CAST your mind back to the beginning of 1994. Hardly a day passed without one global telecoms corporation or another announcing its marriage to a well-known cable television or film company. The child expected from these marriages made in media heaven was interactive television 鈥 a system that would provide 500 channels of movies and game shows to televisions across America and beyond, along with home shopping and banking, information and education services. No longer a passive couch potato, the customer would become an active partner helping to create the entertainment and information environment.
But a year after the hype peaked, little has materialised. Some of these marriages went straight into annulment. One casualty was the $4.9 billion joint venture between Cox Cable and the Texas-based phone company Southwestern Bell. Another was, the mother of all mergers, the $33 billion deal between Tele-Communications Inc, the largest cable company in the US, and Bell Atlantic, its most active and ambitious telephone company. Of those unions that survive, many have remained sterile so far. The Disney organisation鈥檚 interactive television deal with the telephone companies Ameritech, BellSouth, and Southwestern Bell, announced last August, is so complex that they are still working on the contract. And nowhere does anything seem to be happening quickly. In Orlando, Florida, media and entertainment giant Time Warner has begun tests of a sophisticated system combining high-speed communications links and high-powered computing technology. But these tests did not start until last December, nearly a year later than originally promised. A much vaunted trial by telecoms company US West in Omaha, Nebraska, is still three months or more away from playing to real customers. Pacific Bell鈥檚 test in Milpitas, California, has been cancelled.
Around the world, the armies of financiers, television executives and programme makers struggling to build a brand-new industry are finding numerous obstacles looming up in their path. The technical difficulties have become more obvious and the solutions fuzzier as disparate technologies compete for adoption as the true interactive television. Marketing experts offer wildly differing, and sometimes dire, forecasts about what the demand for the product might be when it finally becomes available. Meanwhile, some observers are even claiming that the whole thing is a blind alley, and that what people really want is something more like the Internet.
Advertisement
In the US, where excitement over interactive television was greatest, only three projects have so far come online with interactive programming for real, paying customers 鈥 and even these are all limited in significant ways. More than two years ago, in January 1993, GTE of Stamford Connecticut, the world鈥檚 fourth-largest publicly quoted telecoms company, switched on a commercial service that offers home shopping, play-along games and information 鈥 but no movies. This service will be carried over ordinary telephone lines using Asymmetric Digital Subscriber Loop (ADSL) technology, which allows compressed digital video signals to be sent at the same time as ordinary analogue telephone signals. The company now offers the service to some half a million homes in New England, but can claim only 4000 actual customers.
Testing times
Southern New England Telecommunications started a trial bringing interactive digital video services over broadband fibre and cable lines to 370 homes in West Hartford, Connecticut, last April, and is now expanding that to 150 000 potential customers. And Time Warner鈥檚 Orlando trial, mounted in conjunction with telecoms companies AT&T and Scientific-Atlanta and computer company Silicon Graphics, is offering what it calls a Full Service Network of movies, news, games and even an interactive health system. It hopes to have 4000 subscribers by the end of the year.
Other projects have still not progressed beyond the promises phase. US West is doing 鈥渢echnical testing鈥 of a similar system in Omaha, and intends to start taking customers this summer. Companies elsewhere in the US have similar projects on the stocks, and there are various plans for interactive television services throughout Europe and Asia.
British Telecom is already operating a trial system that reaches 60 of its employees in Ipswich, Suffolk, and expects to expand that to 2500 paying customers in the summer (see: 鈥淓uro channels test the water鈥). Deutsche Telekom, the German telecoms provider, and national television broadcasters in France, Italy, Sweden and Norway, among other European countries, are likely to start trial interactive systems this year.
Meanwhile Singapore Telecom, as part of its push to build a state-of-the-art broadband communications infrastructure, has just announced the start of trial of its own video-on-demand (VOD) project this summer. Singapore is a natural ground for the new technology (Culture clash in cyberspace, New 杏吧原创, 25 March). All its telephones are already digital, and all its exchanges are linked by optical fibre. Some 99 per cent of the population owns a colour television and 84 per cent owns a VCR. By 2005, every home in the country will have a fibre-optic link. A pair of Japanese companies, Mitsui and Fujitsu, are the project鈥檚 leaders. Yet in Japan itself, no company has yet signed on the dotted line to bring interactive television into the home.
