Web forecaster: Forrester interview

George F Colony is President of Forrester Research, a leading American technology consultancy. Colony founded the company back in 1983 ? indeed, the ?F? in his name stands for Forrester. The consultancy (www.forrester.com) is now unquestionably one of the world?s leading centres of expertise on Internet computing, and Colony himself is a widely-quoted guru. We caught up with him while he was in London for Forrester?s first UK conference, a forum about Making Money on the Internet.

Q Do you think it is possible to carry out innovative Internet development while still taking care of the basic business? There are very difficult problems like the Year 2000 and European Monetary Union competing for resources.

A What we are advocating is what we call the curtain strategy. You pull the curtain down, and on one side you deal with the year 2000 and other legacy problems, but on the other side you have to continue to develop new systems. If you don?t you are essentially stealing from tomorrow.

The biggest impact of the year 2000 will be organisational ? an entire generation of CIOs (chief information officers ? a common US title for head of IT) will be fired because of the year 2000. The CEO will turn round and say, ?hey, you knew this was coming for 20 years, and you didn?t fix the problem?. We estimate that 30 per cent of the year 2000 problems will not be fixed by the year 2000.

QHow are distributed applications going to be managed in the future? There are various competing technologies, such as Corba, Java and ActiveX. Which do you think is going to come out on top?

AWe believe it?s going to come down to a Java world. Java is essentially a language, but it?s becoming much more than that now. There?s also picoJava, a chip design, as well as Javabeans, which is an object model. So there is now what I?ll call the Java environment versus ActiveX. That?s really the battle that?s shaping up.

Q You think Corba is sidelined now?

ANo, Corba is being integrated within Java.

Q Is there a chance that Microsoft will do a U-turn, in the same way that it did on the Internet in general? Do you think Microsoft could really get behind Java?

AI?d say the chances are slim to none that that would happen. Yes, Bill Gates will adopt the language, but as far as adopting what I?ll call loosely the Java API (application programming interface) or Java definition of the object model, he would never do that. He perceives that to be such a point of control.

What terrifies Bill Gates, what wakes him up in his big house at three o?clock in the morning, is that Microsoft loses control of the object model. Because if it loses that it?ll lose a great deal of control over how applications are developed over the next five years.

QDo you think Microsoft can continue to grow, particularly as the emphasis is moving away from individual desktop applications to enterprise-wide computing?

AGates has killed off so many of his competitors, he?s bought so many businesses, where?s he going to go next? How does he scale this thing to 25 billion? One of the ways he has to do it is through packaged business applications.

Gates looks for markets with three characteristics. One, high complexity. Two, volume. And three, he wants a world which is very programming intensive.

Now the packaged applications market lacks volume; but if he could commoditise it somehow and radically lower the price, then he could get volume.

He could use ActiveX to build business objects. Then a third-party partner of Microsoft could strap the things together and go against SAP with a solution which is a hundredth of the cost.

QIs Peoplesoft also a potential target?

AYes. Companies like Baan, Peoplesoft and SAP are likely targets. They are like the last dinosaurs roaming the earth, and Gates loves to kill dinosaurs.

QWhat about the really traditional dinosaur, IBM? It seems to be adapting quite well to this Javaized universe, what with Lotus Domino and all the heavy IT Java developers it has.

AMost of the recovery that IBM has shown is really linked to the general economic recovery in the US.

IBM is more like GM than Microsoft ? it is a cyclical company. When business is up then more mainframes are bought, more CICS (a transaction processing system) is bought and IBM does better. In a downturn I?m not sure it would perform as well.

Q Where does Novell fit in?

AYou?ll see it become part of the Java world. You see NDS (Novell directory services) done in Java, you?ll see Netware done in Java. Just as IBM is rather like a secret member of the Sun, Oracle club, I think you?ll see Novell evolving towards that club ? particularly now that Eric Schmidt is at Novell (Novell?s new CEO was formerly chief technology officer at Sun).

