Where did the reasoning for IBM's licensing strategy come from?
Well, both AT&T and IBM settled antitrust suits with the Justice Department in 1956. Both of those consent decrees required that IBM and AT&T license their IP to all-comers. So in the IT world, which followed those two companies at that time, licensing was a very common practice. There was no software industry then. It was basically born on June 27 1969, when IBM unbundled software from hardware and put a price tag on the software. And IBM continued with that model until 1992, when I took the job of managing IBM's IP portfolio.
I had come to the conclusion that the competitive landscape had changed dramatically, and that it was appropriate for IBM to really up-tick its efforts to license out its technology and its IP - but at a different price level than it had done in the past. One of the reasons why they hadn't done that before was because they were making all the profit in the 80s, but then IBM's business model changed and they were getting beat up pretty badly in the marketplace - in fact, I think IBM filed an $8bn loss in 1991-2. But my view was that, irrespective of your fiscal health, you're spending a lot of money on research and development (R&D). Why not see if you can get people on your R&D base, which helps you make some money, and keeps researchers on their toes to move ahead and stay ahead and its great money to get in internally. In 1991, IBM was a net payer for the IP of others, and we turned that around pretty quickly. At that point in time I was offered three different vice-presidencies at IBM, one in the personal computer business, one in government relations, and this one in the IP side. I had a made such a pain in a neck of myself on this issue, that I took it because I said that this was an underperforming side of IBM and I can make something of out it.
As part of that, was there an emphasis on not just patenting more, but getting the R&D guys to think more in terms of things that may not be immediately relevant to the business?
Not just the research guys, but R&D broadly. Most of R&D is D - development - by a factor of 90:10. So there is a huge education component involved in this on a number of levels, and two in particular. One is making sure that they protect whatever they develop, and they hate to write patents. Developers want to develop, they just want to research. They don't want to sit down and write patent claims with somebody. The second thing is convincing the company that it was in their interest to out-license as opposed to keeping things internally. The intellectual misadventure is that most people - and the higher you go they think this more - still think about patents as "a way to for us to stop somebody from doing something", not necessarily "could you take this chunk of IP and go out there and do something else with it, something different". A lot of CEO types and senior executives view patents as a brick wall that you put up there and people will run into it. Well, in our industry its more like a fog-bank - you look out there and you try to drive through it. And you don't know what you might find when you get in that fog-bank, whether it's going to be difficult or not difficult. So you have education on both sides of the equation.
Some people would say that if you're IBM, it's easy to take the licensing route because you can dictate terms to people, but for a smaller company using the patent defensively is more effective.
I'd be the first to admit that one size doesn't fit all. It may well be that as a small company, if I have some really good IP, I actually have quite a bargaining chip against a big company because the leverage is on my side; I have IP that a Microsoft or IBM really needs, if I choose to use it. But I would use the argument that, especially in the software world, there's precious little that can't be designed around over time. If you accept that premise, then licensing the technology and the IP starts to make more sense. If you're going to spend $7 billion a year on IP - which is what Microsoft will be going forward - then I think an outward-facing system is probably in the best interests of Microsoft.
You get a virtuous circle: you have all this R&D that hives off a whole bunch of IP. I can take that and I can go license that to all-comers. When I do that, I create standards, I create relationships and I collect money. I can then turn that money back into the corporation and I can spin around that wheel and so on. I view that as a behavioural model for the industry, whereby a competitor gets that technology and starts to create its own wheel and they're leveraging that; the industry gets the technology and they start to create their own, and pretty soon you've got these things working all together. It's a model that works and I think that most of the high tech is following that model right now. Are they doing as well as IBM? No. But do they have 60,000-70,000 patents? No. (Which is another thing we decided to do in 1992, to really up-tick that piece of it, so we had the capability of doing this kind of programme with a straight face).
So what's the difference with Microsoft?