They may be wise to be hanging back. It is difficult 鈥 some would say futile 鈥 to attempt to build a business on a technology that is changing so fast. 鈥淭he first one鈥檚 not always the right one,鈥 says Rich Frank, chairman of Walt Disney Television and Telecommunication, which has been slow to enter the game. The technical complexities of this new landscape sometimes seem limitless. Tom Mandel of SRI International, a consulting firm in Menlo Park, California, expresses it as a 鈥減lug and play鈥 problem of gargantuan scale. 鈥淥rdering up a network, some ATM switches, set-top boxes, video servers, and the like doesn鈥檛 mean that you can just plug them together and they鈥檒l work.鈥 In fact, just how to plug them together is as big a problem as what to play on them 鈥 and to make matters worse, the two problems are horribly intertwined.
The first step
In TV programming terms, the traditional cable networks that are well established in the US and many European countries hardly differ from ordinary television: both 鈥渂roadcast鈥 from a central studio on a schedule that the consumer cannot influence. One of the questions that strategists are now pondering is how much influence and interactivity consumers really want. The bottom level 鈥 the first step up from the traditional cable systems 鈥 merely injects a small amount of choice into the traditional scheme. The cable conveys a constant stream of small packages of information, such as games, stock reports, weather or news. An introductory programme catalogues the services and provides the viewer with a series of choices displayed on screen. With a remote control not much more sophisticated than the one you already use, you choose one category from the menu. That might bring up further menus until you have singled out the programme you want. The next time that programme comes past in the bit-stream, the set-top box grabs it, installs it in a memory chip, and begins to play it for you. For a small programme, the delay will be so small as to be practically unnoticeable. A chip in the box logs which programme you have used, and every month a computer at the cable system鈥檚 central office sends a query over the cable to the box. The box adds up the cost of everything you have used, sends the reckoning back to the central office through the cable鈥檚 鈥渂ack channel鈥, which handles all communication back from the set-top box to the central system, and the company charges you accordingly.
Simple system
Playing movies this way is called near-video-on-demand鈥 (NVOD). The movies come in a continuous stream from large hard discs on a video server, through a high-speed bus that connects the various components, to an output stream conditioner, which converts the signal to the speed and level of compression that your set-top box is expecting. The movies start in continuous rotation, each one starting anew, say, every five minutes. If you decide to watch one, you will never have to wait more than a few minutes for it to start. You get charged only for what you watch. Technically, the system is simple because all programmes and movies are playing all the time, ready for any customer to start watching. There is no need for complex switching to start only those movies and programmes demanded by particular customers, and to send them only to those who chose them. It will also work without bringing fibre-optic links into every household. The optical fibre can stop at a junction box in the street, or in the basement of your apartment block, leaving an ordinary twisted pair of copper wires to take it from there. The disadvantage of this system is that the choice it offers is relatively restricted. Since all the movies and programmes have to be online all the time, it can only supply a limited number, not a huge library.
True video-on-demand (VOD) calls for a different architecture. When you click on a choice on your television remote control, the set-top box sends a query into the system through the back channel. The system routes the request to the video server, where the video is stored in a compressed format. The medium the programme is stored on would depend on its popularity and its size 鈥 which might range from a two-hour movie to a 30-second description of a restaurant. It could be on CD-ROM, digital audio tape, laser disc, videotapes in a jukebox-like machine, a large array of computer hard discs, or a random access memory chip. The server finds the material you have chosen, and sends it back through the cable system to the set-top box, which decompresses it and starts to play it, usually no more than a second after you click on the remote.
A network is born
This requires a lot more memory and processing capability in the set-top box, and greater capacity in the back channel. But the most important difference is that now the system must be 鈥渟witched鈥: it has to be able to send the video only to the particular box that requested it. For this, most systems will use the new high capacity 鈥渁synchronous transfer mode鈥 (ATM) switches. Putting switches into a cable system moves it across an important divide, transforming it from a 鈥渢ree and branch鈥 broadcast system like a traditional cable system into a network, like a traditional telephone system.
