Speculation is rife these days about the future of Internet portals. Once seen as gateways to the Web, search engines such as Yahoo!, Excite, and Lycos were some of the first places visitors came when they logged onto the Net, making them prime real estate for advertisers. But it isn't enough anymore to just draw traffic to your site.

In fact, Excite CEO George Bell predicts that direct marketing -- reaching out and selling something to your Web customer -- will soon surpass advertising as the primary revenue stream on the Web. At several recent industry conferences, Bell has emphasized direct marketing as "the next logical step in an environment that is driven by personalization."

In early March, Bell spoke with Business Week Online reporter Stefani Eads about how Excite is positioning itself to capitalize on the commercial potential of the Web -- in part via MatchLogic, the Web-based direct marketer Excite acquired a year ago. Bell also spoke about what it will take for Excite to win the portal war. Here's an edited transcript of his comments.

Q: How soon will we see a shift from impression-based advertising to direct marketing?
A: I think it's happening now. What we're able to charge for [Excite's direct-marketing program] through E-mail generally ranges from $300 to $1,000 per thousand ad impressions. The click-through rates -- people opening up an ad or E-mail -- are 1% or less for banners compared with 5% to 10% in E-mails.

While you get about a 1% response rate to banners, you also get about a 1% conversion rate. With E-mail, or direct marketing, you have a 5% to 10% response rate, and then something between 10% and 45% conversion rates, which yield the ability to charge a much higher amount for advertising. So it's happening now. We sent out 25 million pieces of E-mail in the last quarter of 1998.

Q: To Excite's 57 million profiles?
A: No. In our company there are three kinds of databases. There are imputed profiles, where you don't know the name and you don't know the E-mail address. What you know is that by setting a cookie on the person, each time they come to the site, you're able to track some behavioral attributes. So MatchLogic [Excite's direct-marketing subsidiary], in serving advertising for advertisers across Web sites outside of Excite, sets cookies and determines demographic and behavioral patterns of usage on the Web associated with each individual cookie. So it doesn't know the name, but it knows a fair amount about what the behavior has been over a period of time. So we call those behavioral or imputed profiles. And we deploy them to improve ad targeting.

The next set of profiles is the people whose names and addresses we do know, and we have them categorized in two ways. One is the people that registered for the My Excite start page, where they filled out a registration, or maybe they have a free E-mail account with us, or downloaded games, or Java Chat, or something like that. We have about 20 million of those people, and we know their names and E-mails and some behavioral characteristics.

The third kind of list is the opt-in E-mail database that MatchLogic has built on its own. That's where it goes out on the Web and runs contests in places other than Excite to induce people to give characteristics about themselves, including name and E-mail, for use as an opt-in E-mail database.

Q: What are the margins for the direct-marketing business?
A: The opt-in E-mail business that is run by MatchLogic, which is called DeliverE, has 91% gross margins, and about a 50% net margin. We exposed those two numbers this past quarter for the first time, and we said it was also a growing part of our business. But it's still substantially under 10% of our total revenue. It just means that one business segment, which is growing very rapidly, happens to also carry very high margins, because you're able to charge so much.

Q: How do you see rich-media advertising, like that provided by your relationship with Enliven [an interactive ad agency owned by @Home, which is in the process of acquiring Excite], fitting in with this big direct marketing push?
A: I'm not sure I know the answer to that yet. Don't forget that impression-based media will still, by and large, be the gasoline that powers the whole set of services behind Excite and @Home for a while. Enliven ... is getting click-throughs and conversions that are much higher than banners, because it's eye-catching, it's entertaining, it's animated. How Enliven will play into the world of opt-in E-mail direct marketing, I don't know. I just haven't gotten that far.

Q: What do you think Enliven will do for your advertising revenue?
A: We have about 1,200 advertisers over this last quarter, and the top tier of advertisers keeps asking us when they can experiment with rich media and hold the experiments stationary against a control group, which would be narrowband media. But because the commercial is so different in rich [media] than it is in narrow, that's not really a controlled experiment. The creative is so different that the proposition changes to the user.

