Former Time Warner Cable senior vice president of technology Jim Chiddix now heads up AOL Time Warner Inc.'s Interactive Personal Video group, part of former MSO chief Joe Collins's new Interactive Video Division. The group operates from Time Warner Cable offices in New York. His staff currently consists of nine engineers and business-development types, although a couple of dozen employees will likely tackle what lies beyond video-on-demand: the ability to offer any video content ever created at any time. Well, almost. Broadband Week editor Matt Stump recently caught up with Chiddix. An edited transcript of their conversation follows:
MCN: What have you been doing the past six weeks?
Jim Chiddix: My charter is very clear. We're building a new, video-centric interactive service, which will be available to all of our customers who have digital set-top boxes. The idea is to take the bandwidth we have in our cable systems and our installed base of digital set-top boxes to their logical conclusion, which is building on our ability to offer an individual video program to an individual customer.
I'm developing the next on-demand service, which makes use of low-cost storage to provide some additional functionality. It's a much larger and lower-cost on-demand library.
MCN: How does that get achieved?
Chiddix: I think scale and time are achieving that. The server vendors are going to be able to deliver much lower-cost systems in the next one to two years at much larger scale. So instead of provisioning a few thousand streams in a cable system, [it's] having the ability to provision tens or even hundreds of thousands of streams in a cable system at a much lower per-stream cost than currently.
MCN: Let's get into personal video recorders: What's your view of the network-based versus set-top based PVR? Will there be a hybrid approach?
Chiddix: I don't think there will be a hybrid approach. I think both will exist, but I think for a given customer, it will be one or the other, probably.
The real difference is [the] time frame. The quickest thing to do is to build a me-too, DBS-type, local hard-drive solution. And we've committed to buy such boxes, and those will appear shortly. And those will do exactly what you can do with a DBS service today — which is useful, but it has a number of constraints.
By the same token, if you centralize at the headend, we believe you can get the cost per user much lower. Instead of costing an extra $150 to $200 per box, we think the shared costs of a headend solution will be much lower than that.
MCN: With a network-based PVR service, you could use the 2000 series of boxes out there right now, I presume?
Chiddix: That's one of the big points. You don't have to wait for the uptake for a whole new generation of hardware to get the service out there. It just shows up on the existing installed base of DCT-2000s and Explorer 2000s.
MCN: Can today's server technology and cable-industry server vendors handle network PVR capability? What's it going to take, technology-wise?
Chiddix: It's going to take people building much larger servers, using commodity technologies whenever possible. Obviously, we're talking [about] something that is quite a bit less expensive than today's VOD servers, but we think that's quite achievable.
The economics for today's VOD servers [are] fine for a high-value service like movies-on-demand, or HBO [Home Box Office]-on-demand. But the current volume of VOD servers is very, very small. We have three systems installed and we're going to install a dozen more. And those are provisioned to handle peaks in the single digits.
We believe you can build much larger servers using commodity components with much lower costs per stream, and that lets you begin to build some interesting businesses where you don't have to charge a couple bucks for a movie or $7 or $10 a month for a subscription.
MCN: Ninety-nine cents a shot, and I can make money?
Chiddix: Maybe less than that. Maybe we start offering continuing education courses, things that are not used very much, one at a time. But in the aggregate, they form a narrowcast service that really has some power.
MCN: Many of the basic networks seem mesmerized by the idea that subscription VOD will work better for them. But if you had 20 basic networks with some form of SVOD service, for today's servers, that seems to be a mind-boggling amount of content that wouldn't work under today's economics. From what you're talking about, you want to be able to handle 20 SVOD services from 20 different content players. Could you make those economics work?
Chiddix: Yes. We believe so. Remember, it's hard to overemphasize, today's VOD equipment is being sold into a tiny, tiny market by the standards of computer-industry volumes.
MCN: In the world you're looking at, is there a ton of server capacity at the headend, and how widely would you distribute servers throughout the system?
Chiddix: It's not clear yet. That's purely an economic question. The two extremes are [that] you centralize everything and originate every stream from the headend, then you have a lot of transport costs. The other extreme is you have all the servers in the hubs and you distribute assets to them, but all the streams originate in the hub to minimize transport costs.
The answer may very well lie in between. We have not yet worked our way through that financial equation.
MCN: Does your experience in Columbia, S.C., with the session set-up and breakdown and SVOD usage, alter anything in your thinking about network PVRs?
Chiddix: Setup and tear-down isn't traffic. That's just software in the controller. These services do generate more traffic, more streams. If you have a movie-on-demand service, you watch a few movies a month, so for a few hours a month you've got streams in use. With HBO-on-demand, you may use many more hours than that every month, which means more system loading.
Our existing VOD vendors are interesting players in this world, but I think we may be surprised at the more traditional computer industry players who show up.
MCN: Can you make it worth Sun Microsystems Inc.'s while to come into this business? Is there a number that is their sweet spot?
Chiddix: I don't know what the number is, but I bet you this kind of video-on-demand service will get their attention.
MCN: How do you look at the world of gaming and music content, since you have set-top boxes and cable modems through which this content can flow?
Chiddix: Gaming is a difficult puzzle. The gaming world changes so quickly that owning the gaming platform is at the very least a risky strategy for cable operators. We've been through this with the Sega Channel. By the time we really got it deployed, the fashions had changed and the kids were off to a more powerful platform rendering all that Sega Channel hardware into junk.