But suppose you want to be able to do more than choose something and watch it? Suppose customers want to play along with the experts, calling plays in a football game, trying to beat the contestants on a game show, or voting on a political debate? Suppose they want to enter a virtual world in which the computer interacts with them? Now the central computer and video server has to react much more quickly to a much wider range of demands from the customer. The architecture of the system remains much the same, but all the equipment, from the server to the ATMs to the set-top box, must be much faster and have much more capacity. A fully interactive system would need one flbre for each 2000 homes, assuming that no more than one in five of the homes on each fibre would use the system at any one time. To deliver compressed video and interactive responses through a back channel, the fibre鈥檚 bandwidth would need to be 400 megahertz, and it would have to be fitted with some 67 modulators 鈥 quite some system.
Finally, what if customers want to interact with each other? What if you would like the other inhabitants of your virtual world to be other real people, exploring the world from their own homes and maybe stopping to chat, flirt or fight with you? Then the amount of information shuttling back and forth through the system rapidly multiplies. Every player has to get the information about every other player鈥檚 movements. At this level of complexity, managing the system becomes a major feat.
Another big question then arises: where should the intelligence, the computing power and the memory go? At the edges of the system, on the customers鈥 televisions, or in the centre, on a high-speed computer at the cable company鈥檚 main office?
Suppose the virtual world of a particular interactive game includes a mountain scene. The image of a mountain can be stored on the central computer and sent down the line to be displayed on the player鈥檚 television. But images take up a lot of data. A single still picture of the Himalayas may contain 0.5 to 5 megabytes of information, so this strategy puts heavy demands on the system鈥檚 data 鈥減ipes鈥 鈥 its optical fibres, cables and buses 鈥 and its switches. A video server鈥檚 stream controller handling 400 customers at 4 megabits per second will not be able to pump data through the system fast enough to satisfy the demand if a significant number of those customers are taking 3D trips through the Himalayas, or even through virtual shopping arcades. With all the data stored at the centre, the only solutions are to cut the number of customers on each server, or add even more capacity to keep the system up to speed.
Bigger boxes
But there is an alternative: each customer鈥檚 set-top box could be given the capability to draw its own mountains. Then only a few lines of computer code giving the instruction 鈥減ut a mountain here鈥 would have to pass down the line, and the data pipes could be kept a lot smaller without any risk that they will be overloaded. However, the set-top box will have to have much more memory for storing mountain-making instructions, and more computing power for turning those instructions into mountains on the screen.
So the set-top boxes, which every consumer will need, will become a lot more expensive. To display video sent to it in the old-fashioned way, a set-top box does not need more computing capacity than one of today鈥檚 Nintendo game-playing machine. But to generate its own video, on instructions from a central computer, it needs the power of a mighty professional graphics work station, such as a Silicon Graphics Indigo 鈥 which is just what Time Warner has built into the set-top boxes in Orlando.
The choice of architecture is pivotal, and its consequences are not merely technical, but financial, political and even cultural. Put the intelligence and power at the centre, and you end up with traditional broadcast networks: powerful and wealthy corporations or national organisations such as CBS or the BBC setting the agenda for much of the global discussion. Take the alternative 鈥渋ntelligence round the edges鈥 model to its logical conclusion and you get something that looks like the Internet. The interactive television services now under development lie at various points between the two extremes.
Clearly, the companies pouring money into these systems believe that they will recoup their investment at the very least, and that people will be prepared to pay for what they offer. But not everyone is confident that they will. 鈥淭here are no data showing that you can make money at this yet,鈥 says Donna Hoffman of Vanderbilt University in Nashville, Tennessee, who studies marketing for the new media. Questions remain over how much of the system鈥檚 computing power should be put in the hands of the consumer. How much would the consumer like to have? How much would be confusing? Time Warner鈥檚 set-top boxes, for instance, are more powerful than most home computers 鈥 but consumers do not get a keyboard to access all that power, only a set of fixed menus. 鈥淯ntil the results of the tests come in,鈥 says Mandel, 鈥渘o one really knows what consumers will do with the new services.鈥
Lessons of history
If history is a guide, it is not a particularly encouraging one. In Columbus, Ohio, in the late 1970s and early 1980s, Warner Communications (one of Time Warner鈥檚 predecessors) ran an experimental interactive television service called Qube, which offered home shopping and information. The company declared the trial a success, but apparently not enough of a success for it to continue or go on trial elsewhere. Other videotext and teletex services in the US, Canada and Britain in the early 1980s failed to attract a critical mass of paying customers. Pay-per-view events over traditional cable, which deliver scheduled major events such as boxing matches only to viewers who sign up to pay for them, have rarely attracted more than 10 per cent of the potential audience, according to Hoffman.