[Through Enliven] for the first time, we'll be able to give advertisers a relatively controlled environment, where we will deploy a technology called Speed Select that allows us to "sniff" for the bandwidth you're using to connect to the Web. And if it's narrowband, we won't show you Enliven advertising. If it's middle or rich band, we may serve you Enliven advertising.

For the first time, an advertiser will be able to talk to one salesperson from Excite/@Home and say: "What I want to do is experiment with narrowband delivery, middleband delivery, and broadband delivery. What I want is one consolidated report that allows me to compare the effectiveness, because it costs me a lot more, perhaps, to make a rich-media ad than it does to make an HTML static ad." Now we'll be able to deliver all that.

Q: But how much do you think rich-media advertising will add to Excite's bottom line?
A: I wouldn't speculate about that right now.

Q: You've said before that you see the Web more as an application, a task-oriented tool. In what ways is Excite going to develop its content in a task-oriented way?
A: The first thing you'll notice on the entertainment page, which we've done on a number of other pages, is a little bar near the top that says "tools" with a little Excite emblem there. The purpose of that is to say that there are five or six things that people keep saying they want to do when they see the name of a channel. So in entertainment, it obviously will be listings -- TV listings, celebrities, videos, those sorts of things.

The first thing we've done to be more application-centric is to say: "Don't make [a Web page] like a table of contents, and don't favor news, per se, which is more precisely defined by other media like magazines and newspapers, but favor getting the content 'above the fold' that says 'How do I address your issues here.'"

So, in this entertainment [movie] application you can put in your Zip Code, and we'll tell you movies for your local area because that's the No. 1 thing people want to do. After that, there are links to reviews of that movie, a filmography of any of the stars in that movie, a synopsis of the movie, information about what else is playing at that theater, and so on.

Those are all applications, or task-oriented approaches to content. You can't get that in a dynamic fashion from any other media. That's what the Web does well, and companies like ours miss the opportunity if they're not thinking about content in that way because no other media can replicate that advantage.

Q: Is that the key then to getting subscriptions as a revenue model?
A: I doubt it. I suspect that the key to subscriptions is still going to be in forms of content that are not available to the next guy unless he pays. So it's not so much the application that helps me find a movie or a restaurant, or do my taxes, or buy a car, or connect with others online; it's the proprietary nature of content that drives subscriptions.

It has got to be super-proprietary, like very sensitive information about stocks or industries. Maybe someone puts together a news-tracking function, and I really want to follow poverty in Africa for six months, and I'm willing to pay $50 to do that. But I can't subscribe to 30 newspapers, and I can't be bothered setting bookmarks all over the Web.

There are forms of content that are going to be different from why people buy magazines that may work on the Web. But I don't think they'll be applications. I think they're going to mimic more traditional content.

Q: What is Excite's strategy going to be in developing this proprietary content? Is it going to be acquisition based, or...
A: Yes. We will not develop it, by and large. I think we're going to take a view that says it's not our expertise; we're not particularly expert in any kind of content. We've become experts at certain kinds of functionality -- we're an expert at service, we're an expert at personalization, we're pretty reasonable at chat, pretty reasonable at communities, those sorts of things. But it's not content. I mean, there the audience is the content.

So our view would be much more like network television, where we would license content from content providers, and do so only because the user experience was so disproportionately compelling to what we could have offered on our own that we've got to get people to stick around for longer periods of time, and/or charge more to advertisers or direct marketers.

We're not experts in sports; that's why we just did a deal with SportsLine. We're not experts in computer news; that's why we did a deal with Ziff-Davis. So you're going to see us do more of those things. We are the media distribution companies for those types of content, just as a network is.

Q: Would you consider bringing a content developer into the Excite fold? Or do you see the strategy being one where you only make licensing deals?
A: Let's go back a step. I think content is probably the most misused word on the Web. Content is buying tickets. But content is also the movie application. But content is also my stock portfolio. But content is also [the Web magazine] Slate. We use that one word to cover all that stuff. So I think you will see us focus on the kind of content that drives to the leveraging of our technologies, like applications. I don't know whether you would consider the movie application content or not, but we'll do more of that kind of stuff.

