Consumer Research and Demand Forecasting For Wideband Telecommunications Services: Some Perspectives

Murlidhar Rao, GTE Laboratories
Gregory E. Wester, Data Resources, Inc.
ABSTRACT - In the near future, consumers will be exposed to new telecommunications services beyond simple dialtone and long distance calling. With advances in distribution techniques, consumers' homes will be "wired" in such a way as to send and receive more than just voice or low-rate data over modems. As fiber optics pervade the residential arena, households will have access to new services such as video-on-demand, videophone and wideband videotex.
[ to cite ]:
Murlidhar Rao and Gregory E. Wester (1989) ,"Consumer Research and Demand Forecasting For Wideband Telecommunications Services: Some Perspectives", in NA - Advances in Consumer Research Volume 16, eds. Thomas K. Srull, Provo, UT : Association for Consumer Research, Pages: 619-628.

Advances in Consumer Research Volume 16, 1989      Pages 619-628

CONSUMER RESEARCH AND DEMAND FORECASTING FOR WIDEBAND TELECOMMUNICATIONS SERVICES: SOME PERSPECTIVES

Murlidhar Rao, GTE Laboratories

Gregory E. Wester, Data Resources, Inc.

ABSTRACT -

In the near future, consumers will be exposed to new telecommunications services beyond simple dialtone and long distance calling. With advances in distribution techniques, consumers' homes will be "wired" in such a way as to send and receive more than just voice or low-rate data over modems. As fiber optics pervade the residential arena, households will have access to new services such as video-on-demand, videophone and wideband videotex.

The demand for these new services, known as wideband services because of the wide or large amount of information and richness, is tricky to forecast. The difficulty arises because of consumers' inability to react adequately to concepts and services that are in a domain in which they have no hands-on experience, These services are technologically driven and are "truly new and unique".

This inability to assess demand accurately ultimately can affect the design. delivery and marketing of new wideband services. For example, technological solutions that entail lower fixed costs may be implemented even though they result in greater total cost at levels of demand greater than those indicated by initial estimates. Another major technological and marketing problem is that any one new wideband service will not justify the capital expenditures for the necessary network upgrades while there are no existing (non-wideband) offerings that necessitate these upgrades.

This paper examines previous field trials and wideband service offerings, and then addresses the relationship between technology and marketing in the realm of new services. Potential problems are identified and solutions are suggested. Specifically, the use of problem quantification (rather than benefit evaluations) is recommended as a means to uncover areas of consumer interest. Since consumers have difficulty articulating the desired benefits since they are unfamiliar with them, problems in broadly defined areas are explored and quantified. These problem-areas feed a series of "ideation" groups comprised of specialists who use their technological sophistication to flesh out the attributes of potential new wideband services. Tine resulting concepts are again screened by consumers for interest and willingness-to-pay. The use of stand-alone prototypes to test consumer reaction to technological implementations is also discussed.

1. OUTLINE

In this paper, we will first evaluate several of the more important field trials of wideband residential services nd provide a marketing critique of these trials. We will then describe a process for designing new telecommunications services, and indicate how a process such as this one would add substantial value to the trials by employing a multi-step strategy incorporating marketing research, systems engineering and rapid prototyping. At each step of this process, we attempt to assess what customers want and how much they might pay for it. We also measure how much they pay for existing solutions in an effort to anchor their stated willingness-to-pay in more concrete spending levels. Field trials, in this process, become marketing tests of hypothesized trial and repeat, leading to more reliable estimates of demand at stated price levels.

Our orientation throughout this paper is that of consumer researchers and marketers. Engineering terminology is kept to a minimum, and is used only if necessary.

2. DEFINITIONS

The usual services we obtain today over our telephone lines fall into the domain of "narrowband" services, i.e., they do not require high transmission rates to move from source to destination. The reasons they do not require high transmission rates are (1) the amount of information contained in each transaction is low, and (2) the number of transactions per unit of time is small. Examples of narrowband services include standard voice telephone calls, modem connections to computers and credit card verification - found in most department stores.

By contrast, "wideband" services require much higher transmission rates. There does not appear to bc industry consensus as to what rate is required to support these wideband services, however, a data rate of 140 million bits of information per second (140 Mb/sec.) is commonly quoted as sufficient for transmitting high-quality video. For high-quality digital audio, a data rate of 2 Mb/sec. is sufficient. In comparison, narrowband services require data rates of 64Kb/sec. or less.

