From Computer Business Review, a sister publication

America’s business leaders are road-weary. Last month, Chief Executive magazine talked to a thousand US CEOs and found that, compared to research carried out 10 years ago, they are now spending much more of their time traveling – to company sub-offices, to shareholders meetings, on customer calls and to conferences. However, the survey also showed that executives were well aware of a single technology that could dramatically reduce their journeys. Two-thirds recognized that videoconferencing systems, the combination of video camera, computers and communications, may be the answer to their too-frequent flying. That should have sounded like music to the ears of the hundred or so vendors active in the videoconferencing industry, all of whom have been patiently waiting for their market to catch fire. But perhaps influenced by ‘Multimedia Computing will Change your Life’ articles found in in-flight magazines, CEOs and other senior managers have been making those kinds of noises ever since AT&T launched the first videophone in 1970. The trouble is, their sentiment has yet to turn into the kind of stampede to videoconferencing technology that the market and many of its analysts has anticipated for half a decade. Trying to stir the notion that every major corporation and public organization is embracing videoconferencing, vendors like to quote sector analysts, like the Gartner Group, which say that the market for visual services will be worth an astonishing $7.5 billion next year. Behind the scenes, the same vendors find that figure a little embarrassing, given that their own realistic estimates for 1995 revenues for all companies in room-based and desktop videoconferencing products and services still falls shy of the $ billion mark. The fact that most corporate executives have still to convene their first video meeting has the sector in contortions of explanation. After all, vendors proclaim, the various hardware and software components that make up a videoconferencing package are in place – and they work. Not only that, they are increasingly able to mesh with groupware functions such as document co-editing, file sharing and ‘white-boarding’. Key standards are also now set in stone and have paved the way for communications between products made by different manufacturers. The necessary telecommunications – in almost all cases public Integrated Services Digital Network (ISDN) lines – are available in all major European countries, and throughout most parts of the US. And as semiconductor and other component prices have fallen, the cost of videoconferencing systems has slowly been coming down.

Points of resistance

Nevertheless, the barriers are real, and center on issues of quality, cost,connectivity, applications and psychology. The market is very much technology rather than demand driven. Picture and audio quality varies widely from system to system, with displays ranging from a jerky picture with sometimes-unsynchronized audio (the lower-end systems need to squeeze the 10 to 20 frames a second video image into a window a couple of inches wide in order to maintain a level of fluid movement) right through to TV quality relay of multiple talking heads at 30 frames a second. The former is obviously of little use to people keen to conduct important online meetings. And for many customers, the rule of thumb is that if systems do not allow them to see the difference between a smile and a grimace, there is probably a good case for getting on a plane. What is clear is that the components for a video system are all there. The group of standards – H.320 for videoconferencing, including the key video element H.261, and the T.120 standard for data interchange – are there also. Unfortunately, suppliers’ implementations can vary both in quality and in depth of adherence. Half a dozen years down the line, says the Forrester research group, when high bandwidth networks are ubiquitous and PCs come equipped with a video camera and software that can handle all aspects of videoconferencing, the additional cost for moving pictures will fall to zero. In the meantime, users will need to throw high-cost technology at the problem and pay for the privilege. The prices of installing and operating videoconferencing capabilities are high. A multinational, which has three regional centers around the world that it wants to hook up regularly, may lay out $100,000 to equip three board rooms for videoconferencing – and that does not include the installation or running charges, or the cost of using a high-speed ISDN line for several hours a week. Point-to-point connections between individual PCs can be established, again by hiring ISDN lines, but that again is costly and inflexible. Even more expensive, the prospect of video-enabling a LAN is, in most cases, prohibitive, especially when stacked against the business benefits. The psychological and practical problems of operating within this new medium are equally important, but less understood. Most of the photo-opportunities and the early demonstration projects for videoconferencing have centered around desktop systems. Typically, a main video processor card or unit, known as a codec (coder/ decoder engine), takes video streams from a camera mounted on top of a multimedia PC as well as audio and data feeds, and transmits them across a telecommunications network, invariably ISDN. Most systems also provide document collaboration software and document interchange facilities. Estimates by industry analysts Personal Technology Research put the value of the PC-based videoconferencing market at $54.1 million in 1995, with 31,750 units shipped over the year, but that is split between hundreds of suppliers offering everything from individual cameras and boards to PCs and workstations pre-configured for video, such as the Silicon Graphics Indy Unix workstation. Basic point-to-point communications systems like Intel’s Proshare, which costs around $2,000 for a camera, video card, ISDN card and headset with microphone, offers one-to-one PC-based video and data conferencing across an ISDN line. Its accompanying software offers application and data sharing and there is also a 24-way multi-point ProShare server. Picture windows are quarter screen, and run at 15 frames a second, fed by a bandwidth of 128Kbps that is shared by sound, data and voice. As with many low-end systems, priority is given to sound and data transmission, so smooth video flow is sacrificed when sound and data need to hog bandwidth. Other similar products, such as Panasonic’s KXC range, offer transmission speeds of up to 384Kbps. More sophisticated systems which fuse multi-vendor and multi-point PC video stations, require some kind of video switch and LAN/WAN gateway, such as Video- Server’s MCS, which can link between 8 and 48 users from dispersed sites. The desktop may be where the publicity shots and the greatest long-term potential for videoconferencing lies, but it is not where the money is today – or even for the next few years. Take-up of videoconferencing has been strongest for so-called room or group videoconferencing systems. These are either set up permanently – typically in a company boardroom – or a trolley is used to ‘roll around’ the kit consisting of a camera, multi-directional microphone, loudspeakers, display screen, a wireless keypad, ‘Look-at-Me Buttons’ for camera tracking, and at its center, a powerful computer and codec. A vanilla system of this type costs on average $36,000, according to research by Telespan, which reckons 20,000 such units shipped in 1995. Until the technology is cheaper, faster, and works better on PCs, this is where the best systems will reside. With a large boardroom system you can have groups of people around the world conferring at 30 frames a second with perfect speech lip-sync, drawing on on-screen whiteboards, sharing schematics and sharing documents, says Stephen Crummey, vice president of worldwide sales at PictureTel. As shown by the rise of market leader PictureTel Corp, demand for such systems is growing fast, reaching around $700 million last year. With revenues of $346.8 million in 1995, up 36% from 1994, PictureTel can claim 51% of the market for group or room systems, according to Personal Technology Research, largely through sales of its flagship room products Concorde 4500, System 4000 and Venue 2000. A few other roll-about videoconferencing vendors also stand out – VTel, Compression Labs Inc, Sony and GPT Communication Systems. As an endorsement of how difficult it is to get the technology right at this stage, several major players – most notably IBM and British Telecom – have given up competing with PictureTel’s group videoconferencing kit and have become system distributors for the company. MCI and Zenith Data have also joined the 350 distributors that bring in two thirds of PictureTel’s revenues.