Q: Multimedia content?
A: Sure, given the broadband environment. But the data that is the content in those applications doesn't come from Excite. We go and license that data from a bunch of different people, in some cases multiple providers, to enrich that application. So, I guess you could say it's new, in the sense that we're providing the architecture and some of the editorial thinking behind what the thing should look like, and why. You know, the look and feel, how you navigate through it. But, by and large, we are not creating the content, in the way I think of it.

Q: Is aggregating content going to be a big focus?
A: Well, I don't know. I would say that one of the ways we make the service better for consumers is to focus on richer forms of content that keep people around for longer periods of time.

If we can't figure out how to do that on our own, we're always happy to partner with someone on the outside to do it. And we co-brand their site. You can see many examples of that in our service. We've certainly done it with Intuit and the other companies I've mentioned, obviously. So, there are plenty of examples of how we are willing to partner for content and give up some of our branding.

Q: Are you aggressively courting anybody in the content area?
A: If you're asking whether we need to buy these guys vs. license them, I think our first instinct would be to say that we would attempt to license content. Ownership doesn't seem to be strategic. It's building the programming value of the network that's important.

Content is fickle. For example, the best doctor database today isn't the best doctor database in two years, if we don't keep up with it. And if we buy it, we don't improve the chances of those people staying current with their content because we don't bring them anything internally -- other than distribution and exposure to the consumer market, and the look and feel -- but we don't bring them anything that makes them better researchers.

If you do sports content, and we buy you, we don't have anybody who knows sports; we're not going to make you any better. So I think, in many ways, those companies are going to be better if left on their own. But there may come a point of consolidation, which I don't think will come very soon -- it could come in 8 to 12 months or more -- where some of the bigger content sites need to make deeper alignments. Those could be just sort of longer-term licensing agreements, or they could be equity swaps with the portals, or they could be outright being acquired by a portal.

But at first blush, I would tell you that it doesn't seem to me to be strategic to own that content as long as you can license it. And if I can integrate it in ways that are productive for my user base, I probably would rather do that.

Q: You mentioned the words portal and network. Excite is becoming much more of a network, as you say, not a portal.
A: Yes. I don't know what the word "portal" means. We didn't think it up; I don't know who did. I'm not sure it's a particularly great word for what we do, but it seems to have stuck.

Q: But you're much more of a network, now?
A: Absolutely. I mean, the whole pitch that we hastily assembled when we merged [with Internet cable-modem access provider @Home] was the Media Network of the Next Century. Media Excite Network @Home. So, yes, we feel that we're more of a network because we also have the capacity for distribution in a more robust way than Excite had on its own.

Q: Explain what you mean when you say "the device becomes a commodity, and the value becomes the profile" in relation to direct marketing.
A: If I buy a computer, and let's just say I start with Excite when I go to the Internet, and I build a profile there, and over time I substitute my view of it as this sort of newspaper of me -- the thing I look at all the time that updates my content. If the computer goes away, I don't want my profile to go away just because I buy a new computer.

I want to know that it's resident somewhere -- on the server side, not on the client side -- so that it's indifferent as to what device I take it through. That's what consumers are telling us that they want: The simplicity of being able to say, this is me, this is my thumbprint on the Web, and I want it to travel with me lots of places. I don't want to have it die when I make a device change in my life.

So over time, obviously, the value is that features improve and speeds improve and lots of things happen to the market. But at the end of the day, the loyalty of the consumer is to the content.

My own view is that over time you'll see more and more pressure on device pricing. You've already seen sub-$1,000 computers, $25 cell phones, free cell phones if you sign up for long-distance service -- where the service is the revenue stream, not the device. All I was trying to say is that what I see coming is an acceleration of that trend. And you've already seen it, in fact, with the Dell-Excite deal.

