IBM

9099 IBM4 - CONSOLIDATION & INSPIRED BUREAUCRATS

The final act of Tom Jr. was to plan his own succession. After his brother Dick fell from the race, stumbling or pushed, the one contender was T V Learson. It appears from contemporary descriptions that, although  he was an outstandingly capable man, he was a consummate politician and in his time the upper echelons of Armonk caused a number of casualties. He did however, like Tom Jr, recognise the need for a strong team; and Buck Rodgers, for one, believed "…he was tough and demanding, but fair". Thus in 1974, after his retirement, he said "Success is having good men around you", though, perhaps typically, he added "…all these fellows had guts, plain guts. They were willing to risk their job on any decisions". It is rumoured that Tom Jr. planned to keep him from the succession by the simple ploy of instituting a compulsory retirement age of 60; whereby TV would have to retire before him, since TV was the older. Certainly, as is made clear in his autobiography, Tom Jr. was less than impressed with his over-competitive attitude and the short-cuts he favoured. Once, after competing in a yacht race, Learson would not even speak to Tom Jr. for a number of days. It was not surprising, therefore, that Tom Jr. really wanted Frank Cary for his successor!

In the event fate intervened when Tom Jr. suffered a heart attack in 1970. Even then, what followed was typical of his highly individual view of management. In retiring from IBM, at the age of only 56 years, he said, of this decision, quite simply; "I wanted to live more than I wanted to run IBM". If nothing else proved that he was no ordinary seeker after power - as are many other top executives - this decision did.

Power Corrupts…it is a well known saying that power corrupts - and this is no less so in senior management. It distorts the individual's decision-making capacity; where such decisions can become indivisible from their own interests. Even so, the most influential managers I have met, government ministers as well as senior managers such as Tom Watson Jr, have not been driven by personal ambition but by a wider vision; perhaps that of the future of their organisation - but often the future of mankind. Management theory discretely sidesteps the issue but if you find such an individual, do not condemn them as naive. Help them build a better world - and your organisation (and you) will be guaranteed to play a part in that world.

Thus, after all, T V Learson assumed the chairmanship in 1971; at the age of 58. It was clearly not possible for anyone, not even TV Learson (TVR as he was known), to create a unique impact on a company like IBM in only one and a half years. He did, however, preside over the introduction of the 370 range. Unlike the 360, though, this could not be revolutionary but, constrained by the investment in the customer base, was merely evolutionary. TVR has been seen by some as a follower, rather than a leader, but this is to underestimate his dynamism. As he himself said, soon after his retirement, his philosophy was; "We've never stuck with anything ‑ its the excitement, the gamble, move, move, move".

His truly revolutionary machine, the Future Series ‑ which would have been initiated in his period of office, unfortunately never saw the light of day. Regrettably, though it would have genuinely revolutionised computing, the hardware at the time simply wasn't powerful enough to drive it; though Frank Soltis has suggested that its demise was - like Fort Knox later - due to the inability of of IBM's various labs to work together effectively![1]. His 'epitap'" therefore was the 360 which he masterminded, and which was, until the 1990s, and until recently at the heart of the 'architecture' of IBM's mainframe computers.

The end of the second phase of IBM's history, therefore, belonged to the bureaucrats, T V Learson and his successors (Frank Cary and John Opel), assuming power from the dynasty. Their first achievement was quite simply to manage that difficult transition successfully, and indeed to make IBM equally successful for more than a decade. Where so many corporations stumble when a great leader moves on, this was indeed a historical success. Paradoxically, it was what should have been the later handover between 'bureaucratic' leaders - which should have been infinitely simpler - which proved to be far less smooth!

Their second achievement was that of maintaining product leadership by moving IBM, and its customers, into the era of distributed data processing. Finally, and most critically, the campaign against bureaucracy was successfully extended, and in particular the very successful device of continuous change (something like the Japanese kaisen) was introduced, to deny bureaucrats a firm foothold.

After Learson's brief period of office, and for most of the next decade, this continuing 'revolution' was surprisingly guided by two archetypal bureaucrats working together; Frank Cary and John Opel. Frank Cary, especially, was significantly underrated by the commentators - though not by Tom Jr. who had wanted him to be his immediate successor. His style was certainly that of the self‑effacing bureaucrat. Tom Watson Jr. said of him "Frank was a brilliant business analyst, and totally self-confident...He didn't make heroic moves and he didn't make glaring mistakes; when he ran into a problem he simply figured out how to fix it."  On the other hand, the changes he put in place, both in terms of product strategies and in terms of the continuing development of management style, were as far‑reaching as any of those initiated by his more ostentatious contemporaries. The successful IBM of the 1970s and 1980s was almost as much his legacy as that of the Watsons. Indeed, Tom Watson Jr. said "...the fact was that for eight years he ran IBM as well as it had ever been run."

Frank Cary and John Opel were both of them long serving IBM'ers. Both entered IBM via the sales training programme (Cary in 1948 and Opel a year later; both after obtaining MBA degrees). They had spent virtually the whole of their careers in IBM; a large part of which was spent being groomed for stardom. John Opel in particular had an even more rigorous apprenticeship than Tom Jr, having visited almost every area of IBM during his decades long "training".

Succession Planning…this lesson should really have belonged to the chapter on Tom Watson Jr, for it was he who set up the succession. It was not just a matter of picking, and training, strong managers - something that Tom Jr, unlike many weaker managers, had the courage to do - but it was also about handing them a strong organisation; so that they had time to build their skills, and confidence, without threats from outside. Management theory has relatively little to say about succession planning and, in practice, too many successions are to weak 'yes-men' ( as perhaps was the case later when John Akers took over), especially when the previous CEO is departing to avoid emerging problems!

In reality, therefore, it was Frank Cary who first inherited the true mantle of the Watsons. He was seen by the media as very much the IBM bureaucrat. He was for example described by Katherine Davis Fishman in "The Computer Establishment" (Harper and Row 1981) as "…a man who would neither squander the inheritance by rash decisions nor let it diminish through excessive caution". Within IBM, however, his image was rather different to that of John Opel. Even so, T V Learson said of them "We have a bureaucracy, no question about it, but the fellows who were winning would cut right across it. John Opel goes where the action is, Frank Cary the same. They're tops intellectually"; and Buck Rodgers described Cary as "…a master strategist, who at the same time encouraged creative thinking". Certainly Cary did not have the charisma of the Watsons; and indeed deliberately chose to keep a low profile throughout his term of office. His second in command and successor (in 1981), John Opel, had a rather different reputation, that of being a very shrewd operator. A former colleague said of him "John is the most brilliant guy in the business, but with the brilliance comes a little autocracy". That too was my own impression, and that of my colleagues at the time, but I reversed my opinion when I later examined the record in more detail. Interestingly, with the hindsight which history offers, it is Cary who should now, justifiably, seen by the outside world as having been the most brilliant strategist and Opel the follower!

But Opel too maintained a remarkably low profile through the Cary era and his own. The most interesting feature was that it was - during the Cary era - a genuine partnership. Indeed one of his colleagues was reported as saying "Learson and Cary both wanted him (Opel) to be president, but John wasn't ready at the time. Cary took over". Cary ensured that Opel was involved in all the important decisions; a far cry from the strident political battles that reportedly enlivened TVR's days and the isolation of John Akers' later reign. On the other hand, it is arguable that this spirit of team-working led to the one inheritance of Frank Cary's reign which came to haunt the IBM of later times. Cary accepted, and indeed supported, a number of John Opel's more controversial initiatives - even when Opel was still his deputy. Thus, the new 'objectives', especially that of becoming the lowest cost producer, were announced in Cary's time; and John Akers subsequent progress to the top was also orchestrated at this time.

The low profile hid a continuing revolution rather than a decline into reactionary middle age. Even the absence of personal charisma was part of the new IBM. In much the same way that IBM salesmen were exhorted to sell IBM rather than themselves. Its very leaders themselves concentrated attention on the corporation rather than their own personalities; and IBM itself became the character in the spotlight.

Accordingly, with them putting so little on the record about themselves, it is difficult to write about this important period in the development of the successful IBM; though that has not stopped some writers doing just that and even going as far as attributing dialogue to the key players! In particular, there is the difficulty that the low profile adopted by the new management meant that they were less inclined than the Watsons to go on the public record as to their philosophies. In the case of Frank Cary, and John Opel when he was working with Cary (though not later when he himself became CEO),  one is forced much more to read hidden messages into the events, and this is always fraught with hazard. Accordingly, the philosophies reported in this book were the ones perceived on the ground within IBM, and were the ones that drove it to its greatest successes.

