs. So if your factory goes out of business, at least you can sell the factory and equipment. If your movie is a flop, there is not much of a second market for the script. Furthermore, in general in these cases there are no lasting capacity constraints, it can be very expensive to expand an industrial capacity but if you want to produce a few more CD-ROM's, that's essentially almost costless activity. So the problem is that once you sucked the cost to develop your information goods, there's really no price war that you can rely upon. It is very easy to get in a situation where the information becomes commodities and then the price gets competed down to marginal costs, which is almost zero. We have seen this happen in several cases. It happened in CD phone directory business. The very first CD phone directory, which covered just the 9X region, sold for $10,000, that £10,000 a piece or one CD-ROM. That was fine when there was a single producer but after business starting to expand, there were several producers selling essentially the same commodised product and it was competed down to just a few dollars in just a few years, and now of course in the US, at least, phone numbers are available on the internet for nothing. We also saw this in the encyclopaedia business, where the Encyclopaedia Britannica was selling for £2,000 in its print version, Microsoft came out with a $50.00 CD ROM version. We saw extremely intense competition in encyclopaedia market, pushing the prices down to the $50, $60, $70 range with a lot of discounting going on. So commodisation of information can lead to very low prices and the first challenge that faces anybody that is trying to produce this digital information is hard to avoid that kind of commodisation. Let me come back to that in a minute.

But the next question, the next challenge of these digital goods impose on business processes is that, not only do we have this reduced production cost for producing incremental copies of information, we also have dramatically reduced communication costs. That means you can sell to a much broader, much larger audience but in general, those new customers are going to have different demand characteristics that the customer is currently serving. Typically, for example they would be willing to pay less for your product that the customer's that were in your traditional base. So think of an example like legal information that sold to Lawyers, that's one market, now you can put legal information on the internet, so a much broader market, but they are not going to be willing to pay as much as the Lawyers. Same thing goes for medical information or almost any other sort of specialized information content. So the challenge there is how do you sell to that new market without cannibalizing the old market.

Well, we have an answer to both these questions and the answer is, what we call versioning, so versioning is the idea of differentiating your information product in creating a product line of information goods, allows you to segment your market and sell to those different market segments at different prices. Now this is pretty standard in industrial markets. If you look at automobiles, they come in product lines, you look at washing machines, they come in product lines but, the first time you start talking about product lines of information goods, it sounds like a strange concept. When you think about it, it's not so strange because it occurs even in traditional markets for information. Taking books for example, books come out first in hardback, and then 6-9 months later if successful, they come out in paperback. So they are what you use as delay and a slight change in format to segment that particular market or particular movies, it could play first in theatres and then 6 months later, it plays in video. There are several examples of this on the Internet that are quite interesting I think. One example is stock prices. If you want to get 15 or 20-minute delay in stock prices, you can get that everywhere, virtually nothing. If you want a real time bead of stock prices, it could cost you $50.00 a month. Or images are another nice example. There are lots of stock photo distribution sites on the Internet where you can purchase images. If you want a high-resolution image at 600 dots per inch, suitable for glossy reproduction, that could easily cost $50.00. If you could settle for 300 dots per inch image, that might be $20.00 and of course, they give away the thumb nail version virtually for free.

Another nice example in the software market take mass-market software like Quicken for example. Quicken comes in a high-end version and a low end version or Mathmatica, which is software for processing mathematics. They use to sell a student version and a professional version and the difference was that the student version went about 10 times slower than the professional version. Now there's a kind of nice twist when you dealing with information. If you look in the industrial economy, the way you segment a product line. Either segment your market and create a product line is you add value. You take a Camrey and you add some leather seats and some chrome and you end up with Lexis. So you add value to create those high end products. But the information economy, what's interesting about it is you normally subtract value. So think of the examples that I named, the stock prices. We have to create the real time feed and then what you do is that you delay it to create the 20 minute delayed fee or the photo example. You have to create the high-resolution version first and then you, from that degrade the image to create the low-resolution version and the same thing with the software. What they did in the mathematical case, is they created the fast version and then they disabled the math co processor and just did the folding point calls and software in order to create the much slower version for the student market. And of course, you can see this, another nice current example is with windows NT, they have the work station version, which is about $250.00, they have the server version which is about $500.000. The difference it turns out is a configuration file that basically slows down the other version. So what we do in the book is we go through a lot of these examples of different dimensions that you can verge in your information goods. We look at ways to target different market segments and we give some nice rules for how you design a product line of information goods.