Q: The Dell-Excite deal?
A: We have multiple-year deal with Dell [Computer Corp.], in which every machine that ships, in a certain series of Dell computers, features a co-branded Dell-Excite personal start page when you go on the Internet. And it's got all the personalization features of Excite in the main service; it just carries a lot of branding with Dell, and it carries a module, interestingly, for Dell users, so that there is Dell-specific content -- like chats with [CEO] Michael Dell, software updates from Dell, and user forums.

So we're programming into the window, so that essentially one module of personalization for Dell users is their relationship to Dell. One thing that Dell is seeking to do there is build an ongoing relationship with the consumer, not a momentary one at the point when the consumer buys the computer, and then at the point where the consumer reconsiders the next purchase.

Because what they've found out is that people don't think a lot about the fact that they own a Dell in between getting pregnant and needing to get pregnant again. I mean, in that whole space of time, which might be multiple years, the focus is on the screen, not the picture frame. So [Dell] said, "How do we get persistent exposure to these people and build a deeper relationship so that we know more predictably when they're likely to buy a computer again, either because they're telling us or because we're able to observe some behavior in a more direct-marketing fashion."

What will happen over time is that it will enable them to drive the price down, but preserve their margins, I believe, through other forms of revenue, like sharing commerce and advertising and direct marketing with companies like Excite.

Q: Are you not worried that, as content becomes more important and broadband becomes accessible to more people, that other companies may buy, rather than license, various entities so that more data content becomes proprietary?
A: If that happens, I don't think we'll be any less well equipped than anybody else to buy things. I suspect the combined @Home-Excite stock will not fare any worse in a downturn in the Internet market than any of the other big Internet stocks. So I think you'll find that our currency, over time, especially with the merger, is a stronger blended currency, and will leave us in good stead if the currency of choice to buy things is stock. If the currency of choice to buy things is cash, At Home has approximately $500 million in cash, after a convertible that they did in the fourth quarter; we have probably $65 million in cash. So the two companies, together, have more cash than any portal company that I can think of, at the moment.

So if that's the way the market turns, and you have to be a buyer rather than a licenser, the only question you have to ask yourself is, are your currencies depressed.

Q: But what will you do if something that was available on a licensing basis before becomes unavailable because of one acquisition or another?
A: As far as I can tell, especially in the world of the Web, which is very woolly right now, there isn't one of anything in content. The problem in the first chapter of these things is not the content; it's locking the eyeballs. It's getting people to make a habit out of using your site. And then sneaking up on them with richer forms of content. The way to sort of think about it is that we're the aggregator. It's a very different metaphor.

Q: It sounds like it's very much a double-edged sword. If you can provide something users can't find anywhere else on the Web, in terms of content, that's one way to get people's loyalty.
A: Yes. And I think that will increasingly be the case.

Q: But on the other hand, you're saying most types of content are a dime a dozen.
A: O.K. But what I mean is that content comes and goes. Let's say, for example that the Stones were recently on tour and we paid enough money, whatever that would have been, to get Mick Jagger to do a chat after every concert, every night, in the Excite chat rooms and nowhere else on the Web. After two weeks, of course, that was gone. But during that two weeks, we would have begun to have a branded affiliation with Jagger and the Stones, right? And we would have also got a momentary burst of traffic around a proprietary form of content.

This idea of momentary changes of content that other media aren't very nimble with, the Web can do quite well with. The purpose of it, to me, over time, is to allow proprietary content to deepen people's association of what your brand presents. A brand is a promise of something. It needs to be delivered every day in the service you provide.

So if I, for example, said a lot of our future, given broadband, and where Excite is by way of its identity, and maybe just its name, ought to be around entertainment -- maybe that would be a reasonable thing to do, to go pay what appears to be an exorbitant amount of money to go get a rock star into our chat rooms every night for 14 nights while they're on tour, speak to them live from all these locations, and get people to sample lots of other content around that that we have.

We're hung up in the conversation about how to adjudicate or value the content. But I'm saying, recognize that you have to also understand that content is different not only in substance and form on the Web; there can also be difference in spatial or time dimensions, too.

