San Francisco

One of the perennial dilemmas for policy-makers has always been who to favour - big business or small?

Economic theory would generally have it that it is small business that requires the most help. Big businesses wielding market power can easily use unfair tactics to tilt the corporate playing field to their advantage, and governments are rightly looked to as the appropriate umpire.

The problem for governments seeking to do so has usually been a political one. Large, rich companies employing large workforces can make unpleasant headlines if a government displeases them.

But when it comes to developing a framework to encourage national innovation, things are a little more complicated. Orthodox economic theory would have it that small businesses are most likely to be the source of ground-breaking innovation. See article: The Beginning of Breakthroughs

Yet another school of economic thought has it that there is a little understood form of research and development investment that sits between pure research and pure development. It requires long and patient investment but offers no guaranteed commercial payoff. This type of research has in the past been supported by the vast budgets of companies such as AT&T, Du Pont and IBM when they were in a position where they totally dominated their markets.

But deregulation, new technology, and corporatisation and privatisation has seen these monopolies weaken, and their labs have shrunken significantly in recent years. Also, the Silicon Valley innovation cycle has had a profound impact on investment markets. See articles: The Frictionless Economy and Outsourcing and New Ways of Doing Business.

The technologically innovative entrepreneur has assumed folk hero status, and the start-up company has become the yellow brick road to glamour and riches, especially in the US.

But can 100 high tech start-ups replace one Bell Lab? Jakob Nielsen, for one, thinks not.

Nielsen is in a position to know exactly what goes on inside those endangered research behemoths. He is one of the world's leading experts in hypertext (the computer languages to link pages or files of information) and user interface (the design of the screen with which computerised device users interact). Nielsen has worked for two of the great private labs in the US, telecommunications equipment lab Bell Corp, and hardware and software manufacturer Sun Microsystems labs.

The first myth that Nielsen explodes is that university professors have the "ultimate freedom" to indulge in ground-breaking research. (He began his career as a professor at the Technical University of Denmark before being recruited to Bell Corp in 1990.)

"They just have to show up every now and again and give a lecture. Otherwise, they don't have any obligations, right?" he says.

Well, wrong, actually. "As it turns out you can't really do anything by sitting in your office and thinking. If you're into computers, you need equipment.

"You always need to write up all these grant applications and have some big committee down in the capital sit and ponder it - some ministry will dole out the money.

"Interestingly enough, when I applied for grants to study online information systems, that was turned down every time because people just didn't think that was important. Okay, word processing we can understand. So you can do research on how to improve word processing. But why do you need a better interface for online text? That was not obvious to a lot of people. Now, of course, with the web, it's obvious."

What a contrast he encountered at Bell Corp in 1990. "Bell Corp in those days had a policy that they wanted to get the best. So they went around the world and captured those people who were really good," he says.

"At Bell Corp the policy just was, well if it's something you need, you know what you need because you are the world expert in this topic. So we'll give it to you.

"I mean I had to fight to get cheaper computer equipment than they wanted to give me because it was sort of like they really wanted me to have the fancy Silicon Graphics workstation. I said, I don't do 3D, I only do two dimensional user interfaces, so I don't want this Silicon Graphics!"

The generosity extended to the licence researchers were given to pursue their own work agenda, both at Bell and at Sun Microsystems, where Nielsen was one of a team of 30 distinguished engineers with great freedom to follow their professional curiosity. "The culture really was, we'll get the best people, we'll assume they know what they are doing then we will give them everything they want," he says.

The companies gambled that the generosity would be repaid in research discoveries from left field. Nielsen says he was involved in several examples of this.

"One of the projects I did at Sun was the anti-MacIntosh. (A colleague and I) decided we wanted to look, just as an exercise, at how the user interface would look if it were exactly the opposite of the MacIntosh.

"Apple had just published a very nice book called the Apple MacIntosh Human Interface Guidelines. We reversed every single rule, and this is an example of a pure thought experiment because the result was not a computer anyone would build and send to a customer.

"But what was very interesting about it was a lot of the design principles we came up with very well described the internet desktop, in other words a computer that's optimised to internet access." This research happened to take place on the cusp of the internet explosion in the early 1990s.

The concern about the loss or downgrading of big corporate laboratories is based on the view that it is at these labs where research motivated by immediate concerns sparks new knowledge.