Now that kind of leads nicely into the next part of my talk which is about rights management and one of the implications of this reduction in the cost of copying and distribution is not only does it make it better for you to produce the information, but the unfortunate thing is it makes it better for the users of information to illicitly copy and distribute your content. So you have to set the terms and conditions under which your intellectual property can be used. For example, should you allow copy, should you allow sharing, should you allow links, should you sell your software or your images or should you license it?

So there are many different choices in terms of designing your information goods and the terms and conditions under which it can be used. We contend the content owners are usually too conservative with respect of rights management and the tendency is try to preserve the old business model, rather than to construct a new one. And we illustrate this with several different historical examples and then some current examples taken from the internet today. So let me just give you a little bit of history here.

Take for example, books - kind of a nice example. Back in 1740 the first English novel was created, Pamela. It was so popular that the booksellers couldn't keep up with demand and they started writing out the books. Now the publishers hated this idea of book rental because they were afraid that it would cut into their market books, sold for a very high price. Basically, the first edition of Tom Jones was about the price of a working man's weekly wage. But, luckily the book renters, the circulating libraries as they are called were quite successful, over 1,000 of them were established and increased the literary rate in England quite dramatically, so that the number of readers in the 1980's increased very dramatically and in the end it created many more sales opportunities for the book publishers. Now, perhaps you think that's ancient history, but if you think about it, essentially the same thing happened 200 years later with the video machine, back in 1980 the video machine was just a toy for rich people, it sold for $1,000, the content was $100.00 a tape and somebody got the bright idea of renting those video tapes. Hollywood was aghast at this idea, they did everything they could to try to prevent the video rental market but what they didn't see was that the low price content led to an increased in demand for the video machines and the increase in demand for video machines led to widespread demand for more content and now of course, what we see is that for most movies, Hollywood makes more off of the video home view sales than they do off of the theatre distribution. So, it created a mass market for video in the same way that they created a mass market for books a few 100 years ago. A most recent example of something that is going on just today is MP3, this audio compression format which allows you to compress a top song down to about a mega bit, small enough so you can download it over the internet. Well last year the music industry was very unhappy about this because people were basically stealing content, taking CD's, copying the content, putting it in MP3 files and downloading on the Internet, so they went through all sorts of things to try and prevent this from happening. This year they have realized the big problem is not so much the copying but the bypass. That is, totally bypassing the distribution channel. There's the side of the internet www.MP3.com where the recording artists are going directly to that site, posting their own songs as a means of advertising to promote their albums, people were downloading those on the internet, therefore bypassing the distribution chain that is controlled by the music publishers. So the big problem there is not the worries about so much about the rights management, the big problem is really the business model, and they have to come up with a new business model to meet that particular challenge. We also saw this in the web for texture material. If you look at new stories which are almost a commodity item, use to the last year there was a little rights statement at the bottom that said copy right, all rights reserved, do not re-distribute and so on. Well this year, this had been replaced by a little button at the bottom of the story that says click hear to e-mail this to a friend. Total change in terms of how this content is distributed, not so much a worry about intellectual property but more importantly a worry about the business model of how they are going to most effectively distribute this content and realize the competitive advantage you get from a new medium.