First of all, it wasn't until the mid-90s that the issue of software patents in the US became crystallized. Until then, Microsoft, and all the other software companies, were copyright barons. And they still are - most of the products that Microsoft sells is sold under a copyright license as it is with almost everybody's software. In the US, it became pretty clear that you could patent novel inventions even if they were software, and even if they were only implemented as software. Microsoft got a late jump into this patenting world. They have about 3,000 patents and 5,114 applications which are sitting in the USPTO waiting to be acted on. So there is a pretty broad understanding at Microsoft on the development side of the equation that you need to do this properly. And the good news for Microsoft is that most of its patents are in one specific area, software.
Do you have targets for the number of patents you want in the future?
Yes. Are we good at it? Well, you can file as many patents as you want to spend money. So it's a quality and a quantity game.
You really only want to patent the stuff that's really important, so there's a cut made there. Then there's how much are you going to get for a budget, so there's a cut made there, and you try to rationalize those two things. But it is clear, moving forward, that we're going to want to do more than we have done in the past. If anything, the patent engine is accelerating - it's not static, nor is it decelerating.
Do you have figures for the targets?
Yes [smiles]. I gave you the two numbers - 3,000 exist and there's another 5,100 in the bag and I'll just tell you that the pace is quickening, not decelerating.
You've got two choices: you can keep it as trade secrets if you want, but the problem with trade secrets is that they go home at 5 o'clock at night and once they're gone, they're gone. Or you can look toward legal kinds of protection such as copyrights, patents and contracts. There are a number of ways to do this and you want to find the most appropriate way to do it in almost each case.
Are you also ramping up your patent team?
Yes. There's a very talented group of people over here. If I have to say anything about what I look forward to it's being less internalized on one level - not just using this stuff defensively, but finding a way to have a proactive approach to the industry and whomever might want to use that IP. But secondly, we need to do a lot more internally of the type that we haven't done in the past. We've used a lot of outside law firms to do patent preparation and prosecution; we're going to start doing some of that internally. Why would you do that internally? First of all because I believe that internal lawyers who are sitting physically with the developers are going to do a better job than just throwing the technology over the wall to an outside law firm and saying "make the best of it". If I had to pick where I would start with that it would be Microsoft Research. At IBM, we did a lot of it. We had lots of lawyers doing preparation and prosecution around the world. At Microsoft, we don't have anyone doing that, but we are going to start.
How many law firms do you use at the moment?
Probably about 50 around the world. That doesn't mean that there isn't some surgery that would be good, or Slim-Fast diets that we could use on some of these. Over time you have to look back at all of this and say are we optimizing the extra resources that are available to us, or are we being more haphazard than we ought to be. And that's something you ought to do on an annual basis.
How many people do have in the IP team?
About 125.
And do you have targets for growth?
Yes.
Would you like to tell me?
No [laughs]. But let me break it down for you: there are IP lawyers, there are a bunch of paralegals, and I have a large and talented standards group. I have a corporate development function in my area because everything that Microsoft does is IP-related. We have a licensing function that is building up to do all the things I just talked about, and which heretofore had been pretty small and focused on one-offs, as opposed to a programmatic, generic function. We need to do more overseas than we're doing. We've been very Redmond-centric, US-centric about all of this stuff, and there's a big world out there and Microsoft is doing increasing amounts of R&D in places like India, China and Ireland - and western Europe - and you have to have an IP engine that recognizes that fact. You can't do that well by flying in once a year and saying "tell me what you guys are doing" and then make some judgment that you're protecting your IP adequately. But you don't know in that circumstance. So I think one of the things we will do is upgrade our capabilities overseas to do that.
The way Microsoft is structured, do you work with the general counsel or with the company directors?
Both. If you think about Bill Gates's job of chief software architect - which is one he takes very seriously - he's involved in these IP discussions all day long. I spend a lot of time with him. I also spend a lot of time with [general counsel] Brad Smith, but on these IP issues I don't have much of a problem getting anyone's attention at Microsoft. It starts at the top, and if you've got a guy like Bill Gates who's really concerned about IP and really likes it then you don't have much of a management problem.
