Q&A WITH INTERMEDIA'S DAVID RUBERG

From its start in 1986, Tampa-based Intermedia Communications' stated focus has been on meeting customers' needs, not pushing a particular "platform" or type of technology. The company was co-founded by Barbara Samson, who was 24 years old and selling long-distance service to business customers. After listening to numerous complaints from businesses about the expense and difficulty connecting to long-distance providers via local phone companies, her entrepreneurial zeal took over. "I saw there was an enormous opportunity to make a huge difference in the way customers were being treated and the way they were being serviced," she says.

Initially, Intermedia was a "bypass company," putting in fiber-optic cable and bypassing the telephone companies. It then took advantage of other opportunities in the widening telecom field, as a competitive access provider and then as a competitive local exchange company once the Federal Communications Act of 1996 was passed. The company had just $7 million in revenues in 1992 when it went public. In 1997, it posted $676 million in revenues, including its strategic acquisitions. Bear Stearns analyst James H. Henry estimates that Intermedia will generate approximately $750 million in revenues in 1998.

Samson left Intermedia last year to pursue other entrepreneurial and academic ventures. To get more insights into Intermedia's background and where it's going, Business Week Miami Bureau Manager Gail DeGeorge recently spoke with CEO David C. Ruberg, who joined the company in 1993. Here are excerpts from their conversation.

BW: When did the company start offering data network services?
Ruberg: Instead of being network-focused, we have always been customer-centric. We have always called ourselves an integrated service provider. That's why we got involved with data long before others did. Four years ago, we adopted a product and service portfolio that was customer-centric, focused on providing solutions, not platforms. We sold the solution. Customers didn't know if it was frame relay or ATM [asynchronous transfer mode]. They didn't need to know if they were going over concrete highway or asphalt highway. They didn't need to know any of that stuff.

BW: What is the difference between data and voice networks?
Ruberg: When you ask people about data, they think of numbers. They also confuse it with digital, the 1s and 0s. The real answer is that the old voice network was set up on something called circuit-switching. My call to you is on a dedicated circuit. The wide-area data network doesn't do that. Data uses the network and shares the network, and that's where we came up with the analogy of the highway. Think of a superhighway with six lanes. If you have lane A and Ruberg has lane B, if lane B is congested, I can't use lane A. If it is a packet network [as opposed to a circuit network], we collapse the lanes in the network and share it. "

BW: What does that do for the customer?
Ruberg: The user gets bandwidth on demand. They only pay for what they need. For example, if you needed to send data mostly in the morning, under the old system, you paid for the entire line. [But on a data network], if you use the network for only three minutes, you pay only for three minutes.

One day, you may only need one car's worth [of access]. On another day, you may need to pack the highway. But if you didn't call ahead and reserve another lane under circuit switch, you were out of luck. But with new data networks, it's not just data, but also digital so it can handle variances in traffic volume.

We had automated teller machines, grocery chains worried about efficiencies, huge data networks that were dial-up (under the old system) and not that reliable. A new technology was becoming available, "frame relay," which was a more robust and reliable data technology but also was a packet technology.

We looked at this and said, "hey, we can put networks in place, and they can handle data and then later they can handle voice and data." We saw in the company in 1992-93 that realization of this potential was not 10 years away, it was three or four years away.

BW: Who were your first customers?
Ruberg: A group of companies said they'd give us a try. The State of Florida also was an early customer. We submitted a bid, which we won in 1994, and it turned out to be a tremendous opportunity. We exceeded the user connections the bid anticipated. That was a backbone which allowed us to grow to compete against MCI, Sprint, and AT&T. Since 1994, we have been working to expand the network.... Our connected network has 3,000 switches [between Intermedia and its partners, Ameritech, BellSouth, and the other Bell operating companiess]. What we have now is a nationwide network. We are the fourth-largest provider of frame relay behind AT&T, Sprint, and WorldCom.

BW: What about voice over data networks?
Ruberg: We started offering that years ago.... It's developing the software and the signaling and numbering which allows us -- the industry -- to know and figure out in an astute fashion how we can get halfway using voice over frame and rest of way over circuit. We're developing a hybrid. When that happens cleanly, the world will migrate to voice over IP [Internet protocol]. Once we get the quality of service issues worked out over various networks and get various dialing schemes and make that automatic, then the service will migrate and start to migrate quickly."

BW: When will that happen?
Ruberg: It will start to happen at the end of this year.