Cost not an issue

Such systems are typically used by the upper echelons of corporate management and only for a few hours a day. Thus the primary application is business meetings, and cost is not often much of an issue. In contrast, the business case for spreading videoconferencing to the corporate desktop, to enhance collaborative working or inter-business communications, is less easy to make. The cost of building the necessary networking and hardware into office PC LANs is still prohibitive, say some analysts, and will mean that desktop conferencing will not become part of the mainstream corporate computing infrastructure until the far side of the year 2000. Forrester Research puts the case against LAN-based videoconferencing most vigorously: Lack of broad business applications, insufficient bandwidth, and hidden infrastructure costs will prevent desktop videoconferencing from getting to first base, says Tom Pincince, an analyst with Forrester’s network strategy service. Although Intel and others are pushing down the price of desktop conferencing, the outlay required for a video-ready infrastructure is still prohibitive. According to Forrester, the total cost at the desktop is over $6,000 a seat, with on-board PC video hardware and software making up less than a quarter of the total cost. The rest of the bill is a combination of network upgrade costs and the price of a PC powerful enough to handle video’s bandwidth and computational demands.

Video market flop

Forrester found that three quarters of its Fortune 1,000 clients were against putting videoconferencing on the corporate LAN. MIS is worried about traditional issues like bandwidth and cost, but overwhelmingly, companies just don’t have any business demand for the technology. To put it bluntly, MIS does not consider a video call to a co-worker down the hall compelling. For firms with dispersed locations, buyers say their existing videoconferencing systems can do the job, says Pincince. That leaves desktop videoconferencing high and dry, languishing until the end of the century, waiting for the [enabling networking] foundation to be built around it for other reasons, says Pincince. The other video market flop has been videophones. Poor video quality and sound synchronization in the initial wave of products released in the early 1990s is only just being improved. Customers can now buy ISDN videophones, but unit sales of all videophones is measured in the tens of thousands rather than millions. Market pioneers in this area include Panasonic, AT&T and British Telecom, but with prices in the $3,000 to $4,000 range for an ISDN unit, market penetration will continue to be slow. The only sweet spot for the market for several years will be room systems. Until the year 2000 the big three vendors – CLI, PictureTel and VTel – will continue to reap the lion’s share of their revenues from high-end room systems, while their desktop offerings wither on the vine, says Pincince. He is even less hopeful for companies that promise a software-only solution. Start-ups like Vivo and InSoft will face the double whammy of no demand and strong customer allegiance to room systems.