Do you have revenue targets for your licensing programme?
Yes. What you do, if you want to treat this as a business - which you should - then you line up your opportunities and ask who can use your technologies. I don't approach this negatively; I don't ask who's out there infringing our technology that has to license from Microsoft. What I do is say "how do we make it a win-win situation?". Because if you're spending close to $7 billion a year in R&D, there are a lot of people you're going to effect when you do that. So you say "okay, I don't mind anybody using the fruits of my R&D, but I want a reasonable return for that" and that's fair, because they are avoiding the R&D costs that we have taken on. So you look at that and see who the main beneficiaries are, and you put together a licensing programme to try and persuade people that that's a much better way to go than just the negative aspects which is "I'm going to stop you from doing that" or "you can't use whatever I've got". That should be a process, working with the technical people, telling me where the industry is likely to go. I don't particularly see the difference between software and hardware, so I take a pretty broad look of where I see the industry is going, and working with the tech guys we put together a programme accordingly.
Are we there yet? No we're not. Do I think we can get there? Yes I do. Rome wasn't built in a day, and it's going to take some time to change the inward focus which we're really good at to an outward-facing focus for Microsoft.
In light of proposals for the software patent directive [the directive on computer-implemented inventions], how are you going to be able to use the European system?
Well, I could live with the old proposals, but the new directive is even worse. We had a similar problem in the US where we used to have to find a hardware implementation for the software - but that was possible.
The new directive is, I think, very bad for Europe, in that it's a form of industrial policy that at the end of the day is going to limit European competitiveness with the US and Japan in this area of software. But that process is still going on, and lots of companies are viewing this latest version [of the directive] with great alarm, not just Microsoft by any means. And who knows what will come out of that sticky process that they have over there, especially next year when they'll have 25 members of the EU.
Why do you think that in Europe there is the fear of allowing software to be patented?
I think there's a fundamental lack of understanding that there isn't a real difference between software and hardware. I also think there are large political forces driving this issue. For some reason, it has caught the attention of the Greens, and I'm not clear why. But that's true of this country too, where [consumer activist and Green Party presidential candidate in 2000] Ralph Nader is involved with the Open Source Movement. Some on either side approach these issues with almost a religious fervour. My own view is that we need to get a little more sophisticated. Linux isn't going to go away and Microsoft isn't going to go away. IBM isn't going to go away, and IBM has got one foot in each camp. Hewlett Packard isn't going to go away, not to mention all the financial institutions who write their own software, like Swiss Re and UBS - they're not going to go away, either. So I think we need to enter into a new world with a little more sophistication than this black and white world that some people seem to see out there, which I frankly don't see.
Also, I can give you all kinds of software implementations that used to be in hardware. Apart from the plug in the wall, anything you used to do in hardware you can do in software. Just think about the media player that exists when you boot up your software on your machine. It's looks just like any other media player - it's got fast forward and stop and pause and all that stuff. That's software. Is there a hardware version of that? Sure, any DVD player you want to go and buy. Or if you look at some laptops there's a little button in the middle which is basically the mouse. You think that moves, but it doesn't. it's really a strain gauge that calculates where you want the cursor to go by how hard you push it and in what direction. How does it do that? Well, there's an algorithm for an x-axis and an algorithm for the y-axis - that's software. Is there a hardware implementation for that? Sure, every mouse that's practically ever been built is a hardware implementation of that.
I think people react to software as if it's an animal from another planet, and I don't think that, at the end of the day, the distinction [between software and hardware] is a very useful one.
There is an argument that software is made up of algorithms, so patenting it is like patenting language or formula. The fear seems to be that if IBM goes out and gets 30,000 patents in this area, then anyone who tries to write any kind of software anywhere in Europe is going to get sued by IBM.
Well, not really. I don't think IBM sues many people at all. In the eight or nine years that I ran things there, we didn't sue anybody. Instead, we came up with a licensing programme that gave the technology for people to use - exactly the opposite of suing people. I have heard that argument, but I just don't buy it at all. I think fundamentally if you're willing to make you're technology available, which IBM certainly is, then I don't know how that's anti-competitive.
