On Managing Information Technology -
At one time it was fashionable in management circles to hold that the skill set required for the successful management of one business almost automatically ported over to successful management of another. Good managers of police departments would likely be just as effective managing a community college or hotel, it was said. Whether that view ever held true in the computing profession is in doubt. What is not in doubt is that it won't hold true for the 21st century. Those institutions which fail to take note of this will not endure.
The general reason behind my claim is that the management of information technology (or perhaps all advanced technology) requires a very different skill set than would be brought to bear on the management of different institutional resources. With the smokestack and extraction industries, the service industries, even agriculture, it could always be claimed that the industry consisted of a triumvirate of sorts: production, marketing and infrastructure. These entities could function more-or-less independently of one another. One didn't need to know organic chemistry to market coal or understand much of telephony to supervise operators.
Information technology just doesn't work this way. It probably never did, but geometric market growth conceals a multitude of management errors. Now that the growth of information technology has slowed from phenomenal to respectable, the "natural selection" provided by competition is thinning out the ranks. The absence of an executive info-czar in the 21st century will set technology-oriented institutions on a collision course with extinction. Failure to observe this truism almost sank IBM, the computer industry behemoth which could do no wrong for the first half of the computer revolution. It is costing major academic institutions dearly at the moment as well as enormous sums of monies taken from ever-shrinking budgets are wasted on technology bounty hunting expeditions which produce hardware and software with minimal generality and shelf life. This is due to a general and pervasive lack of understanding of and context for current technology. I can illustrate this culling of the herd with IBM's well- publicized loss of $10 billion in one year. This doesn't happen by lethargy alone. To produce the largest corporate loss in recorded history one has to cultivate incompetence as one would a penchant for art. This cultivation only culminated in this debacle - it was already reaching fruition by the early 1970's.
IBM didn't find itself in this predicament through the cunning of off-shore clone makers, or the Japanese theft of trade secrets, or the failure to license DOS, or the Justice Department's anti- trust litigation, although these things didn't help their cause any. They got into this predicament through their own mismanagement. In particular, their failure to understand the technology envelope they were in. It seems hard to believe.
Illustrations abound. IBM lead the way in multiprocessing, CMOS thin-film storage and multi-chip modules with their ambitious F/S project in 1971; their RISC architecture in the ACS and ServiceFree prototypes antedated Sun's effort by more than a decade; their Mermaid GUI shell for DOS lead Microsoft's Windows by five years; and this is just a short list. From a technological point of view, IBM continued to lead the industry through 1990. What happened?
The problem was that the genuine understanding of the technological frontier was limited to the technologists and not management. The reason that few have heard of the above initiatives is that they were killed or hampered by management who saw them as distracting technologies. That's right. IBM's lead in RISC, GUI, PC, distributed computing, 32-bit multiprocessors, and sundry other cutting edge technologies was wasted because the advantages were not fully understood by the major decision makers. And this was a corporation which had proprietary interest in these technologies because of its multi- billion dollar native R&D program!
The experiences of Control Data, Unisys, Wang, Apple, Digital and other corporate giants is not that dissimilar. They all lost their competitive edge because of poor technological forecasting, which was due to the technological ignorance of the executive teams involved. At bottom the problem is skill inversion where over time management becomes dominated by individuals who have little understanding of the technology developed under them. This weakness leads to unrealistic and damaging decisions which ... You see my point.
The information technology windows of opportunity are measured in months, not years, and this is what makes the skill inversion problem critical in this industry. One mistake can take you out of the game for good. There is little time to recover. Corporations and institutions which fail to take this into account will become extinct in short order. Enter the info-czar; the executive vice president for science and technology. This is a person whose skills are first and foremost technological, not managerial - a person who is genuinely in touch with technology and is widely recognized from within and without the organization for personal achievements in the area. This person will play a major role in technology policy and corporate strategy. Had IBM had a few of these, the aforementioned mistakes would have been avoided. Without such an individual, many of our current high-tech institutions will face a similar fate.
The management model which I have in mind is similar in structure to the centralized production system which many of the major motion picture production studios used during Hollywood's heyday. Perhaps the best example of this was MGM under Mayer and Thalberg, with the latter, as chief of production, providing the artistic and technological direction to the corporation. That system made MGM the industry leader. Coincidentally, when it was abandoned after Thalberg's death, MGM began its slow but steady decline. While the model doesn't withstand microscopic comparison, it illustrates how successful an executive team when it has "mastery of the domain" fully enfranchised. This mastery cannot be absorbed or delegated. One of the worst strategies is to try to maintain technological omniscience through information agents and brokers. Slightly less undesirable, though seldom more effective, is the use of internal R&D watch lists. Both techniques are virtually assured of producing what I have called technology bounty hunting as corporations race over the technology landscape for a closer view of the horizon which never materializes
This model is not limited to the executive ranks. It should also surface in the boardroom and, other things equal, should be sought after at other executive levels in the organization. The operative concept is that organizations which intend to maintain a presence in information technology must have genuine technological expertise represented in each major decision it is to remain viable. This simply cannot be achieved through token appointments.