On the other hand, the fact that we have a very high bandwidth, very low latency connection service with our cable modems is pretty interesting to enable network games. Set-top boxes could be part of that equation as well but it's more likely we'll benefit through cable modems on the gaming side.
The music side is different. I'm fascinated by fact that we've got millions and millions of digital set-top boxes hooked to the television set in our customers' homes and also, in many cases, in close proximity to the home-entertainment center. Set-top boxes are an intriguing thing to think about with regard to audio service and to on-demand audio services, which could share many characteristics with on-demand video.
MCN: I can't imagine an S-A 8000 box in an Time Warner Cable system somewhere that doesn't have a big MusicNet icon.
Chiddix: Why an 8000? Why not a 2000?
MCN: You can do that on the 2000?
Chiddix: You're going to use MPEG [Moving Picture Expert Group] transport to move music instead of video, the only question is, are you going to output it as analog or as secure digital through the USB [universal serial bus] port? There are intriguing possibilities there. We haven't thought our way all the way through that, but I think that it's an interesting topic.
MCN: Is there enough memory in the 2000 box, or do you need a browser to make AOL TV work at that level?
Chiddix: Depends on how you define AOL TV. The subset of AOL TV which could be chat, messaging and so forth, there's no reason that couldn't be done on S-A's 2000 boxes. Running a browser is more challenging. To run a browser in a set-top box you'd need more horsepower in the box itself.
On the other hand, I'm not aware that anyone has proven that a browser on a television set yields a particularly compelling experience.
MCN: I know the copyright issues aren't in your purview, but a lot of people look at network PVR and think that rights issues would prevent it from happening anytime soon, although there is a lot of content present in parts of your company.
Chiddix: The rules that apply to PVR, whether it's on a set-top box or in the network, are far from clear. A lot of lawyers are going to put in a lot of hours before it is clear.
MCN: Do you go to your content brethren and say, 'I'd like to do a network PVR test in six months, can I get some content?' Is that a doable proposition?
Chiddix: Sure, although six months is too soon for the servers I'm talking about. I'm looking further out. But I'm sure that we will have content.
MCN: Interactive television, walled garden, all those things you got to leave behind … how much of that plays into your division? What about electronic commerce or interactivity, for example?
Chiddix: You mean reading text off the television? I'm not aware that anybody is making any money with text on the television. Am I missing something?
MCN: I just wanted to make sure I asked the question.
Chiddix: The one kind of interactive television that makes money today is video-on-demand.
MCN: Aside from continuing education, what other content categories might work in an individualized video world?
Chiddix: Anything that exists in film or video form has the potential to be sold on-demand. Yes, there's a whole interlocking system of rights, but on-demand usage brings great convenience to consumers and that means it brings value, and that value is something that should be valuable not just to the deliverers but also to the owners of the content. What we have is technology that lets us tap into new functionality and new value.
MCN: Anybody hitting the four major professional sports leagues yet?
Chiddix: It's way too early, although there is value there. There's another whole area where there's value and that is advertising.
MCN: Is Madison Avenue ready for this?
Chiddix: The Bowflex [home-fitness machines] people certainly are. There is power in the ability to delivery to either demographically targeted groups or individuals, messages from people with goods and services to sell and do it in a way that is precisely targeted and audited. There's real potential value here. The question is how do we turn that value into something Madison Avenue can actually use.
Remember, people have had the ability for quite a while to change channels during commercials and, when they videotaped things, to skip through commercials. So Madison Avenue is used to the idea [that] they have to grab people's attention.
In a world where people are using PVR functionality, that's even more true. But one way to counter that is to be able to deliver to people advertising that truly does interest and engage them. For example, you can foresee a world where during a program that you're watching, a short commercial appears and it is sharply targeted towards your interests, within the constraints of privacy law. Within that commercial is embedded a key which lets you — if you want to find out more — pause the programming you're watching using PVR functionality and then come back to the program when you're through watching the infomercial. The infomercial can be as long as you want it to be.
MCN: Some people might wonder why you would want to put so much effort into a project that may not be economical or scalable right now.
Chiddix: We think it is economically viable, and what's more, we're not talking about anything that needs any invention here. The technical mechanisms to do on-demand are quite well understood. The bandwidth is there because of the HFC [hybrid fiber coaxial] upgrades that have been made. What's missing is the will to break the chicken-and-egg relationship between volume and price, and the will to build the management systems, which can allow things to happen at this scale, and the imagination to build new services on top of that platform.
MCN: Some people might say consumers aren't ready to pay a buck here and buck there, hence the allure of the SVOD model. Some may say we can't have a 500-title price list out there.
Chiddix: There are lots of ways to price it. You can price the whole thing in the context of a couple of different subscription tiers. It does not have to be confusing. There are some interesting user-interface challenges here. We need to use an intelligent user interface on screen to let people navigate and manage this much choice.
MCN: When will we see a network PVR test?
Chiddix: I think we can have technology being tested in '02 and I think we could have a system working in '03. Obviously, we'd like to do this as quickly as we can. We don't how long it's going to take, but we think it's important to do strategically.
As larger and larger hard drives go into DBS receivers, it's very important for the cable industry to try to use its bandwidth to provide a more compelling experience than that which has the constraint of local storage. I think a me-too strategy of building essentially a DBS receiver with a hard drive that happens to run on cable is not the optimum competitive response.