New approaches

Such comments highlight the efforts by various specialists to do away with the video board, either by providing the functionality of the codec in software, or by trying to grease the path to mass-market video by embedding video right down on the main processor, as Intel and Sun have done. But Stephen Crummey at PictureTel, sees a gulf between desktop and room videoconferencing that these efforts will fail to breach until the next decade. There are going to be classes of systems, says Crummey. There will be a mass market, eventually, which says ‘maybe put it in the chip or do it in software only and let the average person who does not require the quality, a hobbyist maybe, use it.’ And then there is going to be the professional market. That’s where we are, that’s where the money is and that’s where you are going to need the hardware board. Nevertheless, PictureTel has been playing in that desktop market for almost two years with its lower-cost Live PCS products, and 25% of its revenues come from that PC area. From a dollar standpoint, we are doing more than anyone else in the market, including Intel, Crummey claims. But desktop videoconferencing is simply not the same ball game. There is a big issue with establishing the infrastructure for desktop videoconferencing. Most PCs are already hooked up to LANs and not everybody is going to have ISDN to the desktop. What you can do on a desktop is have a couple of people meeting. You can’t get 30 frames a second, you can’t do applications like remote diagnostics or telemedicine. But can Intel – and Sun with its Sparc chip – get the equivalent of that video board down onto the base chip? We don’t believe it. Having standard MPEG compression on a chip is one thing. Having realtime 30 frames a second video flowing through that chip, no way. You are not going to get good video without using some form of hardware [board], I don’t care what anyone says, Intel included, attests Crummey. Look at Microsoft’s Video for Windows, look at Apple’s QuickTime for Mac. They’re a joke. That’s not video, that’s a toy, that’s a postage stamp. That leaves the room systems in the driving seat. 70% of the market value is today derived from room systems. Even if over the next two years that falls to 60%, desktop’s market share will still only amount to between 20% and 25% with the remaining 15% to 20% split between services and networking. The realistic observer would still note that videoconferencing is a communications medium not yet of the many, still of the few, says Adrain Butcher, general manager of British Telecom’s Visual Solutions division. No-one doubts there will be a mass market for videoconferencing, but there is great debate about the form it will take and when it will arrive. Over time there will be, and the way that it is going to be done is by a $200 software codec you buy and plug into the PC, but it still doesn’t solve the ISDN issues, the networking issues, says Crummey. It’s not like you’re buying Excel or Word, but it will happen. It’s not clear to me [how] you make money in that market. The mass market is equally likely to be set in motion by another trigger, the Internet. Already, software companies are offering phone calls and video clips – albeit low-quality – across the worldwide network at a fraction of the cost of using telephone lines. Although the problems of creating a quantum leap in quality are formidable, few experts doubt that in the early decades of the next century, videoconferencing will be Web-based. A few companies, like White Pine Software with its CU-SeeMe technology, claim they can provide videoconferencing across the Internet. In the meantime, expensive fast lines will be the standard, according to Butcher at BT Visual Solutions. ISDN is the primary network for the foreseeable future for this technology. The slower than expected take-off of desktop videoconferencing has left the market highly-fragmented, as many companies experiment with price points and package combinations in search of customers or applications that have appeal. Certainly, vendors have come to realize that there is no horizontal market for desktop videoconferencing. Most are chasing vertical sectors such as education, medical, finance and retail. Schools in remote areas are using videoconferencing to deliver better education, designers are showing early versions of their products to clients, engineers are collaborating on projects, judges are considering using videoconferencing to interview witnesses, farmers are viewing livestock, and so on. But most demand comes from the business meeting market. In the meantime, nobody seems certain of when the long-heralded surge in videoconferencing will come. Analysts such as Forward Concepts, which said the market was worth $866 million in 1995, believe a $3.6 billion market will emerge by 2000. Personal Technology Research reckons the market for PC-based systems will be worth $1.5 million units in 1999, when 3.8 million units will ship.

Latent demand

The turning point, according to bearish analysts, will be when videoconferencing is installable for $200 per node – and is simply a mix of camera and software. More realistically, analysts at research houses Ovum and Insight say their feedback from customers suggests the mass market threshold is closer to $1,500. What is clear is that there is huge latent demand.