accesses since August 25, 1999
copyright notice
link to published version: Communications of the ACM, November, 1999
accesses since August 25, 1999"There is no business in the world which can hope to move forward if it does not keep abreast of the time, look into the future and study the probable demands of the future."
The quote above is from Thomas J. Watson, Sr. on the occasion of the opening of a new IBM laboratory in 1932. Few would disagree of the need for sound and accurate technology forecasting in any organization which seeks to remain competitive. However there is little evidence that the full ramifications of this observation are consistently understood or widely implemented even today. T.J. senior’s advice may be falling on deaf ears – and this is will become even more of a problem as we begin a new millennium where network/distributed processing environments will be even more unforgiving of technological blunders.
Many years ago, a senior IBMer told me that Thomas J. Watson, Sr. founded the Yorktown Heights research center primarily as a means to avoid technological surprises, and only secondarily as a leading research center. In a sense, this same theme was behind the Manhattan Project where the original goal was to prove that it was impossible to make an atomic bomb – in this case, the "reductio ad absurdum" approach literally "backfired." However, in retrospect, there seems to be much to recommend the strategy of using research to prevent being blind-sided – either by competition or foe. In many situations, it is more important to know what we don’t know than to know what we know. I assume that Watson, Sr. and Leo Szilard were thinking along those same lines.
History has shown that many of the giants of our industry – including those as formidable as Xerox, Digital, NCR, Control Data, and Unisys seem to have failed to take this to heart. In fact, IBM itself is no foreigner to technology-apoplexy – Future Systems, micro-channel architecture, RISC, OS/2 - need I say more?
In this column I want to emphasize the importance of accurate technology forecasting and to attempt to give some explanation of why we have so few industry and academic leaders who are really good at it. The sidebar illustrates this point. (Note: this list came to me as unsolicited email some time ago so I can’t personally guarantee the accuracy of the quotes. But the point is made, even if the names are changed to protect the innocent.)
… is the inability to spot technology black holes - how many organizations are trying to extricate themselves from costly thin-ethernet connectivity in favor of 10baseT networks which use inexpensive copper and are far easier to trace and debug? How many failed ISDN sites will have to undergo retrofits to DSL systems, that capture the unused frequencies of the existing telephone copper infrastructure to achieve several times the bandwidth of ISDN. How many organizations jumped on the CP/M, Next and OS/2 bandwagons? The point is that these sorts of decisions call for informed leadership and not management by consensus. Trend analysis, poll taking, flash lists, and outside consultants will not provide accurate information quickly enough to be reliable in this information age. Organizations which have much involvement with high-technology (which is almost everyone nowadays) need the internal capability and adaptability which comes from such leadership. So it behooves us to determine the strategies that provide an organization with the greatest likelihood of success in understanding and controlling its technology future.
In The Mythical Man-Month Fred Brooks deplores wasteful practices of corporate software managers like throwing bodies at software projects to speed up completion. To quote Brooks, "I think it's important to have a system architect who's different from the boss. It's also just as important in the implementation of an architecture to have a chief designer who maintains personal intellectual mastery of the overall design." (cf. Sasha/Lazere, Copernicus, NY, p. 168).
Why not throw bodies at software projects? Why not put people in charge of technology projects who do not possess intellectual mastery of the overall design? Because it won't work! Technology can't be managed like inventories. The skills required are precise and, to a large degree, non-portable.
Over the years, I have argued for this realization before industry executives and academic leaders, alike. I have suggested that the information needs of modern organizations are so variegated and complex that a executive positions should be created just for science and technology oversight. In information-intensive industries and academia, the appropriate model might be to create an executive position for strategic planning of information technology. But I’ve been disappointed. While, I have noticed that the title of "info czar" has certainly caught on, the substance of my idea hasn't. Modern organizations that have created these positions routinely violate the spirit of my argument that this calls for a technologist. My concept was that the placement of skilled and successful technologists in the organizational power loop is essential to successful strategic planning. It was never my advice to create additional managerial overhead by throwing bodies at the problem. The central theme throughout this column is that the skills of "technology executives" are first and foremost technology-related.