What are these wideband services that will require these enormous transmission speeds? A useful typology is found in Rothamel (1986) and is displayed as Exhibit 1. Note that the services that require the highest data rates are all video-based. All document and audio services can be transmitted at the lower rate.

At this point, it may be useful to define a few terms. Exhibit 1 lists four types of Service Classes: Dialog, Messaging, Retrieval and Distribution. Information must flow between a source and a destination, typically the user, and the breakdown in Service Classes is based upon how the information flows.

In Dialog services, the source and destination interact continuously via the same mode. For example, in video telephony, both source and destination continuously send video images to each other.

EXHIBIT 1

SOME WIDEBAND SERVICES AND THEIR DATA RATES

In Messaging services, the source sends information; such as a video image for picture mail, which is saved and then retrieved at a later date by the destination.

In Retrieval services, the user interactively controls the information content that is sent by the source. The request for information may be of one mode and the information requested may be of another. For example, in Image Retrieval, the request could be sent by pushing buttons and the information retrieved could be a page of the Sears Roebuck catalog displayed on a video monitor. Another example of a Retrieval service is Video-on Demand, in which a user may request a specific program or-movie that (s)he wants to view and also specify when the show is to be viewed.

The last group of services, Distribution services, gives the destination selection of, but not control over, what is sent by the source.

There are several unique features in many of these wideband services. First, they are all outside of the consumer's experience. Second, there is no directly competing product or service with which consumers can compare the potential new services. Third, consumers' price expectations are likely to be inaccurate (because they have no idea of the underlying investment or operating costs). Fourth, adoption of any one of these services is likely to be influenced by the mix of services offered broader penetration will result from offering many different services that appeal to different market segments with different needs.

The net result of these features is that consumer research becomes a tricky exercise, and demand forecasting even more so because the services offered arc inter-related.

Ho v, then, does one conduct useful marketing research and generate meaningful estimates of consumer demand? For inspiration, we first turn to some of the laboratory and field trials that have been conducted to date. These trials were not designed as marketing tests, but rather as technology tests. However, we intend to point out that there are substantial weaknesses in technology tests devoid of marketing concerns.

EXHIBIT 2

SUMMARY OF LABORATORY EXPERIMENTS & SERVICES OFFERED

3. LABORATORY EXPERIMENTS

Laboratory experiments have usually been conducted on very small numbers of homes/users (hereafter referred to as subscribers), usually under 10. These subscribers have been given access to wideband services in a laboratory setting, rather than at home.

The very limited number of subscribers make these laboratory experiments interesting only for exploring the possibilities of wideband technology. While it is possible to undertake consumer research for services such as TV reception over phone lines and high-definition TV, it is unclear whether any research was actually undertaken. With a service such as video telephony, the small number of subscribers limits consumer research to tests of usability of the service. Clearly, demand forecasting is impossible without a fairly large number of subscribers hooked up with video telephones.

4. FIELD TRIALS

There have been several medium-to-large scale field trials to date, of which we describe and evaluate four. The trials arc distinguished by the size of the subscriber base and the range of services offered

- Biarritz. France

- BIGFON in 7 cities, F.R.G.

- Tokyo and Mikata, Japan

- Westminster, U.K.

4.1 Biarritz, France

This field trial began in Biarritz in 1984, with 1,500 subscribers (1,200 residential and 300 business). Subscribers were given "access to a wide range of broadband services -- videophony, audiovisual databases, TV and stereo sound program distribution, and an on-line TV program library -- in addition to conventional narrow band services like telephony and videotex" (Touyarot, 1987).

This trial is unusual in one respect: it is the only field trial in which no free services were offered. Despite this strength, there are several problems with this trial.

First, the French Minitel experience has probably predisposed the French public into accepting both the concept of videotex-like services and the additional "box" in the home. The Minitel terminals have been given away free in the past to educate the public and to avoid the problems of having consumers pay for their own terminals.

Second, these services were offered to subscribers at very low prices, almost equivalent to regular telephone service: one additional telephone call unit was charged for each 10 minute interval after the first 10 minutes (Touyarot, et. al, c 1985)

Third, the French billing system provides only an "all-in-one" figure produced bimonthly. This aggregation, in contrast to the U.S.'s itemized billing, makes it difficult to interpret the cost consequences of using the "higher priced" wideband services. For example, the published data state that the average bimonthly phone bill rose by 30% and then stabilized. Moreover, the French (Touyarot, 1987) claim that: 1) as of March, 1986, 64% of the calls between any pair of videophone terminals were "always" conducted with videophony; 2) an additional 16% were "in many cases" made with videophony; and 3) that "within the first months following connection, some 40% of subscribers use [videophones] for over two-thirds of their calls, and this figure is increasing".