Let me turn now from talking about information goods to talking about the enabling technology. I want to talk about the idea of lock in. So information technology is typically part of the system or network and these networks can be real networks like the telephone network, fax network, internet and so on. Or sometimes they can be what we call virtual network, so you think of the network of windows users or the network of Mac users, and so the network users of IBM frame users. And so what happens is the components typically work together. So a fax machine isn't very useful unless other people have fax machines you can send faxes to. Telephone isn't useful unless other people have telephones and of course, if you look at computers and applications and the hardware itself, this is not worth entertaining without applications that can run on that hardware. So we see these systems of hardware, software, servers and browsers, DDD's and players, all sorts of complimentary goods that fit together. Take a word processor for example. You have to have the computer, you have to have the software and the most important component of all is the web ware. The stuff hear, the skill, the knowledge you can build up in using that word processor to produce effective documents. What that means is that the cost of switching any one component of these systems may often involved switching the entire system. If you switch your hardware, you have to switch your hardware and you have to switch your web ware. So in many cases those costs of switching were so high that they could leave to technological lock in. You are pretty much stuck with this technology choice that you made. So that leads to a kind of challenge for both buyers and sellers of information technology systems. The seller of course, likes to have a loyal customer base; they like to lock people into their technology. The buyer wants to avoid lock in but in many cases you really cant avoid it so, what you try to do is get compensated for it up front. So there's a, we call this a dance, an intricate dance between the buyers and sellers and that occurs at the individual level, at the organizational level and even at the societal level. So how can you use this kind of lock in or switching costs in terms of your business? When we go through this material we look at both sides. We try to look at buyers and at sellers. Take the seller side here. One thing you can do is you can try and sell complimentary products to your installed base. A nice example there is HP, Hewlett Packard has a very good business in making printers but what's not so commonly recognized is about 25% of the profit in their printer business actually comes from the sale of B-jet cartridges. If you look at Iomega, which makes disk drives, proprietary format for disks, what, happens there is almost all of their profit comes from sales to the media. This is as old as the story of giving away razors in order to sell blades but it's critical for understanding how to best exploit opportunities in the information age. Another thing you can do is, if you don't want to sell products yourself, you can sell access to your installed base. So a nice example there is Dell, they are selling on site system administration for their computers but, now you see what happens when you are selling both the system administration and the computers and changing any one component to change the computers, you have to get new system mains. To change the system mains you have to get new computer, which dramatically increased cost of the entire system that you are using in your business.

Another example is that your telephone, your wireless, your cable and your internet access often on one provider, but of course, it also makes you much more loyal. And it makes you loyal in the sense that the cost of switching to an alternative supplier for any one component can be quite large. So we hear a lot about how friction such as search costs are being removed. In the internet we've heard this idea that friction free economy. The argument is that you are going to see an increase in frictions in some dimensions because, Companies are going to want to add artificial lock ins to create this kind of environment where they can make their customers loyal to a particular supplier, so a nice example of this is providing complimentary services like assistant administration I mentioned earlier. Another example that we think is going to be very important is the idea of loyalty programs. You want to offer discount, bonuses, coupons, all sorts of extra rewards to consumers who are large purchasers of your product or who are particularly loyal long-term purchasers. And of course, this is something that is enabled on the information technology because its so easy, its so cheap to track use, you can reward your loyal purchasers by increasing their switching costs. Making it more costly to switch to an alternative supplier because they loose the accumulative usage that they build up by staying with you.