But I think it borrows from traditional media metaphors to be framing the questions the way that you are, and not recognize that the capacity for proprietary content on the Web will exist in ways that it doesn't exist in traditional media.

I've already said that one of the reasons it's hard to see the Web is that most people are condemned to ask the questions with a point of view. And the point of view has been formed by what they did before the Web.

So one of the things you negotiate with, as you start to move into this world, is the form of proprietary content-for-the-moment. Let's just say it's personified in the name of someone, and he or she says, well, how much promotion are you going to give it? They understand the fees on this thing are not going to be big for a while. But it's like if you're going to make a lot of noise about me, that's a kind of value.

Q: The last thing I want to talk about is your international strategy. How are you going to take advantage of the global market? How does that strategy mirror or differ from the one in the States?
A: Boy, that's a big question.

Q: You've stressed that personalization functionality is important to Excite, but I would think that overseas, you would have to figure in a certain kind of cultural adaptability.
A: More than that, it's turned out to be the availability of local content feeds. Our presumption is that you only succeed with your Web site locally by bringing rich local content into the site. So when you choose your newspapers from which we will gather news for you in your news tracker, or in your personal start page, you're darn right The Financial Times is in there. And The Economist is in there.

How you organize the page is a piece of technology that we can export from Redwood City, Calif., to London pretty easily. But we really felt that the value for people, the consumer, is that [the content] is truly local -- that rugby scores matter, and cricket scores matter, and all those sorts of things.

Q: Do you have in-country presences?
A: Yes, we do. We have about 25 people in Tokyo. We have about six or seven people in Australia. We have about 30 or 35 people in London, who serve Germany, France, the northern European countries. And we have three or four people in Italy. That's about it for distribution around the world.

Now, we have a team of engineers in Redwood City that protects and distributes the international code for Search and Personalization. We don't have personalization up in every one of those 10 versions of Excite that are localized, principally because brick by brick, we've got to make these content deals. I mean, there's nobody that's sort of organized the content in this way, so you've got to go in there and do it, like newspaper by newspaper, or magazine by magazine, or TV station by TV station, and start to bring these content feeds into it, you know.

Q: But the strategy is to stay local?
A: Yes. Well, there's two strategies really. One is to stay local by way of content and focus. The second is, where possible, give away some of our long-term upside in exchange for someone bringing us some strategic value -- distribution, content, marketing, in addition to funding. So we've done that with British Telecom. We've done it with Telecom Italia. We've done it with Liberty One in Fairfax, Australia, which has both Web sites and media properties. And we've done it with Itochu in Japan.

So we've done four joint ventures that account for 4 of the 10 localized versions that we publish. But the priorities are the same in each place. What we're trying to do is leave it up to the local managers what content comes in at what date. So they're researching it now, but they [other countries] are almost all where we were two years ago, where search [functionality] is very important.

The next thing is directory and channels. The next thing will be forms of personalization. The next thing will be communities and chat. Tripod brought in their home page building tools to five or six of the Lycos Western European sites over the last couple of months. And they've had very good success.

So it's just a question of giving the general manager enough flexibility and a clear menu of products that are all functional and scalable in the U.S., and then finding out what needs to be prioritized into that region. And we need to understand the justification. Why is that a good idea? Personalization now, before chat? Why not this one before that one?

But we try to leave as much as that up to the local people as possible, and let them be the judges of what their cultures are going to want. If I come in and try to tell you what I think would make a good Excite Japan experience, I'm sure I'd be wrong by 180 degrees.

Q: What about broadband ...
A: There are about 45 employees of @Home in a joint venture in Amsterdam. So they've got things under way. They've also got, from what we can see, not a lot of conflicts with what we've done. So our four big joint-venture partners don't appear to be in conflict with that. And, in fact, AT&T, as a big owner of @Home Excite, has joint-venture agreements with British Telecom. So that's actually a plus. I think everything is fine with AT&T and the other guys, with the exception of, possibly, Telecom Italia.