What about the argument of being against the principle of paying for these algorithms?
Well, are they novel or not? If they're novel, somebody went through a lot of work, a lot of research, to create them and make them useful. If I had invented that, and you wanted to use it, I don't see anything wrong with me licensing you to do that. You didn't do the R&D, and I have the negative right not to license that if I don't want to; I can stop you from doing it exactly the same way that I did it. Again, I don't see the difference between just because it's written down as software - I could take a chip and I could hardwire anything in there that I wanted to do. Or, I could take a general purpose chip and I could write software that overlays with that chip that does exactly the same thing. What's the difference? I don't know, but somehow in Europe there's a big difference.
My concern with Europe is that what they're going to, writ large, is develop an industrial policy for this kind of thing in Europe. The US isn't following that. The US is off to the races on software patents and that's not going to be turned around anytime soon, if ever, and so is Japan. And why Europe would want to be uncompetitive with those two areas of the world, I really don't know.
If you were to ask [European company] Nokia about this directive, they're very upset about it because they write all kinds of software for their cell phones.
As it stands now, would the licensing project that you're embarking on be hindered in Europe? Would it be a market that you'd focus away from?
No, I think if anything, we'll focus on Europe that bit more. Again, I think we've been way too US-centric, and I don't think we joined the debate [on software patents in Europe] in the right way. Microsoft's got all kinds of over-arching problems with the EU and sometimes you don't want to leap out of your foxhole, which is natural. But in these debates about what is the proper way to go with IP - such as should software be independently patentable or do you always want to have a hardware manifestation - these are really good overarching questions that ought to be debated in a thoughtful manner, with the overall question of are we being competitive with our geography being held in mind, or are we marching to somebody's narrow agenda? We went through this in 1991 with the first software directive, and it was very much looked at as a competitive issue in those days and very much looked at as a form of industrial policy that probably wasn't good for Europe in those days. This latter directive, a lot of people have been very concerned that it looks like Europe versus the US - and as you know there's some real difficult issues about the US that don't help from the European perspective, and I think this falls in with some of that. And of course, don't forget that you've got these huge US companies like Microsoft and IBM, but when [European anti-software patent activists] look at that, they tend to forget about the SAPs and the Nokias of the world, even European car companies - there's an immense amount of software that exists in automobiles these days.
Are you in favour of the proposed fee increases at the USPTO?
I'm broadly in favour of anything that helps the USPTO because its understaffed and in some areas doesn't have the horsepower that it needs to have. Some of that is caused by diversion of their budgets, which has been going on for a long time. I think a good thing for industry to do would be to support the USPTO in those kinds of situations. I don't think you can ever ask the appropriators not to appropriate - that's like telling members of Congress that they can't do their job - but they ought to understand that this again is a competitive issue, a form of industrial policy if you want to think of it in that way. We're very much behind the USPTO on this issue.
The other quality issue is are we issuing the right patents or is the system out of whack...
John Williamson, president of the Intellectual Property Owners' Association, told MIP that his organization is worried that the system is out of whack because the metric for measuring USPTO success is wrong. He said that instead of focusing on the number of patents issued, attention should be given to the quality of the patents that are granted.
It's very difficult to know what makes a good patent. Patents last for 20 years, but what might be useful right now might not be useful in a year and vice versa. It's very hard to say whether a patent is going to be a good quality patent because you made that bet three years ago when you filed it. I think a better answer to that is that it's incumbent on the owners of that IP to scrub their portfolios every year and do triage over the thing, and not pay maintenance fees going forward on patents that don't look like they're going to cut the mustard.
How can you tell if it's not going to cut the mustard?