BW: What percentage of customers or data networks now carry voice?
There's no way to know within the packets -- each packet has an indication that says "don't disrupt me, if congested, route me around." It's as if all the cars had tinted glass. There's no way to know. [One customer] allocated bandwidth large enough to run voice/data together. That was initially. Now, instead of expanding the number of lanes, we have packets that have identifiers. You put a flag on the car's antenna and then identify it as priority, like an emergency vehicle on a highway. It's not channelized, but we can control congestion on the on/off ramps.

BW: What will be first, faxes, video?
Ruberg: All services will start to go over once [the industry has] the protocols and gateways in place. With a digital network, you're talking about signaling. With a data network, you're talking about content. Today, we can send voice, data, and video over analog, nondigital networks, but it's not as efficient. So we are taking all of these signals, converting them to digital, and shoving them over the network. It's digital and packet that makes the difference.

On intranets, we're already handling voice, so we're seeing more and more voice, and faxes, coming into these networks. The expansion will allow more and more to come on. It's already happening -- it's expanding the reach, as we take it from intranets to extranets. Once we get a packet-switched network, then we will see voice going over that network, and we'll see voice prices coming down significantly.

BW: What new features could you have that you can't today?
Ruberg: For example, video phones. I could be drawing some of this out for you in diagrams. Now we would need a dedicated line. If we had bandwidth on demand and had phones, and we would pay only for the time used, one hour, it would make economic sense."

BW: Why is Internet protocol more efficient?
Ruberg: [In addition to the differences between circuit- and packet-switched networks], the traditional voice network had all intelligence at the center and was at these big switches. If we wanted to change technology, we had to change the center. An IP-routed network makes it easier to introduce new functionality, new products, and services at the endpoint rather than the centerpoint. It allows for more rapid introduction of technology to have the intelligence at the end, not the center.

There's one other element: directories. Today, you have to remember phone numbers. There is a "logical" number, a hard translation to a physical location. What I'd really like to do is remember your name and have the system find [your number] for me. A routed system remembers addresses by names, like Web sites, not numbers.

[With names on a network, that becomes the responsibility of the network and that currently chews up bandwidth], but a routed network is more flexible in reflecting change because people move around. [That's why we need] intelligence at the end and not the center.

BW: What kind of cost/price differences can customers expect?
Ruberg: Today, if you make a long-distance phone call from a business line (not residential) and you beat up your long-distance carrier, you can get rates of 12-15 cents a minute, end to end. Qwest has announced 7.5 cents per minute in selected cities and selected terminations, geared to residential, [though the] quality of service may not be same.

IP-based telephony utilizes the network more efficiently, and so the consumer, addressing cost not the price, will see long-distance rates come down by half, primarily for business. We will see a decrease in residential rates, but I don't know how much. A lot has to do with next year on access charges. Voice prices have dropped by 50% over last two years, they will drop by half again by the year 2000.

BW: What does Intermedia offer?
Ruberg: We charge by the minute, but we don't normally sell long-distance. We sell it with guaranteed services. I could sell you IP voice today, but you wouldn't like it. You can buy long-distance service to Venezuela, and one company will take 27 seconds dialing time, another will take 10. We'll ensure voice quality is there, we'll take 10, we'll go by cable then satellite. We talk to the customer on what quality they want.

Qwest's quality is not there. It is the technology, though, not the fault of the company. It's like the early days of dialing a POP or access code. I don't mean it's going to sound terrible, but there are different things that come into play. Certain people don't mind sitting in back of bus and having the bathroom next door to them."

BW: What is the impact of all this on the Bells?
Ruberg: It means they have to upgrade networks, and they're well aware of it. [Intermedia partners with many of the larger telecom companies, including a recent alliance with U S West.] It's a strange new world. These companies are not homogenous. They have a retail arm and a wholesale arm.

U S West has an "out of their territory" technology called "Enterprise," which is a combination of Cisco, Microsoft, Intel, and Intermedia coming together to provide solutions, not platforms, for their customers. If someone has a corporate headquarters in Denver and branch networks and want one-stop shopping and guaranteed functionality, they've put together a quasi-alliance to serve as these one stop shops, and we're part of it."

BW: What are you telling customers now about new technologies?
Ruberg: We don't want them to worry about it. If we sold them a platform, we'd tell them. We sell them a solution, so they shouldn't care. If they can't tell [the difference in quality], they don't care.

But with IP/voice now you can tell. So we will recommend they do only their intranet, and then they make the choice. You have to match their needs. You don't get them involved in the technology. If you do that, you're not a one-stop shop, you're a peddler of technology.