Attempting a rational integration of these usage patterns with the stated increase in the bimonthly bill becomes impossible due to the "all-in-one" characteristic of the French billing system. How could so many calls use the videophone and yet the average phone bill increase by only 30%? Could it be that 30% of the videophone calls last more than 10 minutes but less than 20 and thus accrue an additional telephone unit, or do 15% of the videophone calls last between 20 and 30 minutes? We can't tell.

To gain some perspective on the cost of delivering videophony relative to telephony, we use the cost of video conferencing relative to that of audio conferencing (Carey and Moss, 1985). The relative cost of video-conferencing, at commercial rates in the U.S., exceeds 50 times the cost of audio-conferencing. Granted that the cost basis of U.S. rates may be much higher for video than audio, and that conferencing costs may be higher than simple two-way phone service, the Carey and Moss calculations suggest that the prices charged subscribers in the French trial were subsidized to such an extent as to mace dubious any estimates of demand for the wideband service.

Informed estimates at GTE suggest that if pricing is based upon the bandwidth requirements of video versus audio telephony, the price of videophony is more likely to be some 100-250 times that of plain old telephone service.

Whichever estimate is used, it is clear that the Biarritz prices charged were far too low to bear ny resemblance to the actual prices that might be charged by a for-profit U S. telecommunications firm.

The fourth and final marketing weakness in the Biarritz trial pertains to an apparent lack of concern with the marketing mix. Without knowledge of the marketing mix employed, it is very difficult to judge what demand estimates might mean in the real world.

4.2 BIGFON, West Germany

This field trial began in 1984 and connected 320 subscribers in 7 large German cities (Kneisel, et.al., 1986) with optical fiber. Of these, 68 subscribers were equipped with videophones at no charge. The test was clearly labeled as a test of wideband technology and not of customer demand, which was assumed to exist The overall number of subscribers who can be connected to the videoconference network, i.e., a maximum of about 300, is rather modest compared to the immense investments required for the basic infrastructure. However, it is assumed that there is a latent demand for additional connections in trade and business centres" (Kneisel, et. al., 1986).

While the BIGFON trial appears to have been a good design for testing technology, it failed as a piece of test marketing. For all the reasons given above regarding the weakness of the Biarritz trial, and because the videophone services were free, BIGFON remained another wonderful wasted opportunity to conduct careful marketing experiments.

Especially noteworthy is the fact that investments to the tune of DM 2 billion have been made or planned for the 1984-1988 period on the basis of testing free services to 68 subscribers. This network is anticipated to connect a maximum of about 300 subscribers, making the investment cost (assuming that 300 subscribers actually sign up and pay real money for the service) a staggering DM 6.7 million per subscriber.

4.3 INS, Tokyo and Mikata, Japan

This trial began in 1984 (Takeda, et. al., 1986), and provided a total of 310 "service monitors", both business and residential, access to a variety of wideband services: videoconferencing, videophony, television, video monitoring, high-speed fax, and interactive video services (VRS, or Video Response System). The actual number of subscribers is unclear, since the 310 service monitors represent the total number of terminals needed for the different services offered. Also, 45 of the 310 monitors were located in the showrooms of NTT, the Japanese telephone company.

Early results reported below were derived from a total of 20 residential and 60 business "service monitors". In discussing videophony and their Video Response System, the authors state: "In spite of the fact that it is free of charge, only a few subscribers were using the service because for the most part they were strangers to each other. About 50% of all service monitors used VRS more than twice a week." (Takeda, et. al., 1986).

Again, as in the previous trials reported above, free service was provided, and questionable estimates were derived for the demand for wideband services. All criticisms of the BIGFON German trial and the Biarritz French trial are applicable to the INS Japanese trial.

4.4 Westminster, U.K.

The Westminster trial is probably the most extensive of its kind, serving a customer base of some 50,000 subscribers. The trial began commercial operation in 1985 in Westminster Borough, London.