Now this leaves me to the next topic I want to talk about which is positive feedback and network effects. At the beginning of my talk, I mentioned the idea of network externalities, the idea that the value of the network depends on the number of users and its kind of interesting to look at where this concept entered economics. The very first paper written on this was in 1973, over 25 years ago. It was in the Bell Journal of Economics, written by an economist named Geoff Rolls. He went to work for Bell Labs and they asked him, well could you come up with a model of why picture bone was such a flop and so he worked out a little model. He said well, if the number of adopters with the value of the service depends on how many other people adopt, it turns out that there's a critical mass, if you can get to that critical mass, you rush on to success. If you fail to get to the critical mass, you just fall back towards the zero market share. It was a very nice paper; quite interesting model but nobody read it. Nobody read it for about 10 years and then my colleague Karl Supero, along with another colleague Mike Cass, wrote a little model in the early 80s about how to use what were the strategic implications of this. If there was a critical mass, you had this positive feedback, what could you use to get there and avoid that sink of failure. Well it turns out it was very nicely timed because in the 80s, there were a lot of successful networks and the difference between Rolls paper written in the 70s and the Cass and Supero paper written in the 80s, was really nobody cares much why the technology failed but they were very interested in why technology succeeds, and what we saw were fax, ATMs, videos, internet, web, e-mail, all of these networks that have thrived in the 80s and early 90s and now of course, everybody is interested in this network idea. So the value of the network to a user depends on the number of other users and Metcass Law, for example in the Engineering literature says the value of the network grows by inch squared, so both of these ideas, network externalities of Metcass Law are describing essentially the same phenomena. Now, what's interesting about this in the case of information technology, is not only do we have this, what we call demand side economies of scale. How much your product is worth, depends on the number of users. We also often have supply side economies of scale. A very large cost to produce the first unit very cheap cost to produce incremental units. So we get a kind of double laming out of this. It's both the impact on demand and the impact on supply that matters. Supply side economies of scale, says the bigger you are the lower the unit cost. Depends side of economies of scale says the bigger you are the higher your value to the users and accommodation of these can be particularly intense.

We've heard a lot this morning about first mover advantage. There really is an advantage there, both on the demand side and on the supply side, so it's important to get in there first. We talk about a lot of strategies for pre-emption in this context. We also talk a lot about expectations management, because for these technologies in a very real sense, the technology that people expect to win will win. So it's important for you to manage your expectations to convince your users that you got the winning technology. So last year, there was a very nasty standards battle going on between two proprietary standards for 56 kilobits modems, there was the XA standard and the 56 flux. We poured out a lot of ads from the computer magazines. Each of them claimed to have an 80% market share and what they were trying to do of course, was manage expectations. Same thing was with the Netscape and micro soft browser wars. The statistics on browser use reported breathlessly in the Press as that will continue.

Now what happened in that 56 kilobit per second model wars. It's quite interesting. It ended in a truce. What happened is that they agreed finally on a common standard, and its important, we have heard several standards were mentioned today because what standard do is that they allow these networks to interconnect. So, what they do is they change the nature of competition from competition for the market to competition in the market. When there are 2 different standards for these 56-kilobit modems, they are each competing for the market, once they agree on a common standard to be keen competitors in the market.

Now, there are very strong patterns of behavior that apply to both real and virtual networks. And the critical thing is to look at this issue of interconnections, so the final topic I want to discuss is the standard in connection.

It's also quite interesting to look at a little history there. Nice example is to look at Marconi wireless. As you know radios were originally viewed as a point to point medium. It wasn't until substantially later that people recognized that you could do broadcast. Why would anybody want to send a radio message to no one in particular, because they didn't realize that there was a way to support broadcast radio in the advertising. But the Marconi International Marine Course was quite interesting when they did this point to point, when they were primarily selling radio for point to point purposes, they wouldn't sell the equipment. They would only license it because they didn't want to let the equipment outside of their control and they would not connect with any system. You could only send Marconi signals to Marconi radios. They subsidized access to key clients, for example newspapers, especially cheap access to wireless services and they engaged in exclusive viewing for example, they had a deal with Lloyds of London where Lloyds would only insure ships that had radios but not just any radio, they would only insure ships that had Marconi radio. Now, finally in 1912 the US passed the Radio Act, which made inter connection compulsory that was for 2 reasons, one because of the shipping accidents - The Titanic had something to do with it - and also because of the political situation. They wanted to ensure that the military radio transmissions would be carried even if they had to go over multiple networks.

