The Silverlake Project: Transformation at IBM. By Roy A Bauer, Emilio Collar and Victor Tang. Oxford; 219pp; £20. Review by Tom Lester.
The earthquake that first hit computers a couple of years ago rumbles on, with the continuing possibility of further major shocks. IBM, still by far the biggest structure, is showing nasty cracks and is a few storeys shorter than it was. Its critics believe many of the company's ills were of its own making. It was oversized, over-rigid, and over-confident when flexibility, speed and humility were needed.
The fairness of these criticisms can partly be judged from this latest addition to the extensive IBM lexicon. The Silverlake Project analyses some of IBM's more entrenched failings, and - given that it is an inside job - with unusual frankness. The authors were and are senior IBM executives, and presumably writing with the blessing of the top management. With some medical conditions, getting the patient to admit to his problems is the first step on the road to recovery. Much of the same applies in business; on the evidence of this book the prognosis for IBM must be hopeful.
The Silverlake Project was the codename IBM gave to the programme to develop its highly successful AS/400 range of minicomputers. In the mid-'80s, the corporation was fast losing its share of the market for machines between mainframes and PCs, where prices ranged from $15,000 to $1 million. There were some 250 rivals in this segment, which was dominated by Digital Equipment's VAX range. Against VAX, IBM could field no less than five mid-range computer lines, all of them old and slow, and none of them - in marked 112 contrast to VAX - compatible with any of the others. The three authors acknowledge that this chaotic state of affairs had been brought about by departmentalism - also partly blamed for the failure of the massive, 16,000 man-year programme to replace all five lines with one compatible range. This was the ill-fated Fort Knox project, which cost IBM over $1 billion but proved beyond its management and technical resources.
In 1985 some of the managers who had been deeply critical of Fort Knox-and sidelined for daring to speak out - started their own more limited project. They had been exiled to a development centre in Rochester, Mionnesota, a thousand miles from the corporation's Armonk headquarters and more famous as the home of the Mayo Clinic. On the shores of Silver Lake (actually a reservoir), the centre was able to operate as a business unit without undue interference.
Assisted by a professor of marketing from the Wharton School, Bauer and Co have arranged their experiences of the project so as to align with 10 management principles. None is new, as the authors acknowledge, but they make a useful device for hammering home the moral of the tale. Thus, you need a leader with vision, and the right people to enthuse with that vision. Your staff must be empowered, and, where appropriate, organised in cross-functional work teams. You must research, segment and model your markets, and position your product accordingly. You must allocate resources according to agreed priorities, and cut development times by conducting processes in parallel and getting it "right first time". You must form partnerships with outsiders, particularly customers, and shape and exceed their expectations.
There's nothing startling here, either. But the AS/400 was duly launched in June 1988, and within two years was registering phenomenal world sales of $14 billion annually. The team went on to enter for, and win, the top US quality accolade - the Malcolm Baldrige National Quality Award. Yet the reader is left with the feeling that for all the project's success, the authors are more concerned to make a point within IBM circles than to explore the wider issues raised. For example, the failings which they noted in Fort Knox were peculiarly IBM ones, and the actions they took to avoid repeating the same mistakes were pretty elementary. Unavoidably, their solutions sound rather glib. Had IBM really never analysed its markets by industrial establishment rather than by company? Perhaps it hadn't. But having caught up with the world, the authors don't hesitate to congratulate themselves. "Most of all, we went from being a "nice little business" for IBM to being a nice big one ... We wanted IBM Rochester, as an ongoing concern, to stand out as the best of the best."
Parochialism is not dead in IBM. There's no mention here of the other major IBM project of the time, the RS6000, the work station which overlaps the AS/400 range of applications and to some extent competes with it for customers. There's no discussion, either, of the crucial marketing issue of "openness" and whether the AS/400 offers the ability to interconnect as effectively as the market now demands. No discussion, indeed, of whether all that effort might have produced a product for the 1980s rather than the 1990s. Good as the AS/400 undoubtedly was, it may not have been enough to shore up the great edifice in time of earthquake.
Tom Lester is a consultant and writer on business affairs.