There was clearly, however, a commitment to continue the philosophies of the earlier years. Thus Cary stated to a reporter "If I were to select one single business procedure that was most important to our success in the early days it would be that we only leased equipment. This put a discipline on the business that was excellent. It motivated IBM people and it built a great relationship of trust between the customer and the company. The customer knew he had leverage". This statement holds a considerable poignancy, when you realise that it was the abandonment of this policy not long after he stepped down, which was a major contributor to IBM's later problems!

But it was still a period of consolidation. The new IBM could no longer be the personal property of one individual. Moving from a dynasty to more normal management, has been the downfall of innumerable other companies. Fortunately, Tom Watson Jr. had felt personally secure enough to develop a very strong management team, with a clear succession. It will be interesting to see what happens to Microsoft when Bill Gates eventually is forced to relinquish his power of control; for despite his seeming to hand over the position of CEO he is still very firmly in charge.

Certainly there was a considerable degree of financial security at the time of IBM's handovers. By the end of 1976, for example, IBM already held around $6 billion in very liquid assets (usually cash), and this made it in its own right (and with its own money!) one of the largest "banks" in the world; and it made, in the process, vast sums playing the newly volatile currency markets.

Consolidation…the obverse of the succession coin, however, is that the new team must have the courage to consolidate, and to build upon what has gone before. Too many new tenants immediately want to change the organisation so that it is clearly made in their own image. Yet continuity is usually the most effective policy, until such time as any changes (and then probably only incremental ones) can be justified. Change, for change's sake is far more destructive than no change at all! Once more, overall management theory does not focus on 'dynastic' successions.

IBM was now a mature corporation. Its structure was such that the senior management at Armonk could distance itself from the rest of the IBM world, confident that the business as usual would be well run, to spend its time planning the future.There are few senior managements confident or brave enough to take this step.

Expansion

At the start of the era the extended process of unwrapping the various evolutionary developments hidden under the covers of the 370 range was well under way. At the same time the development programme for the Future Series (FS) had just started. This was intended to be just as revolutionary as the 360 had been. It was to be the most open‑ended system ever, with flexibility built in that would be able to handle virtually any future developments. It would for the first time use a "relational database" to give the greatest possible freedom in data storage and handling. Finally it would be "user friendly", so that end‑users, increasingly communicating via terminals, would not find their way to computing barred by the technicalities.

It was a brave dream, but it was unfortunately not to be fulfilled. Thanks to Tom Watson Jr's investment in R & D talent, the technical capabilities were there; apart from just one. The computing overhead needed to drive such a complex system (even, and particularly, where it was to look simple to users) was just too great for the hardware available then; or, if you prefer Frank Soltis' version[2], the disparate teams could not work with each other. Whatever the reason, in 1975, therefore, the project was quietly abandoned.

It was a dream that haunted IBM long afterwards, and arguably interfered with the development of less ambitious (but still radical) schemes. It reappeared, at least in part, in 1977 in the S/38. The overhead to run a smaller computer under FS, particularly with a reduced scope, was thought to be containable; where such overhead was observed to be exponentially related to the basic size of the system. The architecture under the covers of the S/38 was a dream for the computer buff. As a member of the UK announcement (new product) management team I thought at the time, and even now, that it was the most elegant architecture that IBM ever produced. Almost every aspect of it was totally flexible, 'device independent' in the jargon, to the extent that it appeared as it would even support networked CPU's (something almost unheard of at that time, especially for a relatively small system). But then IBM always built a wide range of features into its computers and did not necessarily activate them. For example the largest S/3 (the predecessor of the S/38) reportedly had 'virtual storage' built into the hardware (a feature then reserved for the 370 range) but this was never activated. Most important in the case of the S/38 was that it had a true relational database; the first on a production mainframe. Even on the (relatively small) S/38 the problems were still to prove far from simple. The overhead was even more massive than expected, with the main operating system reportedly growing from 12 million to 20 million lines of code during the last few months; and absorbing ever greater numbers of development personnel, until it eventually matched the 2,000 used on the 360. Largely as a result it did not become the range that it was originally planned to be. The smaller S/36 that followed it reverted to 360 type architecture.

Accordingly, the machine ranges evolved. The 4300, 303X/308X and later the 3090 series (though significantly different in detailed hardware technologies) from the users point of view incrementally added power and marginal features; though even this growth allowed users to move ever deeper into the use of widespread data communications.

The real machine revolution of the period was outside the CPU, though driven by it. This was the explosive growth of teleprocessing (TP), as it was called at the beginning of the 1970's. Over the period this grew into distributed data processing (DDP) and moved on to true networking.  It is strange to recall that only a few decades ago, at the beginning of the Cary era in the mid 1970s, most computing still originated with, and revolved around, the humble punched card. Thus the end‑user of that time first created a written 'punching' document which was duly punched by a trained operator into the ubiquitous punched card, which was then hand carried (something of a problem when the user was in Edinburgh and the computer in London) to the computer room and processed (in a batch) on a CPU. The output from a printer was then hand carried back to the user by much the same circuitous route (and typically several days after the process started).

Distributed Data Processing

The revolutionary approach which appeared at the beginning of the 1970's was simply based on the concept of allowing users to talk directly with the computer. The user (wherever he or she was in the world) sat down at his or her own visual display unit, the now ubiquitous VDU (or more likely now his or her PC), and entered their own data. This was instantaneously transmitted by wire to the CPU; where the answer was just as quickly transmitted back to appear on the screen, in 'real time'. Not merely did this cut out unnecessary processes, and the related costs, but business information was available in minutes not days as previously; and a whole new industry, the provision of instantaneous management information systems, was born.

The move in the later 1970's was to distributed data processing (DDP), where part of the processing power itself was put closer to the users, often in mini‑computers based in their own departments, with the large central computers still running the main databases. In this scenario the user departments did not just communicate with the computer they actually controlled their own part of it and could determine how it was to be used. In the 1980's this moved towards its logical conclusion, with processing distributed to the individual who now had their own personal computer sitting on their desk. All these elements are now increasingly being linked by local area networks (LAN's) where the processing is automatically handled, locally or centrally (without the users needing to worry about this), by the most suitable processor. Beyond that, of course, we now have the world-wide Internet; but more of that later!

General Systems Division

In terms of Armonk strategy, it was Frank Cary's stated prime objective, from the end of the 1960's, to move IBM into DDP and beyond. The first obvious element of this was the creation of the infant General Systems Division (GSD) in 1969, to market smaller (S/3) computers. It has been argued that this division was formed to compete against the then emerging DEC mini‑computers. Indeed this probably was part of its function. It did not in practice keep DEC out of the scientific area, but it did stop it penetrating the commercial area; and in the process GSD in its own right became the second largest computer company in the world. With perhaps a quarter of IBM's overall business it eventually generated around $10 billion in revenue annually. It has also been argued, perhaps more cynically, that it was a possible sacrificial spin‑off, if needed, for the anti‑trust suit; and the dates of its creation, and indeed dissolution, do parallel those of the suit!

But the most reliable evidence shows that it was the first step in Armonk's, and Frank Cary's, drive to DDP. Thus as early as 1971, when many discussions elsewhere revolved around whether it was preferable to use 80 or 96 column cards, Frank Cary stated the revolutionary proposition that "The market will move to remote computing and non‑central processing unit equipment will be a continually increasing portion of the business".

The policy was consolidated in 1974, after Cary came to power, with an enhanced GSD adding responsibilities for creating new small computers and attempting different approaches to DDP. It was given the quite specific mission to compete with DPD (Data Processing Division ‑ the large mainframe division); healthy rivalry was to be the order of the day.

Distributed Data Processing was not a policy limited to the General Systems Division. DPD was also given the same strategy. In terms of hardware this eventually resulted in the 8100 series of distributed  mini's (after DPD's first attempt with the rather poorly received 3790 series). Most important of all was that it resulted in SNA (Systems Network Architecture); the basis for all IBM data communications in the 1980s. This was the concept, the 'architecture', behind the software which was to allow distributed data processing to become a reality for the larger systems and within which framework all IBM systems, with very few exceptions, came to operate. As with most such IBM 'architectures' at the time it became the industry standard. More recently, of course, Bill Gates has managed to impose his own standards on the PC software industry.