The second distinguishing feature (other than its large subscriber base) of this trial is the fact that the cable network is switched in such a fashion as to allow for considerable expansion in the range of services offered. To date, TV channels, FM radio stations, an interactive videotex program guide, and a videolibrary have been implemented (Powter and Fox, 1986). The videolibrary was being tested in pilot form for payment-based service due to start in late 1986 (Kerr, 1986). Services listed as "awaiting commercial exploitation" in 1986 included a larger videotex service (at a cost to the subscriber), a gateway to other databases, and a pictorial videotex system (Powter and Pow, 1986). While a recent brochure (British Telecommunications plc, 1988) suggests that the pictorial videotex service had been tested and was commercially available, we have not uncovered any test results.

It appears that British Telecom is selling its technology as a package to cable television operators. In light of this "hard-doll r return", it is surprising that customer-research data have not been collected and widely disseminated. The only prior research we are aware of is the results of a market analysis conducted in rather perfunctory manner, in which the cost and availability to the customer of broadcast TV, pay-cable, movies and videotape rentals and purchases were calculated to get ballpark estimates of the competitive frame for videolibrary (Kerr, 1986). An analysis of traffic levels from viewership data for television is also reported in the same source.

We have no data that suggest that the Westminster trial was any more "marketing oriented" than the others reported on above. However, the British have simply put an infrastructure in place and are awaiting the development and commercialization of wideband services. Major investments in actually providing wideband sen ices such as videophony have not been made.

5. EVALUATION OF THE FOUR MAJOR FIELD TRIALS

The most noticeable fact about all these field trials is that they were all tests of technology, and not of customer demand for wideband services.

It is commonly believed that wideband networks will become financially justifiable only when they support extensive video-based entertainment to the home, thus allowing the recoupment of network cost. However, for such ubiquitous demand for video-based services to pervade the marketplace, sufficient consumer experience must be accumulated, which requires that the infrastructure to deliver the services be in place first.

This Catch-22 situation has generally been resolved in favor of first putting m the expensive network and switching capabilities, and then offering services that can be distributed over this network. The Germans have recognized this in stating that the "unconventional approach adopted by the Deutschc Bundespost in undertaking substantial investments without an accurate picture of the demand to be anticipated may at first seem rather unusual. However, from the Deutsche Bundespost's point of view, it is the only way out of the chicken and egg dilemma demand can also develop after investments have been made in a network since potential customers can very soon expect other customers to use the network, and secondly, because customer interest can be stimulated by demonstrating the facilities of the new service.' (Kneisel, et. al., 1986)

What does this mean in practice? While the Germans were willing to invest almost DM 7 million per user in the infrastructure. it is unlikely that private enterprise in the U.S. will do likewise. In fact, the most likely outcome is that trade-offs between network designs with different fixed and variable costs will be resolved in favor of solutions with lower fixed costs. Such technological decisions may, in the longer tenn. limit the number and quality of services that arc offered. Amortizing large investments over a smaller number of services and subscribers will also lead to higher total cost in the longer term.

Our contention is that wideband network decisions should be driven by customer need for wideband services, not d technology. We will describe a process that, from the marketer's standpoint, will enable the firm to make huge investments with reasonably accurate knowledge of customer demand, and with a base of careful consumer research underlying the estimates.

Other problems that have plagued previous trials are discussed below.

5.1 Absence of prior customer research

From the published literature, it is questionable whether the trials reported above were preceded by any "reasonable" consumer research. Can truly new services beyond the experience base of consumers be "researched" in any reasonable manna prior to the service concept being fully formulated? We think they can, and that the process described in this paper is a good example of how to do it.

5.2 Undefined marketing mix and free services

For all of us in this audience, these problems are glaring. For the telecommunications engineers typically in charge of these trials the problems are recognized, but not viewed as being critical. Recall that these individuals have backgrounds in regulated businesses where services are guaranteed a rate-of-return on investment. FigurIng out the cost of doing business is adequate to build a business plan, because the rate of return to the firm is mandated by law.

5.3 Underutilized research opportunity

These trials also have one other feature in common: they present remarkable opportunities for marketing experiments. The published work suggests that the research conducted was limited to a few focus groups and some survey work.

In summary, it is probably fair to say that as telecommunications companies emerge from the protective shroud of regulation, these "directly-to-field" approaches to assessing consumers' needs and forecasting demand will prove disastrous. More careful and thorough approaches will be needed to assess customers' needs and how much demand they will create at given price levels.

6. A PROPOSED PROCEDURE FOR NEW SERVICE INTRODUCTION

The ideas to be discussed in this section should be seen as the way marketers view the issues in telecommunications services research It represents the marketer's view of the mandate of the marketing concept, which is to propose and refine a process for designing new telecommunications services which will improve the quality of business decisions by focusing on consumer needs.