Preliminary market surveys deliver conflicting evidence. Many business plans assume that VOD is the 鈥渒iller application鈥, the service that potential customers will find irresistible. Also in contention for the title of killer ap are home shopping and other forms of entertainment, such as games. Dataquest, a Californian research firm, estimates that VOD will become a $5.2 billion per year industry in the US alone as early as 1997. An Associated Press survey commissioned by the three major American broadcast networks and the National Association of Broadcasters showed: 鈥淥f proposed interactive services, 30 per cent of respondents were interested in 鈥榤ovies on demand鈥. No other service offer came close: information got a 12 per cent response, sports got 12 per cent, video games 11 per cent and home shopping 7 per cent.鈥 One question these surveys did not apparently ask, however, is how much people would be willing to pay for these services.
By contrast, a survey of 600 Americans published last October by the computer magazine Macworld showed movies ranking only 10th out of 26 possible uses of the 鈥渋nformation highway鈥, with other shopping and entertainment uses clustered near the bottom of the list. People were much more interested in going to school, voting or using reference materials online. Three-quarters of the people surveyed said they wouldn鈥檛 pay more than $10 per month for any service except for at-home education.
Tom Grieb, general manager of GTE鈥檚 Main Street interactive service, which does not carry VOD, accepts this view. 鈥淲e don鈥檛 believe there is a killer ap,鈥 he says. 鈥淭o really embrace this, people have to change their behaviour. That usually takes a long time. It鈥檚 just like automatic teller machines, which were widely held to be a failure in the mid-70s.鈥
Nobody doubts that interactive television represents a big investment. George Forrester Colony, head of Forrester Research in Cambridge, Massachusetts, estimates that it will cost $1700 per home to bring interactive television to Americans. It will take 10 to 15 years to pay back that investment, let alone make a profit, Colony says. He foresees maximum market penetration this decade of only 5 or 6 per cent.
The corporations planning this huge project see increased consumer choice as the driving force. The old model was: 鈥淲e broadcast it, you watch it.鈥 The new model is: 鈥淲e broadcast it, you choose it and watch it whenever you want.鈥 The idea that the consumer might want to do more 鈥 make social contacts, start clubs, send messages, interact with other people, rather than with their television 鈥 finds few takers in the industry. Some observers think this is a fundamental mistake. 鈥淭hey don鈥檛 know what they are doing,鈥 says Hoffman. She believes that the planners who are sticking with a one-to-many model are missing the point. 鈥淲hat the people who are running these systems need to know is that people don鈥檛 come to them for choices, they come for deeper psychological reasons.鈥
Sense of community
Hoffman points to the popular computer network America Online, which has huge retail sections. 鈥淏ut few people use them,鈥 says Hoffman. 鈥淭he system earns a minuscule amount of its income from the shopping venues. People spend the majority of their online time on e-mail, chat rooms and forums. They want to talk to each other. People are searching for a sense of community.鈥 One of the key attractions of the World Wide Web and other Internet systems is that they allow the user to be an information provider as well as a consumer, she says. 鈥淐urrent implementations of so-called interactive multimedia like interactive television do not offer that.鈥 Allowing people to choose their package of pre-produced entertainment is not real interactive television, Hoffman says. 鈥淚t鈥檚 just new and improved television.鈥
Hoffman is not alone in seeing the limitations of the current rash of proposals. 鈥淭he set-top box and what it will carry, is only the tip of the iceberg,鈥 says Phil Anderson of the Tuck School of Business at Dartmouth, Massachusetts. 鈥淭here is no question that the many-to-many model will ultimately prevail. The question is only how long that will take.鈥 Which leaves an intriguing question: when fully interactive broadband services can be channelled into every home, will they be provided by the cable consortia that leapt into the limelight last year, or will services that grew up on the Internet have already swept the field?