The new division GSD, deliberately based on headquarters in Atlanta (away from the mainstream), was however also used as a vehicle for things other than DDP technology. In particular it was used by Cary to stimulate innovation within IBM, and to provide some real competition for DPD; where at that time the outside competition was relatively weak. Its role as gadfly was reinforced by its initial staffing; with a number of leading 'dissidents' from other divisions being deliberately included .

The effect was somewhat diluted a year after the formation of the division, when in 1975 GSD was included in the larger grouping GBG (General Business Group). This possibly started with the best of intentions, increasing its competitive size and muscle power so that it could compete internally with DPD; though once again the possibility of a consent decree - as a settlement of the anti-trust suit - cannot be discounted. The dilution, in practice, was because the other piece of GBG was OPD (Offices Systems Division); the ailing typewriter division which had been declining for more than a decade, with corresponding problems to divert management. The result of this forced marriage was at best lacking in synergy; but despite this GSD still managed to remain successful in its own right.

Microsoft approaches the problem from a different direction. Its organisational structure is made up of a collevction of teams - based on its main product lines - which regularly change as the staff within them move to furrther their careers. Much as for university sports teams in the US, reputation sems to be all and the heads of these teams use their successes to keep their stars on board and bring in new talent. When growth is rapid, such ('market') freedom is effective in allocating resources, but as growth slows - and fewer new recruits are hired - resourcing the more 'support' work becomes ever more difficult.

Non‑Financial Controls…at the same time Armonk developed the panoply of non‑financial controls that are generally ignored, since they are normally very blunt instruments. Used in a very sophisticated manner, they allowed Armonk to steer its newly created bureaucracy. It is paradoxical that these ideas were later subverted, in the time of John Akers, to actually stifle the process of change!

At the end of his term in office, the real strength of IBM lay in the structural factors lying unobserved beneath the surface, and in particular in the culture. The one greatest success of Frank Cary (aided, at that time, by John Opel) was to consolidate an idiosyncratic, though 'revolutionary', family-run company into a workable bureaucracy; with all the positive features predicted by Max Weber. It was just as efficient and flexible as it ever was. This part of its history shows that a bureaucracy (using it in the sense that Weber originally gave it - and not in the popular, pejorative sense) can actually be efficient and indeed innovative.

The time of Frank Cary was, therefore, one of great achievements. I would, indeed, argue that his achievements rank with those of the Watsons. On the other side of the balance, however, must be weighted the fact that the seeds of IBM's ultimate decay also were planted in his time. He, along with the rest of IBM (myself included), supported John Opel as the next CEO; and, more important, he even sponsored some of Opel's more controversial initiatives. He even supported Opel's choice of successor, John Akers, and did nothing to oppose Aker's disastrous management decisions during the next decade when he remained on IBM's board. It was only after he retired from the board, and was released from the 'groupthink' which seemed to afflict that body, that he at long last recognised Akers' weaknesses - but by then it was too late.

History may, therefore, record that (as has been the case with so many leaders) his personal achievements were indeed great - but they were betrayed by the colleagues, and friends, he chose!

Capable Succession…its is not sufficient to merely pass on a business in good shape - but it must be passed into hands which will maintain it in good shape. Unfortunately, the selection of CEOs in general is still an art not a science!

Organisation by Controlled Anarchy

The time when Cary was CEO was the period when IBM, at its peak, demonstrated its greatest richness of management skills - particularly (in advance of its time) in Human Resource Strategies - which has never been equaled. For comparison, Bill Gates has achieved incredible success based on a much narrower range of talents; albeit ones which have come to importance in the 21^st century.

People Strength…at the seeming peak of its success, IBM's great strength lay in its 400,000 employees; who had been carefully recruited to be the best in the world. They had then been just as carefully trained, and retrained, to develop their full potential. It is arguable that Microsoft's comparable strength also lay in the recruitment of an especially excellent group of key employees, albeit smaller and with a much narrower range of abilities. But IBM's greatest genius lay in managing delegation so that this full potential was realized for the benefit of IBM; where almost the opposite is true of Bill Gates.

This delegation would in any case normally be demanded by personnel of the caliber of those working within IBM, but IBM ensured the process happened. It created a subtle form of 'anarchy' which forced lower level management to act as 'leaders' (co-ordinating the co-operative efforts of their team members) rather than as hierarchical demagogues. Indeed, a lack of formal structure had originated much earlier, with Thomas J Watson. As Tom Jr. explains, in his autobiography "We had no organisation chart because Dad didn't want people to be so focused on specific jobs that they concentrated only on those jobs...He wanted everybody to be interested in everything."

Structural Blindness…the (hierarchical) organisation charts so beloved of most organisations, can artificially restrict the horizons of those constrained within them. Management theory has recently started to explore alternative structures - in particular networks - many decades after the Watsons did so!

Indeed, many organisations - especially the high-tech ones such as Microsoft - are adopting peer to peer networking; where the collegial atmosphere (for example on Microsoft's  Richmond 'campus') removes the need for hierarchical control (though Bill Gates does still demand a modern form of 'droit de seigneur' at higher levels!).

Cellular Organization

These individual, 'anarchic', cells were supported by a powerful horizontal communications network; where such communications are more normally vertical. In IBM at the time this network revolved around a 'grapevine' of great sophistication; but this was being increasingly supplemented by electronic data networks (with, even during the 1980s, each employee having his or her own terminal). 

Based upon the material contained in Tom Watson Jr's autobiography, it appears that this outcome came about by pure accident. There is no evidence that, as traditional management theory might suggest, it undermined IBM's management strengths, and led to its downfall. Very much the reverse. The later picture on the ground was confused, but it appears that senior management later attempted to impose more disciplined traditional management practices - of the kind taught by the Harvard Business School - and this did in part lead to the subsequent downfall.

This horizontal communication was aided by the (encouraged) existence of large numbers of 'fixers'; senior personnel (but not managers) who knew their way around the very complex (and bureaucratic) IBM systems, and had built networks of contacts in other departments (in the same way as the Japanese managers, who perform the same function). The evidence is that it is precisely this group of individuals who IBM later got rid of first! For example, in the UK's key staff group, supporting the marketing of large mainframes, all the middle and senior management, between them representing almost all the expertise, left in one three month period during 1993! Once again, Microsoft - working on one main campus with closely related work groups - does not need to formalize any of this (or even to create informal networks).

Fixers…management theory makes little or no mention of the role of those individuals who (usually acting informally, without any authority) co-ordinate the activities across various parts of the organisation. Yet, in increasingly complex organisations (especially those making use of networks), these individuals become an essential part of the problem debugging process. Management theory simply does not allow for this role in any guise.

Corporate Planning

Up to the time of John Akers, the other element of IBM's organizational strength lay at the opposite extreme, in the senior management; at Armonk - much the same as, in Microsoft, these reins are held by Bill Gates. However, IBM's normal, world-wide, planning horizons routinely covered 5 years, and its longer term planning (particularly in terms of new plants) regularly stretched beyond 10 years; yet its planners admitted that their financial models (which were just about the most sophisticated in the world) were only accurate up to 13 months ahead! Possibly as a result, Armonk had a (relatively unique) awareness of its own potential fallibility; and as a result planned for almost every eventuality (so that when the worst did happen there was usually a contingency plan to limit the damage, and when the best occurred IBM could make the most of it). You will not be surprised to learn that this, too, has gone. By the beginning of the 1990s, at the country level at least, the planning horizon was reduced to a maximum of one year!

Below Armonk, however, the management structure was probably weakest at the most junior level (but this was contained by the "team leadership" approach). As the level of management rose so did its quality (where the reverse is true in many companies); for IBM chose its senior management very carefully indeed (measuring their track record by their subordinates comments in the Opinion Survey just as much as by their performance against revenue or profit targets), and it trained them by moving them around almost as much as their Japanese counterparts.

Unlike the Japanese, IBM implemented (and still does, for once) strong financial controls; of which perhaps the most stringent was that each group, and each country, should not merely be profitable but should also fund its own growth (not an easy task where high growth rates are endemic). But, like the Japanese corporations, it was the non‑financial controls that were in reality the most important.  Once again, this approach went back to the early days of IBM. Tom Watson Jr, in his autobiography, said "...Dad didn't particularly trust numbers. He thought they could distract a businessman from the real issues at hand."  In happier days, however, it was the "Sales Plan" (directed at the sales teams in the field), rather than the overall Marketing Plan, which most directly shaped how IBM's revenues would be generated. It then controlled its expenses very effectively by strict controls on capital expenditure and numbers of personnel (headcount), as well as by more direct means. Ultimately the shape of the organisation itself was used to exert control; and in the longer term manipulation of the culture itself was perhaps the most powerful control of all. It is difficult to identify exactly what has happened on these fronts since that time; though it was obvious that, by the early 1990s, IBM senior management was no longer in control (but whether by choice or by force of circumstance or, perhaps most likely, by ignorance is not clear).