You sit with your technical people and say "did we make the right bet here, or did the technology zig while we were zagging?" For instance, you filed a bunch of patents on a DVD drive and went down that road thinking that's where everybody's going to go, and all of a sudden the technology veers off this way and all the IP you had related to going that way is basically useless to you. And then you have to make decision about you're going to do with that IP. Can you sell it to somebody? Can you donate it to a university? But at the minimum, ask why you are paying maintenance fees on that IP. So I view the quality issue on being on every IP owners back, not so much the USPTO.
Having said that, there is a huge need in the PTOs of the world for high-quality individuals who are quite versed in the areas that they are saying grace over. In the software area, it's been especially difficult because until recently in the US you had to be a chemical engineer, an electrical engineer or a biologist to become a patent examiner, and computer science backgrounds have only recently become appropriate. That's actually all well and good because you now have more women in the field because they don't go to engineering school, especially electrical engineering schools for reasons I'm not clear on, and you had an imbalance. So the USPTO needs very good people, and it needs lots of them. It needs to have its fee structure maintained, at least on some level, and also see if we can draw halt to the diversion.
But moreover, and I don't want this sound like the USPTO is not careful because it is, but it seems appropriate to me to recognize that the patent system can get out of whack every now and then. It got out of whack when electricity was invented, but heck, the system's been around since article one of the Constitution was written; it self-corrects over time and when it has excesses, they're generally short-lived. Has that occurred? I think you can argue today that there are some business method patents that with hindsight don't look so good - we can remember Compton's Encyclopaedia, when they got a patent that they said basically controlled the entire internet, and what did the USPTO do? They reexamined the patent and said "no, this is way too broad". So, the system does correct over time, and just because there are a couple of bad examples, and there are, doesn't mean we should throw the baby out with the bath water. This is another argument you get in Europe - they mention some of these overarching things [such as over-broad business method patents], but I still think the best way to do it is to fix the system that is there; I would always err on the side of allowing protection for inventions that are novel irrespective of their form. That's the way Japan is, that's the way the US is right now.
Does Microsoft has a position on the fee structure and how it relates to ending diversion?
We would probably support the USPTO in what it thinks it needs.
I know a lot of IP associations have said they won't support the fee increase unless diversion ends.
Again, they're pushing on a rope. You cannot tell the appropriators that they can't do their job, if you just think about it in practical terms. That doesn't mean that they can't help the situation by lobbying Congress and the Executive branch - because the Executive branch tends to pull some money off the table also - to say that we think this is an important engine of US competitiveness that needs to be recognized as such, and you're starving it.
Does Microsoft do that?
Yes. We're big supporters of the USPTO and we're trying to take a leadership position on helping in this regard. I've been up around the halls of Congress, and I'm going to be up there at the end of this week doing exactly what we're talking about.
Do you have a position on trying to work around the consequences of the Eolas suit?
This happened before I got here, so I'm probably the worst person to comment on it. But I would say something about patent litigation - 50% of all cases get overturned on appeal. It was obviously a very large award. There are responsible people who argue that there were some things that happened that will probably lead to it being overturned on appeal, but I'm just not the right guy to talk to about that, and I frankly shouldn't in the middle of litigation, even if I could. There are a lot of bets being made out there on the industry developed that this thing threatens. Sometimes the [Court of Appeals for the] Federal Circuit, when they see these massive things that could be really disruptive, are more inclined to take a hard look at stuff.
There has been some legislative activity that seems to put the responsibility for copyright-infringing potential of P2P on the software industry. Is Microsoft taking steps to adapt to this stance?
I'm sure the answer to that is yes. It's not an area that I particularly focus on, but the standards crowd would very much be focusing on that, and we have a very robust standards group under a woman named Michelle Herman, and it is her job to work with various standards bodies and other companies, whether it's P2P or security, so we do it in an appropriate and forthcoming fashion. That's the one area we're reasonably good at outfacing - in the standards world which in the software arena is a big deal. And many of those issues crop up in those things. It goes without saying that you've got to be a responsible vendor in those kind of situations.
What countries do you look at as being a particular problem for counterfeiting for Microsoft, and what countries are good?