Nielsen describes how the distinguished engineers at Sun, and the researchers at Bell Corp practised this.

"At the university there was this notion that it didn't matter what you did, just write your papers, that's the big thing at university. At Bell Corp, it was like, well, you do your papers, then you go and talk to these guys who are developers to make sure they understand, to make sure they are up to speed and to make sure they get the benefit of the new inventions before anyone else," he says.

"But at Sun it was much more direct: Actually go and be with a project, consult with a project and really help them on more day-to-day situations."

He says there was a risk in that approach that the researcher would become "just another programmer".

"But on the other hand I would say the way you get inspiration for new things is from the problems of today, because you see what is bad and hopefully you get an idea of how to improve it. Then you can start pursuing those and start doing some quite futuristic work," he says.

Nielsen does not believe large companies can continue to stoke the fires of innovation merely by establishing their own venture capital funds - in effect, outsourcing their R&D function.

He concedes that waiting for "random entrepreneurs" to develop a specific application for which there is already a market demand, such as speech recognition, and then buying them to gain access to the technology, can be a legitimate way for giant corporations to move their products forward.

But when it comes to advancing the fundamentals of their business, or understanding what big, profound changes technology might be about to bring to their industry, big corporations have to be in the game themselves. "If you want to be able to know what the car of the future will be if you're a car company, that's the kind of thing you can't outsource.

"You just can't wait until somebody else builds that car and then you buy it."

But it seems that the disaggregation of the corporation is an irreversible trend, thanks to the liberation of information management

Who can devote their energies and resources to long-term, commercially motivated research?

One Silicon Valley-based start-up, Intertrust, offers an example of how it might be possible to bridge this gap in the knowledge creation chain within a new business model. Intertrust has developed a new internet security system. It boasts it can deliver any type of digital information - from text to video or sound -- with complete security and provide accurate tracking, billing and usage data. The first generation of its product is presently being tested by Intertrust's commercial partners. But unlike the typical Silicon Valley start-up, Intertrust has been developing its product for more than a decade and, not coincidentally, has received no investment from Valley venture capitalists.

Duncan Davidson, Intertrust's vice president, business development, says the secret of the company's ability to work for so long was the personal capital of its founder, Victor Shear, and his tactic of systematically wrapping as much of his work as possible in thousands of protective patents as he went. "The only other company we know that did stuff like Victor did was AT&T in their labs," he says.

Intertrust is Shear's fourth company and the success of previous ventures allowed him to run it on a shoestring of a couple of million dollars capital until 1994, when he raised another $6-7 million from angel investors. See article: How Angels are Giving Flight to the High Tech Boom.

Intertrust then raised money from institutional investors in 1997 when it was confident that it was close to releasing a product.

Davidson says Shear's ability to avoid venture capitalists was crucial. "It's very clear that from 1990 to 1995 if the VCs had been in and not been extremely patient, this company would never have gone as far as it has gone," Davidson says. "It had to crunch and crack some fundamental computer science problems before it could get to where it was."

Davidson says if Intertrust succeeds from here, it will demonstrate two things: the value of patents in computer software, where they have never been systematically used before, and the value of fundamental computer science versus "quick iteration". He also says people will look at the Intertrust model to see if it can be systematically replicated.

But in the absence of an established serial entrepreneur with deep pockets, how can such a model be funded? Davidson says the biggest problem with getting outside investors into long-term research projects is the absence of a market to give their investment liquidity, so that they can cash out at any time. But he says it is possible to envisage how such a market might be created, once people understand the blue-sky potential of long-term research.

One way, he says, might be to securitise investments in several long-term research based start-ups such as Intertrust, in the same way as debts such as mortgages or car loans can be turned into financial products and sold in units. That way, a package of start-ups can be bundled together, unitised, and a valuation struck for each unit based on the historical hit rate of such companies. The investments could be traded just like any other listed or unlisted unit trust.

It might seem an unlikely device to pay for the work of boffins, but if investors are willing to buy other people's exchange rate hedges, maybe not.

Further Reading

The Productive Edge: How U.S Industries Are Pointing the Way to a New Era of Economic Growth. Richard K Lester.

Bell Labs is Dead, Long Live Bell Labs. Robert Buderi. Technology Review, September-October 1998.