EXHIBIT 3

PHASE I: SELECT SERVICE AREA

Our process does this by both focusing on the needs of consumers and matching those needs with the capabilities and inclinations of the firm to market only those services that provide a profit to the firm.

The group that we feel necessary is a multidisciplinary one, with technologists, marketing researchers and human factors specialists working together to design promising new services. The technologists come in two flavors: those traditionally associated with telecommunications engineering (switching, transport, CPE and systems engineers), and those who specialize in producing the "look-and-feel" of real telecommunications services via creation of prototypes of these services.

The premise underlying the process we about to describe is that we need to filter "areas" in which to develop new service ideas through a coarse screen; then progressively use finer and finer screens to allow only promising service concepts in these areas through to the next steps. Thus, the earlier steps arc intended to be very broad and low-cost, while subsequent steps are narrower and narrower in scope and potentially higher in cost. Our thinking in this are draws upon many points of view (Cooper, 1984, Urban and Hauser, 1980 and others).

Our proposed process is divided into three major phases. In Phase I, we select the service area to be investigated further. In Phase II, we develop service concept(s) in the area(s) we selected in Phase I. In Phase III, we design and test the service(s) which were developed as concepts in Phase II.

We will exemplify our process using the area of "video entertainment services to the home" as our wideband application. There are several reasons for using this example. It is clear from the required data rates that the main justification for wideband networks is the transmission of video services. Also, all previous trials have provided some form of videobased services. Finally, experts in this area suggest that only video-based entertainment services to the home will generate enough usage to justify the huge capital costs of providing the infrastructure for wideband services.

We will point out how the proposed process makes a systematic attempt to build a business case for the service that will best satisfy customers' needs and provide the desired level of profitability to the firm

6.1 Phase I: Select Service Ares

The bulk of the work proposed in this Phase is in the areas of organizing people, collating and digesting secondary sources of data and providing a summary of the different efforts to a management team for a GO/NO GO decision. In the home-video case, the following outputs are considered necessary:

Customer Scan:

For major media (broadcast, basic cable, premium cable, pay-per-view, video rentals and purchases, and movies)

- ownership, viewing habits and trends

- household characteristics of users

Technology Scan

For major components of telecommunications technology (switching, network, consumer premise equipment, etc.)

- current state of technology, features, cost

- likely future costs

- competitive evaluations

Marketplace Scan

For major media (broadcast, basic cable, premium cable, pay-per-view, video rentals and purchases, and movies)

- revenue and trends

- major competitors

- cross-media competition

- competitive strengths and weaknesses

Phase I of this process follows the text-book approach to understanding whether the firm should enter a broad area, and whether the area has the requisite fit with corporate goals and the firm's resources to develop new services in this arena

6.2 Phase II: Develop Service Concept

This Phase is critical to understanding customers' unmet needs and developing technologically-feasible services that meet those needs. There are two major activity streams in this Phase. In the first, we propose development of a Core Service Concept, which is a statement of which unmet customer needs we will attack, what we know about the target customer, and who the major competitors (media and/or firms) are. The intent of the consumer research steps in this stream is to isolate existing consumer problems in the area under investigation. Rough estimates of how much consumers will be willing to pay for solutions are also obtained, to enable us to build the demand side of a business case.

If the Core Service Concept meets new business criteria, we proceed to the next activity stream, which yields a fully-formed Service Concept Statement The consumer research steps include the use of techniques such as conjoint measurement. Testing of the Service Concept Statement is done in the context of competitive offerings, with a price tag associated with the new service concept The Service Concept Statement may be likened to a semi-finished advertisement, for marketing research purposes. Even at this stage, rough estimates of market potential are made.

6.3 Phase III: Design and Test Service

We enter this Phase with a complete description of the service, with features, pricing, competitive positioning and source-of-business estimates. However, the service still does not have a user-interface, and consumers have assumed some characteristics of it. For example, an interactive in-home on-Demand video service with 5.000 shows to choose from may be considered by reasonable humans to be difficult to use, and their reactions to the concept may be based on assuming this level of difficulty. The key question is whether respondents will react to the functionality of the service or how easy or difficult to use they perceive it to be.

We like to say that the service entering Phase m is likely to be useful to consumers, but not necessarily usable.