It was a complex and sophisticated organisation (and one that was ever changing, to avoid the onset of bureaucratic middle‑age),and perhaps that is why control was eventually lost, but at its peak it offered rich lessons of control by methods not even hinted at in most management text books (and certainly not in the texts which IBM management have been reading since losing confidence in their own abilities).

Personnel Policies

While, the apex of the pyramid, global control was exercised by a very small group of people, the Central Management Committee (CMC) located at Armonk, at the other were the 400,000 or so workers at the "sharp end", whose implementation of the policies was never been less than impecc­able.

Recruitment…as with any company, and particularly with a Theory Z company, the greatest strength of IBM ultimately was the outstanding quality of the latter group. IBM was able to, and often did, employ as "workers" people who would qualify as directors in many other organisations; and could afford to pay them as such. The quality was evidenced by the number of graduates in the workforce. In the UK, for example, even in 1982 they accounted for 30% of the total workforce, where only 27% of it had less than 5 "O" levels.  This was just one indicator though; for IBM, as an avowedly "single status" company, held no special brief for grad­uates, and the figure was merely a reflection of the quality of its recruits in general. It was reported that IBM then received in excess of 100 applications for each job it sought to fill from outside the company - the sort of ratio which Microsoft now enjoys instead! Natura­lly it could choose the cream, and it could even afford to indulge the most idiosyncratic choices of its managers without seriously diluting the overall quality. The Japanese corporations were in the same fortunate position.

To reach even the stage of the first interview they had to show a sound employment track record. They would, of course, have had to demonstrate the degree of charm that is an essential prerequisite to passing through a number of such interviews, whatever the company doing the recruiting. Despite Buck Rodgers' claim that "only first‑line managers do the actual hiring in IBM", in practice a whole range of staff and management (usually including relatively senior management if it was sales‑force recruitment) interviewed potential recruits; IBM took the process very seriously. There was, however, one relatively unique requirement that, with very few exceptions, they had to meet and that was passing an aptitude test. In theory this merely ensured a minimal numerate ability, in order that they may be able to understand the complexi­ties of computing. In practice it was really a rigorous intel­ligence test that ensured that all key personnel entering IBM had a high intellectual capabil­ity. IBM was just as prone as any other company to recruit its staff on the basis of the rationalized intuit­ive hunches of its managers but, despite this, the aptitude test ensured that in the end the odds were heavily weighted in IBM's favor; in terms of achieving a high caliber of personnel, even when afflicted by the worst vagaries of the recruitment process. The benefit of this one simple screening should not be underestimated as a contributor to the undoubt­ed high quality of the key personnel within IBM.

The Multi-Million Dollar Investment…the most important investment any organisation makes is in its people. Thus, one of IBM's middle-rank employees might cost in excess of $100,000 per annum (if all the overheads are allocated). Assuming the employee stayed for just 20 years, hiring him or her might therefore lead to a total spend over time of more than $2 million. Yet most organisations hire staff on a less rigorous basis than they use to take the decision on a new photo-copier! Much more important is the fact that staff now represent the major resource of an organisation. Get your hiring wrong and you get your whole operation wrong! HR theory is predicated on the value of human resources; but even it does not usually recognize the scale of the financial investment!

It is interesting to note that Tom Watson Jr claimed that 'The part of my work I loved best was picking and choosing men.' Once again, the lead came from the top.

Training…having recruited the highest caliber of personnel IBM then trained, and retrained, them. It was recognized that all personnel new to IBM must receive suitable induction train­ing no matter what their job. This was no new fad, for one of the very first things T J Watson estab­lished on arriving at IBM (then still CTR) were training schools for its staff; and in particular for its salesmen. The one year plus that this training took for field personnel was just one extreme version of this. It was recognized, in addition, that staff training was an ongoing commitment and all staff were expected to attend regular courses, adding up to an average of 5% of overall staff time and typically as much as a month a year in the case of the field force. At its peak, during 1984, internal education cost IBM some $600 million in direct costs alone.

Maximizing the Return on Your People Investment…having already demonstrated the scale of your investment in people, it should take little further evidence for you to realize that training them to be more effective offers incredibly good value. Everyone, from governments downwards, talks about Life-Long-Education (LLL), but almost nobody does anything about it!

The Cult of Individualism…by the end of the 1970s IBM had, over its three quarters of a century of existence, deve­loped a resource of some 400,000 very high caliber and especially well trained personnel. The final element in the structure that then maximized the effective use of this resource was a commitment to delegation, to the maximum degree poss­ible, at all levels. This came about mainly as a result of the caliber of personnel involved, since they were independent minded enough not to settle for anything less. It was, though, also enshrined in IBM's personnel policies. Every employee had, as of right, the maximum delegated responsibility and authority for his or her own particular func­tion. To a degree in theory, and to a large extent in prac­tice, management was asked first to provide support for the individual rather than control them. Thomas J Watson's approach was reported by his son: "Dad always preached the idea that executives and managers should see themselves not as bosses of employees but as their 'assistants'."

Management as a Resource for Staff…Thomas J Watson's concept turns management theory on its head. Staff are conventionally supposed to be the resource which managers control. But adopting this upside-down viewpoint, whereby managers support their staff, is a very powerful device for seeing what a manager should be doing in practice rather than in theory! Despite the power of the idea, it definitely is not part of management theory.

Following this lead, IBM had to a very large degree managed to break down its operations into cells that could be auton­omously run by one individual. This reached its peak in the field force where the sales-person was uniquely responsible for all IBM's activities affecting their customers. He or she was the single point of contact co-ordinating the many facets of IBM's activities, and they carried the prime responsibility for ensuring that the customers needs were suit­ably met, and sales accordingly maximized; IBM's policies usually had a secure basis in pragmatic self‑interest. More important was that he or she was also given the comparable authority to guarantee that all this could happen. Perhaps uniquely among most industrial companies they were given the privilege of deciding the best policy for his customers, even if nece­ssary over their own management's views. They had the right to set such policy, and management had to have their agreement to any changes; though it has to be admitted that it was a brave, or foolish, individual who risked their career by insisting on their rights in this way. Yet again, the field force was but an extreme example of the general degree of delegation that applied throughout IBM. Once more the position since that time was confused, but sales force delegation probably is less well practiced - not least because it is now difficult to control the overlap of responsibilities with dealers and agents.

Delegation…much lip-service is paid to the need for delegation, but little normally results in practice. In IBM the result was guaranteed - and the outcome was the most powerful form of management obtainable. Delegation is no easy management task, but it is a most essential one. Management theory has, over the years, gone on endlessly about delegation; but the practice still does not seem to have got though to most practicing managers.

For many outsiders the typical IBM'er has been seen as the archetypal organisation man, facilely fitting into a rigid bureauc­racy. The truth, even then, was almost exactly the reverse. To a large extent IBM comprised some 400,000 individualists each with his or her own defined territory which they jealously defended against all comers. In effect IBM consisted of some 400,000 separate small businesses, and it was this "anarchy" that was at the centre of its success, for each business was superbly well run. Each was the personal responsibility of a uniquely well qualified "proprietor" whose expertise in their own special area was unchallenge­able. The creation of a cult of individualism, under the promotional banner of the "wild duck", was Tom Watson Jr's greatest ideological legacy to the company. This form of organisation is more typically seen in 'collegiate' organizations. The most obvious are usually universities, but professional groupings - such as the leading management consultancies - are also often organised in this manner. This may, at least in part, explain its success with IBM - which was also to a large extent a collection of professionals! It may also be, in a rather less well directed and more informal context, one of the reasons why Microsoft's Richmond 'campus' manages to overcome Bill Gates many small mistakes (but the real test will come when it has to handle a real disaster!).

Collegial Organisation…organisations are being increasingly staffed by 'professionals', to whom power is maximally delegated. Under these circumstances the optimum organisation is 'collegial', very much like that which has served universities so well for many years - but now spread to many more organisations. It may seem like anarchy, but - with the correct individuals and correct cultural controls - the freedom of action it offers best matches the rapidly changing environment of modern organizations. This is one model which some academics, at least, are suggesting may apply in the future.