QWhat about Forrester?s way of thinking about the Internet as ?the fourth channel?. Your idea is that the first channel is person-to-person interaction, the second is paper, the third is the telephone, and the fourth channel is the Internet. You are saying that to be really effective on the Internet you have to have done well in the other three channels?

ANo, I?m saying it gives you an advantage. If you look at, say, the top five banking brands on the Internet five years from now, it would not surprise me to see the top ones being the leading brands in the first three channels.

QDoes that mean that a lot of the early successful Internet startups, such as Amazon books (www.amazon.com), will succumb?

AI call Amazon books amazon.toast, because Borders (Borders Books and Music, a big US retail chain) and Barnes & Noble (an even bigger US bookseller) are coming in very strong now.

Amazon?s position is indefensible. They have some nice custom-built software, but that?s about all they?ve got. They don?t have a monopoly over the books, and their technology can be duplicated in six months.

Some of the brands that went on the Internet early on, like Amazon, do have a chance of becoming lasting brands, but they should prepare themselves for a very tough fight.

Q A lot of the most successful Internet applications have an element of self-service about them. Do you think that?s a pretty key element?

AYes. But for me it?s really what I call time arbitrage. In the US now we are working seven per cent more hours than we did ten years ago. So there is no time left. As far as I?m concerned, what Amazon is doing is giving me time ? and I?m going to be attracted to any company that can give me time as part of the product.

QIt is noticeable how few profitable Internet commerce sites there are.

AYes, there are very few ? if any.

QWhen do you think profitable sites will appear?

AI don?t know, but I?ll give you my estimate. I predicted about 200 billion dollars of Internet commerce in the US in the year 2000. And how much profit will there be on that revenue? I would bet a 10 billion dollar loss.

QHow crucial is the emergence of new payment systems? Do you think credit cards are perfectly adequate?

AYes. It looks to me as if Visa and Mastercard are going to keep their monopolies, and it?s going to be a credit-card environment. I don?t think there?s going to be an Internet Commerce Unit. The central banks of most nations won?t allow a scrip of that type to be generated.

What?s more, within three years I think we are going to see at least 50 per cent of PCs being delivered with card readers: you put your card in and download money from your bank.

QWhat about micro-payments? Will that become established quickly now that people like Digital seem to be pushing it? And does it actually matter much ? will it enable new types of commerce on the Internet?

AIt is really a minor issue. There has been the idea in the content publishing world of charging, say, 10 cents for articles and 15 cents for information.

My belief is people won?t pay even small amounts for articles ? it?s just not going to happen. So, in my view, micro-payments is a peripheral consideration.

QWhat role will advertising play in the future?

AI think we?ll find the price of desktop software will fall to zero in a fully Internet-connected world, because of the ad angle. It?s a scenario we?ll have to think about and research carefully. But Yahoo has just done a deal with Netscape, so there will only be one search engine within Constellation and Communicator, and it?s going to be carrying advertising.

Why would I pay for Communicator or Constellation when there?s advertising built in to them? How long will it be before the ads start to roll in Microsoft Word? Or Excel ? of course my Excel spreadsheet will be connected to the Internet so it can get new ads.

Or even Corel Office, written in Java? My point is that the price of software will eventually go to zero. If I have to watch ads, I?m not going to pay you money for it!

QSurely that?s bad news for Microsoft, because it still makes a lot of money from Office?

AYes, half the business is from Office sales. And I?ll tell you another problem Gates has: he rolls the Office base about every 21 months ? his business model is dependent on it. But, if you actually look at Office 97, what is the difference between it and Office 95?

The problem is, there is no difference! There?s no reason for users to roll on to the next version, and the inevitable slowdown of revenue from Office is one of the major challenges facing Gates. Maybe he?ll be forced to run ads in Office!

QAt the moment, the Internet is still a very American world. Is anything likely to change that?

AInstead of whining about how it?s all US-based, I?d do two things if I were Tony Blair. One, I?d widen the Internet pipes and make them cheap in the UK.

And two, I would work on building a strong venture capital culture and community here in the UK. We?re awash with capital in the US. Forrester is a small company ? and we?ve just raised $35m. There?s capital everywhere.