My idea seems to be similar to Brooks’ - the success of the organization depends upon a great blend between job description and skill set. Consider the following question: What kind of skills would be required of a vice president for information technology? -- Knowledge of the capabilities of digital networks? Familiarity with data mining? Working knowledge of information agency? Understanding of distributed database systems? These are all likely to be critical skills for the info-czar of a progressive organization. However, from my experience, these are not the skills which are most widely sought after. I’ll illustrate the point by paraphrasing two recent job ads for executive positions in American universities:
POSITION#1: Vice Chancellor for Information Infrastructure
JOB DESCRIPTION: Executive responsibility for policy-making, planning, development, implementation and overall administration for computing and related technologies in support of the University's ...mission. The vice chancellor will: (1) create and maintain a productive, dynamic environment for the use of computing and related technologies..., (2) create and maintain an organizational climate and a working environment ... that encourages creativity, adaptability and cost-effectiveness ..., (3) have administrative responsibility for academic technology, computing and administrative systems, network and telephone services, (4) participate ... in policy-making, strategic planning, goal-setting and troubleshooting on institutional issues [regarding information technology], ... etc.
REQUIREMENTS: (1) minimum of 5 years of management in computer-related areas, (2) experience in directing and managing an open distributed-computing environment in a university, (3) proven record of success in planning and problem-solving and in managing complex information technology resources, (4) strong interpersonal working relationships with members of diverse constituencies, (5) PhD or equivalent.
POSITION#2: Associate Provost for Information Technologies
JOB DESCRIPTION: Executive position reporting to the Provost. Responsibility for envisioning and planning the effective use of information technologies. This person will be the institution's advocate for information technology. Position includes management of full-time staff of 109 with an annual budget of $8 million.
REQUIREMENTS: Successful applicant must have (1) excellent interpersonal and oral/written communication skills, (2) experience with, and commitment to, participatory management, (3) at least 5 years in planning and problem solving and managing a diverse information technologies staff, (4) knowledge and experience in information and telecommunications technology, at least a Master's Degree in an appropriate discipline. Desired qualifications include (5) earned doctorate in computer science/engineering, MIS or related field, (6) professional experience in a higher education setting, (7) experience with state and federal funding agencies, (8) proven record of obtaining external gifts and grants, (9) proven record of promoting the creative and innovative use of technology to support the teaching and learning process, (10) experience with the delivery of instance learning through technology, (11) experience in high performance computing, (12) experience with distributed client/server administrative systems, (13) knowledge and experience with multimedia technology, (14) record of academic achievement including publications and participation in professional conferences and organizations, (15) prior experience in management of a computing facility, (16) prior experience in management of a telecommunications facility.
Both of these institutions seemed to be forward looking, and they may well have succeeded in what they were trying to accomplish. But if they did so, it was despite their job ads rather than because of them, for the ads betray a fundamental misconception about the nature of the solution and the position which will bring it about.
Let's analyze these ads in turn. In position #1, there is a clear mismatch between what I have argued are the sine-qua-non skills for the position and those sought after in the ad. The problem, which is so typical of our time, is that the job ad betrays a fundamental confusion between the w ay that the university chooses to administer itself with the way it intends to address its future information technology needs. By separating the two, one has much better defined set of objectives and a greater likelihood of achieving them. Failure to separate the two, is most likely to produce confused and conflicting organizational objectives
Let me illustrate in the following way. The 4 elements in the job description define two mutually-exclusive skill sets. (1) and (3) would fall within managerial skills. Perhaps, one might argue that these are portable across job domains. (2) tends to be a stretch for mainstream managers, for it calls for the encouragement of creativity and the adaptability of the group to presumably new technologies and applications. Encouraging creativity is far different skill than recognizing it after the fact! (4) calls for an yet a different set of skills - in this case those which are primarily technological and only secondarily managerial. The intersection of these two, typically contradictory, skill sets creates a target candidate which is almost impossible to find.