Horizontal Communications

The difficulties inherent in any attempt to manage such anarchy are obvious. IBM's formal response might have been in the first instance to deny the merest possi­bility of any anarchy; and then to point out that it employed a very high ratio of management to staff; it aimed to have no more than seven staff reporting to a manager, to guarantee effective personnel management - the level that it is, now, accepted is necessary to manage 'professionals' working on non-routine tasks. There was indeed at least an intention to this effect and, albeit to a lesser extent than claimed by IBM, such a massive investment in management must always have had some practical impact. Recently, with the emphasis on traditional management theory, this might even have become the main mode of management. At its peak, however, more important (though unrecognized by many in IBM) were the very strong cultural factors that encouraged co-operation between the many disparate frag­ments of the organisation. For it was the horizontal communication, between individuals at the workface, that was IBM's particular strength. It may also be a strength of Microsoft; but for the very different reason that it has far fewer staff, spread across fewer 'disciplines' and limited to one main campus - so there is nothing to stop you talking face to face with anyone else!

Despite is global structure, the degree of horizontal communication in IBM was indeed very high and, although the inevitable political games were played, these were less likely to be the most debilitating "zero‑sum" confrontations. At the heart of the unusual willingness to co-oper­ate was a degree of individual security and self confidence, based, most notably, on 'lifetime employment'; one reason why the later revoking of this caused such catastrophic long-term effects! This also reflected on the one hand the caliber of the individuals, and their secure grasp of their own speciality, and on the other the widely recognized rights of the individ­ual to be respon­sible for his or her own special territory; unthreatened by incursions from others. From this strong position it was easy for individuals to build the communication links that were essential to make their part of the organisation work.

This cross boundary communication was made easier because inter‑departmental taboos were largely absent, once more partly as a result of the culture. In addition it was, at least in part, a result of the "single status" policy, which tended to break down the barriers between management and the other ranks to a much greater degree than in other companies. An IBM manager was seen in some respects as just another specialist. In the main though it was a func­tion of the personnel recruited, who tended to be good communicators with a natural desire to communicate with others.

The main communication within IBM, therefore, was horizontal between; peers. This was tacitly recognized, and encouraged, by each individual having their own terminal linked into a world-wide network, even in the early 1980s. On this network the individual even then could communicate at electronic speed with any other individ­ual in the organisation. Interestingly, long before the widespread acceptance of email, the IBM experience showed that it was this person to person communic­ation that was the main currency of such a network, rather than the use of the large centralized databases that was then predicted by many information researchers. This horizontal communication, be it face to face or more frequently by tele­phone (or now increasingly through the Web), is another factor that still diff­erentiates IBM from many other companies. On paper the structure is the usual pyramid, but this hides the fact that the most important layer is at the bottom and horizontal; with the upper layers in many respects uninvolved in most of the day to day communications. Such one or two dimensional representation though does not do justice to the richness of peer group contacts of the average IBM'er.

Much of the above description has its parallels in the organisation of the larger Japanese companies, making up the third of that country's economy that is most visible to the West. Kenichi Ohmae[3], for example, comments that "Most Japanese corporations lack even a reasonable approximation of an organization chart…Grossly simplifying, one could say that in Japan every member of the village is equal and a generalist." In particular, the identical factors of maximal amounts of delegation, degree of job security and amount of horizontal commun­ication are held (at least according to William G Ouchi) to be mainly responsible for the dramatic success of such Japanese organ­isations. The importance of IBM's contribution was not just that - as we have seen -  it was probably the original model but that it was a better, richer, implementation of such philosophies, and certainly was one that was much better matched to Western cultures; particularly those then facing the problems of the emerging "Information Revolution". Its implementation potentially had signific­antly greater depth because ultimately it was based on the individual, rather than on the group which is the basis of the Japanese version. It is better suited to western values precisely because it respects the individual and does not require complete subordination to a group culture. IBM does not require its employees to meet each morning and sing company songs, at least not any more!

The Power of Horizontal Communication…where communication is now so time critical, and its volume has increased so dramatically, vertical communication lines are no longer suitable. Instead, horizontal (peer-to-peer) communications are necessary. This is, in any case, a natural outcome of the increasing use of computer networks. What is not appreciated, however, is - at one extreme - the cultural impact that this has (where it forces Western organisations to behave like Japanese corporations) and - at the other - the cultural changes needed to make it work effectively. Such communications are, indeed, culturally mediated; and if the culture cannot handle them true anarchy may result! The IBM historic experience offers a very good model for Western firms to follow; and also illustrates the very high level of investment - in people rather than technology - needed. Despite the Japanese experience, until recently this aspect of the emerging structures is little featured in theory - though academics are now pursuing the Internet with some zeal.

ARMONK: the Centre of the Earth

We now move from the profane to the sublime, to the rarefied heights of Armonk; and to the small, very exclusive, club that was IBM's Central Management Committee. From the Watson dynasty to the end of Cary's time of stewardship this committee had an enviable record of sound decision making. This was perhaps partly a function of doing all those things that every professor of business management would recommend; though, (as we will see later) the evidence possibly shows that in practice the reverse possibly was true. But IBM was so profitable that it, perhaps alone of companies, could then afford the luxurious overheads of planning over an extended timescale. Where other companies plan for, perhaps, up to three years ahead IBM looked at more than a decade ahead; though, once again, this foresight became another casualty in the 1990s (where country plans were reduced to cover a year or less - rather than the previous five or more!). Such very long term plans were (and still are - despite their abandonment) essential, for although IBM had then managed to reduce the average product development timescale to around three years (and since down to a matter of months) it had been unable to similarly reduce the time needed to bring on stream factories employing new technologies to much below five years. The management committee, therefore, has  to try and guess what will be happening, at least in broad terms, in a decade's time; in probably the fastest moving and most volatile market of all. Although this might, for example, have locked IBM - for several decades - into water cooled technology for its large mainframes it still allowed the freedom to delay the determin­ation of the exact product offering, such as the 3090 range, until just three years before its delivery to the market; and in an emerg­ency, such as that which produced the PC, it could even then occasionally act in a matter of months.

Stratplan…if the very long timescales of plant technology are taken out of the equation then the most important cycle was that of the strategic plan, the Stratplan in IBM terminology, which used to forecast five years ahead (the later one year forecast could not be considered, by any stretch of the imagination, to be a true strategic plan!). The Stratplan was an impressive process. Its basis was a (computerized) financial model that represented not merely a single country, as for example does that of the Treasury in the UK, but it encompassed the whole world; shades of world empire! Depending on the assump­tions fed into it out, it automatically churned the sales forecasts by broad product group and by country over the next five years.

The problem is that, inherent in all such financial models, the output was only as good as the input; GIGO, Garbage In Garbage Out, with a vengeance. In practice, although the financial model(s) used take into account a large number of factors, over a number of years ahead, the IBM forecasters readily admitted that the only really reliable trend indicators in their experience, in the UK for example, were the relatively crude, and short term, 13 month ahead indicators covering Financial Surplus/Deficit, FT 500 Share Index, 3 Month Interest, GB House Building Starts and, in particular, CBI "Business Confidence". In many res­pects the latter was the most important indicator, where - at least in the context of its servers and mainframes - IBM sells capital goods which as investments in the future are highly sensitive to the business mood. The problem is quite simply that no matter how nice it is to accurately predict what your business volumes will be in a years time, this will be no consolation if the plant commissioning plans five years ago or even the product plans a mere two to three years ago were based on different figures, in the event now proved wrong. IBM has contacts world-wide, which it uses to input the best possible informa­tion, but if all the experts are caught off guard so is IBM. It is certainly true that IBM was taken very much by surprise, albeit for the first time in a number of decades, by the downturn in the industry at the end of 1984. It was itself in the middle of a very ambitious exercise, involving large scale investment in new plants, to treble turnover by 1990. The hiccup in the mar­ket left IBM with unsold inven­tories and under-utilized new plants coming on stream. Even then it took nearly a year for IBM to recognize the scale of the problem and take all the necess­ary steps, including starting to talk about reducing the total of its workforce; for the first time in living memory.

So if one compared - as I did - the very impressive, and indeed apparently infall­ible, Stratplan figures from year to year then one sometimes found quite large changes, perhaps by as much as 30%; which unfortunately invalidated many of the claimed benefits of long term planning. It is just such 'failures' which brought long-range planning into disrepute during the 1980s.