Another very interesting story is to look at the history of the Bell System. In 1907 when the original patterns in the telephone expired, Bell got a lot of competition coming in. How did they deal with that competition when they invested in long distance and then they borrowed a very particular inter connection strategy in towns where they were not competing, where they did not have a local presence, it was so access to the long distance service in towns where they did have a local competitor, they would not provide access to the long distance service, giving them a tremendous advantage. They had much bigger network externalities because they had access to that larger network and what happened of course, if that they eventually prevailed in dominating these other competitors. So interconnection is usually bad for incumbence, that why Marconi and Bell have tried to avoid inter connection, its usually good for consumers entrance and complimenters.

In the book we look at two sorts of generic strategies in dealing with this inter connection issue. One is standard wars, examples like Beta and VHS, the 56-kilobit modems, CDMA versus GSM, Netscape versus Internet Explorer, possibly examples of that sort. Then we also look at appliances, things like Job Ads, CDs and BDB's where there is alliance along different competitors or complementors, who each bring something to the table in turns of the standard negotiations. And the critical concept to keep in mind there is a very simple relationship that your share of the profits depends on your share of industry profits times the total value of the Industry. So that's just trivial, an identity that your value is your share times the total industry value. But the critical thing to keep in mind is that some of the choices you make affect your share and some of the choices you make affect the total industry value.

Especially in these network industries where the size of the industry critically influences its value. You have to do the right tuning between choices that affect your own payoff and choices that affect the total size of the industry. There's a nice little insight, its I think fairly obvious once said but its important to keep in mind and that is, if the total industry value goes up from a standard being chosen, there should be some way to reward all the players in the industry. If the size of the pie gets bigger, that means there's a way to cut the pie, so that everybody currently there gets a bigger piece of pie. Now sometimes you need third parties to broker these agreements. So, for example, Hollywood was so badly burned by the Beta versus VHS standards war, that they insisted that DBD makers reach a standard. So they didn't care what the standard was so much, so long as there was a common media standard. And now, of course we have this nice standard for video content. But its kind of interesting to now that there are actually currently seven different standards for DBD writing. That is, if you want to use a DBD writes on your computer, you have to pick among seven different proprietors' standards. It looks like that's being dwindled down to two, but I think its not so clear, whose going to win there. And why is that? Well Hollywood didn't really care about the rights standard. In fact, they like the idea that there were many competing standards, because that made it harder to pirate content. So there hasn't been anybody who emerged and said, well we've got to reach an agreement on this front, so the process of standardization on the write side is much slower than it is on the read side.

So our recommendation is that make alliances not wars because the value standardization is so great in these network industries, it's important to try to realize that value through this inter connection. On the other hand, the terms that you can demand on your strength in terms of negotiating these standards depends on how you can go it alone and how strong you are in terms of your own independent capabilities, so our view is that you should prepare for war but make peace if you can and we discussed several examples of this. One very interesting case history is Microsoft versus Netscape in terms of standard setting in the H2L arena, even though they were bitter enemies and struggling very over the market share in that domain, they still were able to agree on some very critical standards. They were very critical because the cost of fiddling to a degree would be very large, and one example is Virmal, the Virtual Reality Mark-Up language they could use for 3D if they had a potential standard for war, but adverted by agreeing a common standard, even more importantly is SET, the Secure Electronic Transactions, both of them recognized that they were better off getting a common standard, competing within those markets than trying to compete for those markets. So, in conclusion, our view about the new economy is really, its not so new. You can learn a lot about the new economy from the old economy and the old economics. In our view, what you want to do is to seek models, not trends and develop concepts not buzz words and create analysis, not analygics because its really the models, the concepts and the analysis that lead to true insight into this market.

So I want to end with a quote from the Economist, John Maynard Kaines who said back in 1935, the ideas of Economists of political philosophers, both when they are right and when they are wrong, are more powerful than is commonly understood. Indeed, the world is ruled by little elves. The practical man who blew themselves to be quite exempt, many intellectually influenced, are usually the slaves of some defunct economist. Mad men in authority who hear voices in the air are distilling their frenzy from some academic scribbler of a few year's back. I love that last phase. In fact was so impressive is that Kaines could see the advert with a cellular phone way back in 1935.

So, thank you all your attention. I recognize this is the scarcest commodity of all and the information age.