This problem arises when we try to secure the science and technology future of an organization in the same way that we secure its governance and administration. While this was probably never a great idea, it has heretofore avoided disaster because of herd mentality - most organizations try to do what other organizations of their type do. This produces a sort of tidal wave approach to technology evolution - everyone advances the technology front together. Incidentally, a corollary is that herd mentality also inhibits rapid technological change because true innovation has to overcome the momentum of the tidal wave. Large corporations and organizations tend to be unresponsive to sudden changes in environment, as was shown by the analog watch, typewriter, and published sheet music industries. High-tech industries are also vulnerable to such sudden downturns. As IBM showed when it experienced the largest U.S. corporate profit and largest U.S. corporate loss, in that order, within the same decade.
Position #2 is equally misguided, but for slightly different reasons. As with position #1, the incidental and irrelevant requirements are listed as essential requirements. But, what makes this ad so ironic is that the skills which I argue are sine qua non for a successful deployment of an information infrastructure, are demoted to the "desired" category. In other words, there seems to be some sensitivity to the importance of these skills, but with a concomitant lack of understanding about their relevance to the project at hand. This is also evident in the long list of desired skills which are practically impossible to satisfy jointly. It is hard for me to imagine how a claim to have "proven records" in so many robust areas of research and experimentation could reflect more than dilettantism. Entire careers have been spent on small fractions within each one of these areas. The "desired qualities" list cries out for superficiality.
The information age (and, most importantly, the Internet) has changed the rules of the forecasting game forever. Product life cycles in information technology may be measured in months. In the area of browser technology, for example, the first five World Wide Web browsers, Erwise, Midas, Viola, Cello and Mosaic, are of only historical significance – and all were developed in the current decade. Further, the standard protocol for indexing network resources in 1994, Gopher, has lapsed into insignificance as the Web soared past it in terms of both Internet packet and byte count in 1995. The hottest technology in 1997 - push technology - fell into widespread disuse in 1998 as MIS managers worldwide tried to hang on to some vestige of their bandwidth. The information age is very unforgiving of mistakes and misjudgments.
This is the milieu into which new info czars are to be placed. Strong people skills won't get the job done. Neither will the ability to govern disparate groups and constituencies nor the ability to manage open computing environments. If this be our captain, our ship is sunk. The solution is to re-define the role of info-czar into one of a strategic technology planner. The greatest rewards will accrue to those high-tech corporations who experiment with and continuously refine this notion of technology strategist so that they will bring the capability to shape policy around emerging technologies, and adapt to rapidly changing protocols, into the organization. They will be asked to define what is to be proprietary vs. public domain in their organization. They will be expected to anticipate future technological horizons lest their organization fall behind the curve. They will be expected to avoid technological surprises that might blind side their institution. These are expectations which are not managerial in nature. They call for technological leadership.
The solution lies in the recognition of a very basic fact: the way that we organize our institutions for administrative purposes is only loosely coupled with the way that the organization gets its work done. The skills required of our leaders are also very different. Executives in charge of science and technology, if they are to be effective in today's rapidly changing environment, will have different skills than those in charge of marketing and finance, for the former must of necessity be primarily scientific and technical positions.
The information needs of modern organizations are so rigorous and complex that a new position for Astrategic technology leader@ will have to be created at the highest levels of modern organizations. A position which reports directly to the corporate planner(s) without any intervening filtering or revision is envisioned. This position is presently viewed as an Ainternal consultant,@ charged with an objective assessment of business unit plans and objectives, the rapid summary and reporting of technology trends, and the accurate forecasting of relevant technology advances, which will have the ear and attention of the CEO. The solution is to re-define the role of "info-czar" into one of a Astrategic technology planner@ -not a manager or decision-maker, but someone capable of bringing the highest level of technology understanding to the while at the same time being grounded in the needs and objectives of the organization. This position will require both enormous depth and breadth within a technology area.
It is unclear whether our current organizations are adequate to this challenges of the next millenium.