In Phase m, we propose first to build a prototype of the service, which is a "look-and-feel" delivery mechanism. Customers can interact with and evaluate the service for both usefulness and usability. We then add other characterizational elements, such as a name and advertising position, and the positioned prototype is evaluated in extended-use testing. This is the stage where we propose to estimate whether repeat usage of the service will measure up to our expectations and whether the business case is a reasonable one. At this stage, we also assess whether there is any fall-off in usage over time after the novelty of the new service has worn off.

Expected outputs of these steps are more precise estimates of trial, and the first moderately robust estimates of repeat With these data available, service costs and marketing costs are factored into a business plan for the "Test Market Simulation". Note that, until this stage, no investments have been incurred for switches, network installation or modification, equipment at the customer site (CPE) or marketing. In the simulated test market, we are unsure whether the infrastructure to provide the service(s) will be in place or whether we will "kluge" it to give customers the illusion of the "real thing".

We expect to get out of the simulated test market quite robust estimates of trial, repeat usage, customer response to pricing variations, and the usual consumer research measures (awareness, usage, chum, likes, dislikes, problems, etc.). If the numbers hold up, and the new service is likely to meet the firm's requirements for profitability and quality, we then suggest taking the new service into Test Market or roll-out. Since the infrastructure costs of wideband services is quite high, it is highly likely that the rollout has the look and feel of sequential test markets. This characteristic of local markets makes it possible to fine-tune the marketing mix on a local-area basis.

EXHIBIT 4

PHASE II: DEVELOP SERVICE CONCEPT

EXHIBIT 5

PHASE III: DESIGN & TEST SERVICE

7. SUMMARY AND CONCLUSIONS

We have described the usual procedures employed in tests of wideband services in the telecommunications industry. Our biggest criticisms of these are that (1) they appear to be based on little, if any, prior customer research; (2) they make no effort to recognize that pricing is critical to both the level and time-path of customer acceptance; (3) their field results are dependent on the unspecified marketing mix employed in introducing the service, and most importantly (4) that a firm should not make major investment decisions without adequate justification. While the philosophy of technology-driven service deployment was fine in regulated businesses, it is inappropriate in an unregulated environment.

The process we have described is an early prototype of one we plan to test and refine at GTE Laboratories. We have a long way to go before we can state which research approaches are appropriate to what business and market conditions. However, we do have in place a process from which we expect to ascertain certain basic facts about the market at certain stages. These stages are hypothesized to be the most appropriate for obtaining that information, given the amount of time, money and service specificity available. We feel that this process overcomes the four problems that plague the way telecommunications firms have traditionally determined customers' needs and estimated demand.

We will be testing the need for each of these steps, the specific tools and techniques appropriate at each step, and whether or not we can make valid and reliable estimates of the measures we need to make the investment up to the next step.

Stay tuned, we'll be updating you in the next several months.

BIBLIOGRAPHY

British Telecommunications, "Advanced Switched Star Cable TV System", brochure, 1988.

Carey, John and Mitchell L. Moss, She diffusion of new telecommunication technologies", Telecommunications Policy, June 1985

Cooper, R.G., "A Process Model for Industrial New Product Development", IEEE Transactions Engineering Management EM-30 (Feb. 1983), Vol 2-11, 1984.

Kerr, G.W., "On-demand interactive video services on the British Telecom switched-star cable TV network", British Telecom Technology Journal, Vol 4, No. 4, October 1986.

Kneisel, K.-E., W. Gerfen and K Hofig, She optical fibre overlay network of the Deutsche Bundespost an approach from BIGFON towards broadband ISDN", SPIE Vol. 630 Fibre Optics '86 (Sira), 1986.

Powter, J., and J. FOX, 'The Westminster cable TV system", Electronics and Power, December 1986.

Rothamel, Hans Jorg, "ISDN Broadband Services and Applications", telecom report 9 (1986) No. 1, 1986.

Takeda, Hideo, Kentaro Tokikuni and Reiji Furuya, "Broadband Telecommunications Services in the INS Model System", Proceedings IEEE 1986, pp. 506-510, 1986.

Touyarot, Philippe, "First Results from Life Operation of the Biarritz Broadband Multiservice Network", Proceedings IEEE 1987, pp. 229-233, 1987.

Touyarot, P., B. Marc and A. de Panafieu, "First Lessons from the Biarritz Trial Network", SPIE Vol. 585 Fiber Optic Broadband Networks 1985, 1985.

Urban G.L., and J.R. Hauser, Design and Marketing of New Products. Prentice-Hall, New York, 1980.

----------------------------------------