The following text is copyright 1997 by Network World, permission is hearby given for reproduction, as long as attribution is given and this notice is included.

Prognostication as a business.

We put up with a lot of guessing in this business of technology. Much of the guessing is wildly, some times laughably, wrong. But occasionally someone gets it right; Moore's law has been one of those cases. Gordon Moore's 1965 prediction that transistor density of semiconductor chips would double roughly every 18 months has proven remarkably accurate. It is this type of prediction that makes us so gullible for the predictions we get on a daily basis from all corners of the technology consulting world. Take for example, the predictions over the last five years about the extent of ISDN deployment to see how bad things can get. At first glance it would appear that the records of the major consulting firms are rarely blemished by accuracy. Why, in the face of this, do these consulting companies continue to do so well financially?

Once upon a time some technology prediction was easy and produced reasonably exact results. For example, for decades telephone companies were able to predict the growth in the demand for telephone numbers by looking at birth rates and building permits. This has fallen apart. The development of new applications has invalidated the old assumption that there would be a need for one telephone and one telephone number for each housing unit and employee. Cellular phones, fax machines and the Internet have changed the picture to a point where the predictions of future trends that accompany the ever more common requests for new area codes hardly outlive the transition to the new codes.

New applications can limit the reliability of predictions but the most egregious miss-predictions do not seem to fit this model. In 1995 one company in the technology forecasting business estimated that there would be 4.5 million ISDN lines installed by the turn of the century, this year the same company predicts that the number will be about 3 million. Reduced expectations in spite of the migration of usage of ISDN from being a fancy telephone replacement to being an Internet data pipe which should have increased usage. The cause of the inaccuracy in this particular case is easy to find, the predictions did not adequately take into account the perfection of incompetence found in the pricing departments of telephone companies.

A common thread in many technology predictions is wildly optimistic predictions about the adoption of new technology. For the last five years most of the predictions had the ATM market approaching $10 billion per year within five years. These predictions have been a constant, only the target year has changed, receding into the future at about the same rate as time. There are many reasons that the predictions have been wrong but the fact remains that accurate predictions of the acceptance of ATM technology have been few indeed.

There seems to be some deep need to for humans to know their future that flies in the face of experience. I predict that, despite the fact that the only thing that one can accuratally predict about the professional prognostication is that the result will be inaccurate, that this will remain a profitable business.

disclaimer: The best predicting at Harvard is done in the history department but the above is my own.