Long-Range Planning is Essential...the lesson which should have been drawn was that strategic forecasts have to be handled in a very sophisticated way - they are the start of the planning process, not the end of it. The development of scenario planning techniques (which, even now, relatively few organizations - most notably Shell International - make use of) is just one technique which allows planners to examine alternative futures rather than predict fixed ones. The study of the future is no less important (for it is long-range forces which most often destroy organisations), but it has to be properly managed. Because of its importance for long-term survival, this is the area where our own main research efforts have been applied in recent years. Almost half of the 500 organisations taking part in our most recent research had planning horizons which stretched no further than three years ahead and only a sixth looked beyond five years ahead. Moreover, two thirds of them did relatively little scanning or even considered alternative futures.

Awareness of Fallibility…one of the main virtues - an almost endearing one - of IBM executive management in Armonk was, up to the time of Frank Cary (though, surprisingly in view of his intellectual abilities, not in the time of John Opel), that they were sophisticated enough to make the quantum leap to appreciating the limitations of their own figures and of their potentially high degree of fallibility. On the other hand, I doubt that Bill Gates ever doubted his own infallibility for a second! IBM's key documents were not the impressive computer modeled Stratplan forecasts but were (and probably still are) the hand-written summaries that encaps­ulate the management committee's own expert interpretation; and it is these rough hand-written notes that are often the most secret of IBM's information treasures. Even so, to the best of my knowledge, no‑one in IBM or in the industry as a whole foresaw the 1984 down­turn; which had, even for IBM, very embarrassing consequences.

The Real Plans…IBM is, I suspect, fairly typical in that, at the same time as it had very sophisticated ('published') plans for wider consumption, its real strategic plans were held on a few scraps of paper. It was these 'real' plans that (perhaps because of their necessary simplicity) which really decided the future. Apart from Henry Minzberg's 'Emergent Strategy' theory, which does allow for events on the ground to intrude into strategic decisions, almost all theory focuses on the formal planning systems  which - in our experience - are largely ignored by managers in the field!

An awareness of its own fallibility was indeed  perhaps the greatest, and to me most endearing, virtue of IBM management.  It was perhaps best summed up by the comment of Tom Watson Jr that he always "ran scared". I have, perhaps almost sacrilegiously in IBM terms, a mental image of Frank Cary arriving each morning at his office desk and - following the rather odd habit that he and John Opel followed, of working standing up at a specially high desk - head held in hands commencing to worry about the possibility of IBM's imminent bankrupt­cy; his task for the day being to rescue IBM from the brink of dis­aster. Clearly this is pure fantasy, because IBM must then have been more secure than almost any other organisation including many governments; though it must have been very much closer to the truth later! But a very real desire to constantly reappraise the future, and contingency plan for all potential problems, was the virtue that it seems to me must be almost unique to the IBM of that time. Again this was perhaps a luxury afforded by IBM's wealth and actual secu­rity. Its management could contemplate the unthinkable as a matter of course.

Fallibility as a Philosophy…it is a foolish management which sets out to be fallible, but it is a wise one which assumes it will - from time to time - still be found fallible. Such wise managements are prepared for the occasional mistake, will rapidly recognize it and will rectify it before excessive damage occurs. In the middle of the 1980s, at the peak of IBM's success, John Akers must have assumed he was infallible. Managers generally do not like to talk about fallibility, so many business schools humor them by not introducing the subject.

Junior Management

The brunt of bridging the gap between Armonk and the "workers" is in theory borne by the thousands of IBM junior and middle managers. In practice the problems that are endemic in management throughout all large organisations are just as prevalent in IBM. It would be claimed by IBM, for example, that there is a rigorous procedure behind the choice of management candidates. The reality is that at the junior level in IBM this choice is no more scientific than in most other companies. The first requirement of a putative junior manager is often no more than a burning ambition to be promoted, which  needs to be clearly articulat­ed to those of more senior management who can best influence this process; in fact at this junior level sheer persistent bloody‑minded pursuit of promotion may well by itself succeed, regardless of merit. The second requirement is the will and ability to do exactly that which is needed (over the seven to ten years it eventually, on average, took) to overcome the political hurdles en route to the higher level, regardless of the theoretical or pract­ical exigencies of the current job. The popular culture of IBM deplores these practices and Personnel Department might want to deny their very existence, but, as in any larger company, the junior management system thrives on them.

Quality of Recruits…even at this junior level there were ameliorative factors, resulting mainly from the culture rather than IBM's earnest and praiseworthy attempts to train its management. Of these perhaps the single most important was simply the very high quality of the pool of potential management recruits. No matter how poor the selection process, the minimum standard of qualifying employees was so high that any manager was likely to more than match the best found in other compa­nies. This was brought home to me by the fate of a colleague who was not consid­ered to be of very high caliber within IBM, and who recognizing the limits on his promotion prospects in IBM despite his rather frenetic attempts to achieve this by any political means, chose to face the cold winds of the world outside IBM. When I met up with him only a year later he had already graduated from the junior management position he moved to on leaving IBM, to become, after a series of very rapid promotions, the vice‑president in charge of all non‑US marketing activities; albeit in a medium sized company rather than a giant the size of IBM. This brought home to me just how scarce in most companies were managers of the caliber that IBM rated merely average in its ordinary staff. For many years the rest of the industry in effect used IBM as a training ground for its senior management; and, indeed, not a few of the managers at Microsoft and AOL have this background to their credit.

Guaranteed Management Success…if the quality of all your recruits is high, no management system can be ineffective - since (no matter how poorly run - and most are just that) almost every manager selected will still prove sound! This rather crude approach, though effective, is rarely discussed in academic circles.

Forced Delegation…the next factor that ameliorated the potential problems, before IBM found traditional management theory, was that the junior manager's authority was often, by accident or design, "underm­ined"; so that delegation was in effect forced on them. This was accom­plished by a variety of policies ranging from that of "single status", everyone from dishwasher to director ate in the same staff restaurant and was on first name terms, to the inviolable (at least in the UK) requirement that management worked at round tables not desks, so that there could be no inherent status - indulged in by the manager being allowed the position at the 'top' of a rectilinear table - evident in their meetings. Active dissent was even encouraged by the language. In its obituary for Tom Watson Jr., the London Times[4] said "He [Watson] created a policy of internal competition among IBM managers, known as the 'contention system', under which almost anyone was allowed to challenge the decisions of others, even their superiors, and force them to explain their reasoning." I never heard it described as the contention system, but it was true that in IBM one did not even have to go out on a limb to disagree, one instead "non‑concurred"; a much less loaded, if rather esoteric, term.  Indeed there was a formal right to such non‑concurrence!

Non-Concurrence…supporting the right of challenge, and indeed encouraging challenge, is a powerful means of stimulating change. It also opens up debate to all members of the organization - and involves them positively in decision-making. The concept appears in creativity theory, but is only dealt with by the rest of management theory in terms of handling the problems caused by such challenges to authority!

Paradoxically, it was later suggested (most notably by Paul Carroll in 'Big Blues') that these policies actually hampered IBM's ability to react to change in the 1980s and 1990s. If that was the case, and I doubt it, it must have been due to the misapplication of them (which, in the context of the many misunderstandings which occurred during Akers' period in office, was quite possible). In any case, perhaps acting on the then recently published views of Carroll, Lou Gerstner banned 'non-concurrence' almost as soon as he arrived at IBM.

Vigorous discussion was also aided by the fact that there was less emphasis on a "hen‑pecking" order based on management position. As is also the case in Japan, where job grades are also paired with 'status' grades, the status of an individual was, in fact, determined by his "level" (discussed earlier in the chapter on HRS), which was unrelated to any position in management; and (particularly as it was supposed to be confidential) was also correspondingly less easy to determine. The junior manager might thus have been dealing with non‑managers who were actually of higher status than themselves (and, as a matter of safe political protocol, this was the basic assumption they had to make in any discussions). Accordingly, they often became only one (equal) member of a team, albeit usually "primum inter pares". The result was a natural emphasis on teamwork between equals.

Delegation by Equivocal Status…contrary to the theoretical benefits offered by clear structures, in practice delegation to the individual best suited to a task is often optimized where there is some flexibility in the status system - and, especially, if there is doubt about the power positions of those involved. Under these circumstances it is more likely that decisions will be based upon facts rather than on opinions weighted by status. The devices used by IBM, such as single status with hidden levels and encouragement of non-concurrence, powerfully reinforce such equivocal status. Management theory, if anything, tends to stress the importance of 'responsibility' (i.e. authority) levels.