Euro channels test the water
THIS SUMMER, 2500 households in the Ipswich and Colchester area of eastern England will have the option of switching on their televisions, picking up a remote control, tapping in a number, and picking from a range of video titles presented on their screens. These homes are part of the first market trial of British Telecom鈥檚 video-on-demand service.
Just over a year ago BT made a high-profile announcement of its plans for VOD (鈥淐an telephone kill the video shop?鈥, New 杏吧原创, 27 November 1993), which would allow customers to select films or carry out certain banking services across existing phone lines. BT is banned by law from using its network to broadcast TV programmes.
At the time, there was much talk of how the company was tweaking existing technology to meet the demands of tomorrow鈥檚 consumers. But though the trial has yet to start, the euphoria has already evaporated. Alan Rudge, BT鈥檚 head of research, is frank about the practical limitations of the Asymetric Digital Subscriber Loop technology, which is used to transmit programmes on demand over the twisted pair cable that links most homes to BT鈥檚 trunk network. 鈥淎DSL is not economic for 30 million customers. We have 拢15 billion, the price of two Channel Tunnels, earmarked for an optic fibre highway.鈥 This will eventually provide a fibre-optic connection to all residential business addresses. 鈥淲e don鈥檛 want to invest too much in an interim solution,鈥 says Rudge.
BT鈥檚 caution does not, however, spell the end for interactive television in Britain. A number of other companies are experimenting with this kind of service. The cable company Videotron is working with Carlton, Britain鈥檚 biggest commercial television station, on an interactive London News Network. The plan is to allow subscribers to choose between news, and up-to-date reports on weather, traffic and crime, using a remote control handset.
Another British company, Two Way Television, is working on a system that will offer viewers quiz games through an enhanced Teletext-style interface. And Rupert Murdoch鈥檚 New Datacom company has filed a patent application for an 鈥渋nteractive gaming system鈥. None of these companies, however, is keen to talk about its experiments.
Meanwhile, across the Channel, France Telecom plans trials of interactive television this year, using technology similar to that proposed by BT. Similar experiments will take place in Norway, Sweden and Italy, where the American regional telecoms company Bell Atlantic plans an ADSL trial in 2000 homes. Deutsche Telekom intends to use a mixture of ADSL and optical fibre cable to deliver its interactive television service.
But even these experiments have their detractors. Robert Hamersma, managing director of the Media, Programming and Ventures division of the Dutch electronics company Philips, sees the price of the electronics needed to connect a television set to a telephone line as a crucial factor. The threshold price is $750: systems will become economically attractive economically when they cost less than this magic figure, but that will not happen before the turn of the century, Hamersma says.
So some Dutch companies are experimenting with an alternative approach. CD-Matics of Amsterdam has developed a remote shopping system which builds on a Philips CD-i interactive video player, which potential subscribers may already own. Video pictures of the goods on offer are stored on an interactive CD, which is sent to customers by the department stores taking part in the scheme. Customers put the disc into their CD-i players, and then use a pointing device to select their way through a series of menus and pictures of the goods they may purchase. After the customer has narrowed his or her selection to a few items, the CD-i player automatically contacts the store via an ordinary modem and phone link, and the store sends back information on prices and availability. The customer can then make a decision and tap in an authorisation code. The goods are delivered by mail or courier, and the price is debited to the purchaser鈥檚 account.
The Saraxa Group, a media company based in Helsinki, Finland, is also using CD-i as an economical way to provide interactive television services. Television viewers in Finland can use this system to participate in popular video games by remotely controlling a machine at the television station. Viewers send their requests by tapping numbers into an ordinary touch-tone phone. Whoever reaches the remote machine first, gets to see their game selection on their television screen. Saraxa is negotiating deals to use the same technology in Malaysia, Argentina, Spain, Portugal and Britain.