That's a worldwide problem, not just Microsoft's. Any media company you look at, and I don't care what kind of media it is, be it records, movies or software, has this problem. You can even buy $25 Rolex knock-offs on Fifth Avenue in New York, so it's a worldwide problem. Japan has a very good infrastructure where you can do business. China's laws are pretty good, but the question with China has always been what's really going on when you get outside Beijing and Shanghai. I think the UK and Ireland are terrific. Western Europe is pretty good on this stuff.
Does the expansion of Europe into the east worry you in this context?
Not necessarily in this context. I do worry if [the EU] is going to be effective with so many countries on some of these great policy issues. What we're struggling with right now is with the current membership, and when it goes to 25 I just wonder how it will all work. Obviously, there's going to be a heck of challenge there. It's not like they're creating a United States of America from whole cloth; they're patching it together over time. And I think it has a tendency every now and again to go off the rails, certain groups build up a head of steam, but in terms regimes that might be effective, I would think it's probably a good thing that they're going to have 25 countries if they have consistency on these things. All the countries will have to implement the laws, so if we assume they get the right laws, then I think it's a really good thing. If you get the wrong result, you multiply by that 25 and you're not where you want to be. These are good questions that people really need to think about. It's like they say to doctors, "first, do no harm".
Do you have a strategy to combat counterfeiting?
Yes, but it's not my area. But we do all those things. We work with Customs, people stopped at the border because it's counterfeit. There's a very large department that does that.
Looking at the licensing programme on a global scale, what countries, outside of the trilateral regions, where do you look to work in?
Where is stuff made? Thailand, Taiwan, South Korea, China, Singapore. These countries where development follows the cost of labour. So you'd clearly look at those. Obviously you'd look at the developed world, the usual suspects that you'd tick off, but I certainly think there are big licensing opportunities in Asia.
Are you confident that you can enforce agreements in Thailand, for instance, and have people work with you?
That's a very good question. It's going to vary by country and by company. You wouldn't necessarily want to put yourself in that position [of working towards an unsuccessful licensing deal]. You get certain areas of the world that are confiscatory the way they approach this stuff, so you might not in that circumstance, and again one size doesn't fit all. You tend to look at not just the perspective of the country, but you start with the company. Take [laptop maker] Acer: is there any technology that we've invented for the tablet PC that would be of interest to an Acer, for example. I would tend to start not so much from the legal regime, but from the company itself, and there may well be within the legal regime pockets of opportunity irrespective of the legal regime. Obviously where the legal regime is dicey, whether or not your ability to enforce your agreements is open to question, you're going to look at those carefully.
What is it about this job that you like the most?
There's a few things. The first one is the opportunity to try to do this again, and under really good circumstances. Second, contrary to what I expected, I've found the reception and the people to be terrific. That makes it a lot easier than you if you have to push a rock uphill. I guess the third point is the commitment at the top of the business to focus on these issues hard, and to try and make the right decisions, and a willingness to move ahead.
If there was one piece of advice that you would give to IP counsel in a company that is looking to license, what would it be?
You've got to get the commitment of the top leadership of the company. They have to understand what this really means. You can't sort of have a licensing programme, but sort of not have a licensing programme. In my opinion, people have to understand this stuff before they make a reasoned judgment to have a licensing programme. The first question is why don't we just use our IP to stop our competitors, and you've got to get over the intellectual hurdle with people who think "what's a patent for if I can't stop somebody?". You have to explain to people why this is a reasonable business process, and treat it as a business, not as a bunch of lawyers running around licensing people. It's really a business question: do you want to get a return on your R&D? Do you want to spread your R&D out to other people? Is that a good thing to do for your company, or are you better off holding your R&D close to your chest? Even in the drug industry, where the patent and the drug are one and the same, you're now seeing that industry start to spread the wealth around a little. Not in the same way and asserted by licensing, but certainly in joint development because it's so expensive. So you're starting to see this ripple into other industries as well, and I think that's good.