Meetings

The final factor was the workload (where IBM deliberately "overloaded" all its staff) and in particular the meetings load. It was assumed, quite correctly, that IBM management would spend the greatest proportion of its time in meetings; another justification for their mandatory round tables. This meeting time, it was argued, was an essential part of the communication and control processes. The truth was, though, that most such meetings were with their peers rather than their superiors or subordinates. The net effect was that, although it enhanced the horizontal communications processes, it reduced the time that they might have had available to interfere in the activities of the truly productive staff below them who were anyway quite capable of delivering the goods IBM wanted without any supervision. Even so, the clear diversion of management time may explain why Gerstner (in true MBA style) later launched a campaign to reduce the number and length of meetings.

The meeting was a special element of the IBM culture, for not infre­quently an individual's overall performance was most importantly of all judged by their "theatr­ical" performance in front of their peers. As a dyed in the wool marketing company everyone, even the most unsuited, was judged by the standards of sales technique; of which the presentation was the most avidly followed. In the good old days the medium was the flip chart, but technology moved forward and the focus of IBM meetings became the foil (in IBM parlance, but acetates or slides in the language of management in general), projected on an overhead projector; no doubt it is - as with many other organisations - the computerized display projector (with managers desperately trying to outdo others with the sophistication of their animated graphics!). This saved time in production, but the one upmanship race moved to producing foils by the most impressive personal computer pack­age; and thence to computer projection itself. The form of the meeting was, therefore, something rather special to IBM. It is easy to make fun of it, but the end result gener­ally was a well run meeting, with good communications; again, for whatever questionable reasons, a net gain for IBM!

The centrality of the meeting to IBM culture may explain the one really dramatic gesture Louis Gerstner made. It was to ban the use of presentations! Managers thereafter were forced to explain their ideas by the use of words alone. It is the sort of seemingly trivial gesture which is often used by incoming managers to impose their stamp on an organisation - but was it effective in terms of the outcomes from group processes?

Management by Meeting…despite the jokes, management meetings can be - and should be - very productive (and they need to be if the large amounts of time involved are to have any justification!). They are an especially powerful form of horizontal communications combined with communal decision-making. It is a feature of Western management thinking that the brave individual decision is often respected more than that taken by a group; yet it is the latter which is most often the more effective (not least in implementation - by all those involved). A hidden benefit, however, is that the tasks which cannot be undertaken because of time in meetings must be delegated! Most theory would probably subscribe to the view that a camel is a horse designed by a committee!

At the end of the day it is probable that IBM did not deliberately set out to undermine the position of its junior managers. The reality, however, was that their workload coupled with the strong influence of the personnel policies did ensure that their position in relation to their staff was more equivocal than usual, and their role could best be described as team leaders, "primum inter pares" (very much in Theory Y mode), rather than hierarchical managers (in Theory X mode).

Middle Management

IBM was able to control the quality of its middle management processes rather better than those of its junior management. It was indeed a remarkable feature of IBM, before the time of John Akers (who was the exception that proved the rule!), that in general the quality of management improved as their level increased; the reverse of the experience in many other companies. This was a direct outcome of the increasing ability of IBM to measure performance the higher the level. The manager at the higher levels had a much greater degree of involvement in setting, or at least in accepting, the targets he was req­uired to meet. In turn the targets themselves were generally more measurable and less subject to random variation. Against this better degree of control must be offset the lower awareness of the "real world" as management was inc­reasingly insulated from the realities of the "sharp end", only receiving signals that were heavily filtered by the manage­ment culture that passed them up the line; an isolation which significantly contributed to IBM's later problems.

Of course in many management positions, particularly within the staff func­tions, direct measurement of performance is difficult. Charged with introduc­ing such a measurement system into IBM's Marketing Staff I eventually had to throw in the towel. There was no shortage of factors that could be measured and I was offered computer systems to measure these to the finest degree, but in the final analysis none of them were directly relevant to the most impor­tant components of the functions. A measurement and control system would simply have focussed attention on the wrong factors, to the detriment of the genuinely important ones; and paradoxically the more efficient the control system the worse would have been the negative impact. I have to report that my failure was not well received by the board, for IBM despaired of controlling such Marketing Staff which represented one of its most important, and expensive, assets.

IBM had a range of devices to get round this problem, at least in part. Like the Japanese companies it moved managers from job to job, but even faster than the average (for a company that was anyway always on the move) with manag­ers rarely staying in any, save the most senior, positions for more than two years. No doubt the main benefits of this were increased experience, wider awareness and reduced chances of management stale­ness. An incidental side benefit, however, was that they got exposure to a range of functions and to a number of contacts, with both super­iors and subordinates, and any major weaknesses were correspondingly likely to be detected; and this was reinforced by the fact that the IBM culture encouraged detection. As far as possible these moves passed key management through positions where performance could be most directly measured. In partic­ular the most important positions required, as part of the company's myth of the supremacy of the salesman, some period of experience in the field; in the most measurable functions of all. This process of management career movement was taken to its logical extreme with those managers aimed for the most senior positions; whose careers were planned down to the last detail over a decade or more. Their own progress, in the most exotic school of management training of all, even took precedence over the importance of the position they cur­rently filled; and a "guardian angel" (watching their progress from on high at Armonk) would often choose to rescue them from any potential fiasco before saving the function in question.

Exposure Testing…the best test of management ability is for it to be exposed to the largest number of different management situations, and faced with measurable targets - in the full glare of 'public scrutiny'. Those who have never been exposed to difficult situations cannot claim to have been tested; no matter how 'clean' is their record. This is another, admittedly expensive, technique rarely to be found in the management literature.

The Opinion Survey

The most rigorous test of all applied to all members of middle and senior management. As mentioned earlier, they were exposed, at least every two years to the most searching scrutiny from their subordinates, in the Opinion Survey; or at least they were until the 1990s - when, ominously, IBM's weakness led to the first postponement of this important ritual! This survey did not limit itself to the employee's views of IBM in general, it also examined in some detail their views of their management. This was indeed a very rigorous measurement since IBM employees were particularly sophisticated in their apprec­iation of management's roles, and of their shortcomings against these; and at least under the cover of the secrecy of the survey had no difficulty in registering their views. The results of the Opinion Survey were taken very seriously. Apart from obvious poor performance nothing could damage, or even destroy, a management career more effectively than a poor survey result; it was a marvel to behold how assiduously managers looked after their staff in the weeks running up to the survey.

Finally IBM ruthlessly removed managers who had, according to the Peter Principle, risen to their level of incompetence. This surgery was accepted by the culture, and by the managers themselves, since, unlike other companies, the victims were not fired but were removed to posi­tions (occasionally as high-level prof­essionals not even managers) where their real talents could continue to be used. Just occasionally this could result in peripheral departments, to which poor managers were dispatched, being less than effective. This price was, however, then seen as being signifi­cantly less than that of having to accept poorer management of the key func­tions. Occasionally the more capable of these managers sent to the "sin‑bin" were able to claw their way back, and a even a few of IBM's most senior managers survived such an experience and eventually continued their career progress.

Surveying Opinion…the most powerful management tool in IBM, and potentially just as important for other organisations, was also just about the most simple. The Opinion Survey can be easily implemented in a matter of weeks; but not merely does it log what is right and wrong about employee morale, it positively shows staff that management cares! Considering how easy they are to conduct, Opinion Surveys are rarely used by organizations.

Once again comparison with Japan is instructive. The same job security and cohesive culture was obtained by IBM, but it was, unlike Japan, not achieved at the price of an inflexible management structure based only on seniority: an inflexibility that even the Japanese eventually attempted to rectify. Once more the gain to IBM was greater flexibility, but also better caliber managers at the higher levels. Once more, however, it is not clear how IBM's later commitment to a much harsher regime damaged these strengths. My suspicion is that they, too, were fatally undermined.

Co-ordination

This still, however, left the problem posed by less well controlled junior management and the resultant problem of co-ordinating activities across the company, where IBM's official theory that this was mainly the role of manage­ment was not necessarily borne out in practice. In the event it largely depended on peer group contact at the lowest levels. All IBM'ers were then constantly in touch with a wide range of other IBM'ers; it was not just managers who were constantly in meetings. At the lowest level this contact might have been more informal, and often was by telephone rather than face to face; aided by a world-wide internal telephone network that made it just as easy to call Boston as Barcelona. Now, of course, it will be by email - as it is in most companies (where this aspect of the IBM lesson has been very readily accepted by junior staff. Though their managers too often still ask secretaries to do it for them!)

