Abstract

Sandy Kyrish. From Videotex to the Internet: Lessons from Online Services 1981-1996.

The rapid evolution of the Internet has been matched by a torrent of news and prediction about it, making it difficult to keep the medium in perspective. One way to envisage the Internet, and to speculate on its future development, is to think of it as part of a continuing history of online services rather than as a sudden and singular phenomenon.

This study analyzes predictions and outcomes between 1981 and 1996 for US residential online services, beginning with videotex and ending with Internet and World Wide Web. It looks for lessons in the similarities and differences between past and present expectations for, and iterations of, online services.

In this summary, the term ‘online’ refers generally to services that are delivered over a narrowband connection. The study distinguishes three intervals of narrowband services, called (1) videotex, (2) online services, and (3) the Internet.

The study first reports on a strong build-up of predictions about videotex in the early and mid-1980s. Videotex enabled customers to use a keypad to call up screens of content from a centralized database, for display on a home viewing screen. It was seen as a natural result of combining the television, the telephone, and computing power (the modern notion of convergence is only a new buzzword, not a new concept). This hybrid technology frequently evoked notions of social and technological ‘revolution’. Predictions from this interval were often optimistic and technology-driven, though there was also significant pessimism in the business press.

Predictors presented a characteristic review of possibilities. Typically, these included references to shopping, banking, news, travel reservations, and messaging; they often suggested that videotex would ‘change the way’ we play, learn and work. Many predictions emanated from vendors offering or trialling such services.

Market researchers projected billions of dollars of revenue from equipment sales and service usage. But by March 1986, three U.S. videotex providers had ceased operation after consumer disinterest led to combined losses of over one hundred million dollars.

The study then reports predictions made about online services in the late 1980s to early 1990s. (At that time, the term ‘videotex’ largely disappeared as a database subject keyword and was replaced by ‘online’.) Following the videotex failures, the boilerplate of many articles was to recast past predictions in a negative light or to recite lost investment dollars. Often the same writers or publications that had promoted videotex now took a skeptical view of the technology.

Although the launch of Prodigy in 1988 led to another round of optimistic predictions, overall the modest performance of online services dampened talk about their social impact in this interval. By 1993, three national online services claimed about three million subscribers, though Prodigy had already suffered cumulative losses of over one billion dollars.

Interestingly, as the luster of online services faded, similar-sounding technology predictions appeared concerning new possibilities for broadband applications. In the early 1990s, discussion of a residential broadband future included expectations for ‘revolutionary’ household change and enumerated many of the same services predicted for videotex. But by 1993, delayed or failed broadband trials slowed these predictions as well, particularly when the Internet and the newly-created World Wide Web came into prominence.

The study then reports predictions made about the Internet and World Wide Web, from the early 1990s to the present. By 1995, computer users across the world had created more than three million Web pages of information, entertainment, and advertising; there are more than twenty million Web pages today. The business press reacted rapidly to the phenomenon with reporting, analysis, and prediction. More than 1,700 business-press stories were referenced about the Internet in 1995 alone, compared to a peak of about 120 articles about videotex in 1984.

Strongly optimistic predictions have come from a number of well-known individuals. The unifying theme is that networked computers will provide instant, on-demand access to unlimited information and communications possibilities, fundamentally changing personal habits, and therefore society, for the positive. However there is also significant skepticism, not only about the views of predictors such as Bill Gates, but about the desirability of a fully networked world.

Many of today’s business-press predictions for the Internet seem to be a fusion of the two predictive intervals for videotex. Journalists seem unwilling to challenge the technological phenomenon of the Internet yet wish to sound distanced from the hype.

The study offers conclusions from this ‘predictive history’ of narrowband services. It considers two aspects: lessons from the architectural evolution of online services, and lessons from the predictions.

In this study, the three intervals of narrowband service development were based on historical demarcations, but they also demonstrate a pattern of progressive decentralization of the technology and content components. Videotex was a top-down model of centralized providers furnishing centralized services through a specialized terminal. The design reflected the technology of the time, but it also then emphasized centralized interactive services – ie transaction and information oriented. These services were assumed to be desirable, but consumers were not interested; predictors (and suppliers) confused technological capability with market demand. This suggests parallels to recent broadband trials and their problems, since the model for interactive TV has been identical to that of videotex; namely, centralized providers (ie telephone companies) offering centralized programming services (ie video on demand) through specialized set-top boxes.

In the late 1980s, Prodigy was also a centralized provider offering mostly centralized interactive services. But a key component changed with the use of consumers’ own standards-based computers. The change in delivery device – from passive terminal to a multi-purpose machine – changed the assumed model for services, though Prodigy’s marketers did not seem to take this fully into account. The PC as terminal now meant that online services such as encyclopedias and games were often more effective when run directly as a software package. More important, consumers followed the pattern of text-based online services such as CompuServe and showed greater interest in communicative services such as electronic mail and bulletin boards – something Prodigy was architecturally and strategically unprepared for.

Today, the Internet is an antithesis of the first model. It features decentralized content that can be created by anyone and accessed on a standardized platform. This is a new model of services, both in architecture and usage. If the original assumption was that householders would consume online services in their homes to be save time, present indicators are that householders are more interested in using online services to produce or peruse content (Web pages, e-mail, newsgroups) as a new, enjoyable way to spend time.

The still-unsolved problem is that while service providers are stumbling over each other to win new residential subscribers, there is no evidence to show that reaching out beyond the core of already-converted computer users will be profitable.

Several lessons can be taken from the predictions themselves. First, although generalized predictions of sweeping social change are magnetic, they rarely seem to come true.

From a business perspective, these predictions are difficult to empirically judge; they are also the most difficult to translate into concrete products and services to sell for profit.

And predictions of ‘certain’ change may encourage technology spending that is more defensive than strategic – an acknowledged fact in videotex spending, though none of the investors yielded benefits from this. This suggests that companies should not invest significant capital in a large-scale venture that is based on amorphous concerns of being left behind. Companies might set up field trials or joint ventures whereby they can learn, observe, and be prepared to move when real market demand materializes – rather than creating a monster in order to tame it.

Second, general predictions of social change usually rely on an equally general belief in individual change. In other words, transformed societies require transformed individuals. But this vision of continuous self-improvement has not generally been borne out in other uses of media and technology, much less in past videotex and online service offerings and in today’s recreational Web-surfing. Perhaps it is time to quit treating consumer disinterest in long-predicted services as a problem in implementation, and to see it as evidence that people may be more interested in doing new things with a new medium.

Indeed, although the architecture of online services is now fully changed from the fifteen-year-old model of videotex, it is almost distressing to note how similar the predictions are for its use. This study argues that ‘visionary’ thinking about technology is unimaginative when it simply repeats traditionally futuristic views of a changed society, especially since these visions have not been a useful road-map in the past.

Online services have found their success through decentralization – content is now more in the hands of users than providers, and communicative services such as electronic mail and newsgroups are highly popular. Given this evolved architecture and evidence of its power, it is time to move past the ‘radio with pictures’ thinking inherent in many predictions for the Internet. Then we can develop new and better models to assess future possibilities for residential demand for networked services.