This direct communication process covered those issues of immed­iate import to the individual. It was supplemented by a "grape­vine" of great sophistication that kept all personnel in touch with key developments; it was a matter of honor that no announcement, no matter how secret it was meant to be, should come as a surprise.

During a particularly traumatic period for senior management in the UK, at the end of 1985 when an embarrassing number of the UK board members resigned or threatened to resign at a when they were needed to head up the new regions in the reorganization then under way, the grapevine was only a matter of hours (certainly less than a day) behind the board itself in learning of developments! At times it was probably ahead of the board, in terms of forecasting the likely reactions of the individuals involved!

This "grapevine" was of considerable importance as it often provided the prime perspective, sometimes the only one, within which the individual planned their tasks. It was generally accepted that management briefing would usually be inadequate; with the gaps subsequently filled, very effectively in the event, by reference to the "grapevine".

Grapevines…it is almost impossible to overestimate the power of the grapevine in organisations (especially now it is driven by emails), when change is taking place. For some reason, managements want to hide what is happening. Be aware, however, that there is no known device for hiding anything (for more than a few hours) from an active grapevine. Be prepared, instead, to use it. Use it to find out what staff are thinking, and use it to tell them (in return) what you are thinking! This is yet one more tool of management which is ignored by academia - except as a source of jokes!

Fixers…as I have already mentioned, seeded throughout the general medium of IBM personnel were what could best be described as "fixers". These were personnel who because of their length of service and consequent experience, together with a flexible and relaxed if not overtly cynical attitude to life within IBM (often backed by an almost bloody‑minded delight in beating the system), had the wealth of contacts to solve almost any problem. They became, quite informally, the focus for much of the problem solving that the system would claim is handled by management.

Although their functions were not formally recognized by the IBM structure they were informally recognized by a range of policies. Of these the most important was that of providing a career path for professionals, outside of management, which enabled them to reach levels and salaries which could be well be in excess of that of many IBM managers.

The other main element was the most direct implementation of the philosophy that was first propounded by Tom Watson Jr; of the need for "wild ducks". In the philosophy's most directly applicable form it justified support (not least continuing employment!) for the people, usually the "fixers", who management found uncomfortable to live with, but who were recognized (at least by Armonk) as essential to the well being of IBM. A weakening of their role was one of the (possibly) unintended problems induced by IBM later trying to control an aging popula­tion of bureaucrats who had achieved some seniority by length of service only. Thus IBM's eventual tightening up of the structures, to place greater emphasis on costs, typically resulted in the "fixers", who were in many ways indistinguishable to the outsider from the average bur­eaucrat, being forced out of business; with, as explained earlier, some possibly dire results for IBM.

In some areas these "fixer" functions were formalized as Programme Managers, responsible for resources and co-ordination but not for personnel; analogous in some respects to Brand Managers elsewhere. They may have had no reporting personn­el but they did have authority, and many chaired committees with members who were themselves quite senior managers. Needless to say, it is precisely these 'unproductive' personnel that IBM ultimately tried hardest to shed!

Task Forces…where change was formally required, the "fixers" role was also formalized by two devices. The first of these was the "task force"; also a favorite of Japanese management. This collected togeth­er a group of IBM'ers with as wide a range of experience as possible, to bring this exper­ience to bear most fruitfully on the plans for the expected change. The use, and occasional abuse, of task forces was widespread in IBM; as it has now become elsewhere in project-oriented high-tech companies such as Microsoft. It had the virtue of combining the range of experience in the most positive manner, since normal rules were suspended and the thinking could consequently be that much more free ranging. The abuse came about because the task force was only as effective, and its recommendations as innovative, as the participants. There was a natural tendency to use only personnel already involved in the problem, who might acc­ordingly be the least able to develop the wider perspective needed.

Biomedical Group, one of the original Independent Business Units (IBUs - IBM's first attempt at 'intrapreneurship') for example welcomed a new general manager every two years, whose urgent brief was to sort out its profitability problems. The new incumbent in best IBM tradition immediately convened a task force, and subsequently acted on their recommendations. The ultimate problem this caused was that the task force, convened every two years in this way, was always composed of exactly the same members of senior management and came each time to identical conclu­sions; and the problems were once more perpetuated not resolved. In other companies there is an added drawback - the task-force is often carefully chosen to 'independently' come up with the answer senior management wants to hear!

Task Forces…the regular use of task forces (as part of a more complex overall structure than HRS experts normally consider), to resolve critical problems, has a great many advantages. Due to the multi-disciplinary nature of the team involved, in the longer term it helps the managers build their contacts across the organisation (in the process gaining some of the similar benefits enjoyed by Japanese management). Best of all, however, it is the most powerful way of addressing the short term; since the normal rules are suspended and new ideas can emerge. Beware, though, the capture of such task forces by vested interests - their final advantage should be their clearly defined time limit! Somewhat surprisingly, task forces have a  rather poor reputation in HRS circles.

Where the task force is a short term device limited to recommending the planned actions, rather more occasionally a complete project team is set up for an extended period, perhaps for many months. Again, though, this is usua­lly multi‑disciplinary with wide experience and again most importantly the normal IBM rules were suspended to ensure that their actions were as unencumbered as possible.

Institutionalized Change

One of the problems facing any organisation, let alone one in a market‑place as volatile as that of computing, is that of failing to cope with change; of becoming locked into outdated ideas and systems. IBM (under Frank Cary) solved this problem in a par­ticularly effective manner by institutionalizing change, to the extent that the popular culture claimed that IBM stood for "I've Been Moved". As a matter of principle Armonk reorganized IBM every two years; for example, that at the end of 1985 was "regionalisation" which changed the structure from a global split by division based on product to a consolidated structure split by local reg­ion. The impact was more usually lessened by reorganizing different functions at different times, for example field and staff on alternate years; though the more ambitious changes in 1985 covered both - the first time (under John Akers) that they had been so widespread (and, arguably, the first really major mistake he personally made). Even the previously routine bi‑annual changes involved as many as 5,000 personnel in each country chang­ing their reporting structures, managers and even locations. This was so in­grained in the IBM way of life, however, that the disruption lasted only a few days, rarely for more than a week, and then everything settled down to run as smoothly as usual; just as if nothing had happened. Due to the very open structure within IBM the cost in terms of disruption was thus surprisingly minimal. It did mean, however, that the first question any IBM'er immediately asked on meeting an old colleague was "what's your job these days?"; a question that was asked as a matter of fact, and with no sense of irony. This greeting was traditionally extended to everyone during the first weeks of January each year, when IBM almost invariably indulged in at least one significant reorganization!

The benefits were, until John Akers (in 1985) pushed the system further than it could handle, quite dramatic. They allowed IBM to reshu­ffle its organisation - strengthening, and pruning, where neces­sary - whilst giving as many people as possible new experience. Most important, by throwing all the cards in the air it broke any logjams that were building up in the organisation (including resolving the "levels of incompetence" resulting from the Peter Principle) and allowed new initiatives to emerge. Armonk always was able to quote an eminently sound immediate reason for each of these changes, even in 1985, but one is bound to suspect that the means were eventually more important than the end, the process of change itself was the real objective. The end result was almost incidental, it was just one of many such alternative ends each of which would be just as effectively implemented by the ever patient IBM staff.

It is rumored that the dependence on change became so addictive that Armonk created a separate division to handle the process. This was probably an apocryphal rumor; but it encapsulated the perceived reality. Perhaps the only equivalent exposition of such on‑going commitment to continuous structural change were the various "revolutions" initiated by Mao‑Tse‑Tung which kept China in an almost permanent turmoil for four decades; strange bedfellows, the archetypal multinational capitalist and the most fervent revolutionary, both depending on the same management technique!

Institutionalized Change…as developed by Frank Cary, a regular (bi-annual) change in the corporate structure does seem to offer major benefits. As it relates to the power structures, it dislodges emerging constraints ('empires'), and offers a fertile breeding ground for new approaches. At the same time it encourages a culture which welcomes change on a wider front. On the other hand, it must be more symbolic than real and it must loosen the fabric but not recast it in any definite ideology - so that the pervading culture is  untouched. The process unleashes very powerful forces, and if these are also used (as by John Akers in 1985) to introduce strategic imperatives this may destabilize the whole organisation (as it then did IBM!). This almost anarchic process, often with no obvious strategic objectives (other than the covert one of loosening up the structure), is understandably shunned by the academics who constantly search for rational approaches!

Change then was a means by which Armonk facilitated its overall con­trol. With change already in progress Armonk merely had to steer a course; it did not have to overcome the usual inertia to any movement.

Financial Control