IBM

9093 IBM11 - PRIVATEERING WITH MICROSOFT

For once, in this book, this short chapter is specifically about Bill Gates and Microsoft. In the context of the tragic side of IBM's history which much of this book describes, he used his place as an 'illegitimate pretender' to the IBM fortune to usurp its position! That, rather than his place in the pantheon of business gods - or devils - is what really earns him his place in this tome about business giants. In view of his very different values, though, some go further to describe him as the 'anti-Christ' to IBM's previous 'snow white'. Indeed, whatever view you take, you cannot avoid the fact that the early history of Bill Gates and his company was inextricably intertwined with that of IBM (or at least of its PC Group); and maybe, if you want to pursue the Anti-Christ theory, you can argue that he has been the beneficiary of a great deal more luck than any mortal has a right to expect.

On the other hand, Gates has always had a very clear vision of what is important to an IT supplier. In his excellent book[i], which gives the most comprehensive coverage of events surrounding Microsoft's changes of direction at the end of the 1990s, David Bank nicely summarized a lengthier email from him in 1997 as: "…one idea: as hardware gets cheap software gets valuable. Software is the bottleneck. Computer software will always be in shortage because what people want to do with computers is always increasing."

Luck… teaching sales managers in IBM, I used to emphasize the four factors they would need to achieve an outstanding performance. In descending order of importance, they were:
Achievable Objectives - the first was by far the most important. It was simply the ability to persuade those setting the objectives - be they higher management or the stockholders - to expect less than might be actually achievable. If you beat low objectives (as, for a long time, Bill Gates did) you are a hero. If you fail, even marginally (as IBM very publicly did), to meet higher objectives you are on the way out!
Luck - thereafter, the next factor is luck. This is the one which is written very large on Bill Gates' CV. There is no easy way you can create this.
Hard Work - on the other hand, the saying that by hard work (and Bill Gates has dedicated almost every waking minute to his own project)  you can create your luck (or more accurately the opportunities for some luck to come your way) holds some truth.
Ability - only comes last!
Most performance management systems, however, assume the reverse, indeed that ability, coupled with hard work, is almost all that is needed!

L'Etat C'est Moi

On the other hand, unlike IBM which has long been a corporate identity, Microsoft is Bill Gates; it would be nothing without him - and that stark contrast with IBM is also part of the fascination of his role in the context of this book. So for the record, William Henry Gates III was born in 1995 - in Seattle where he still lives - to wealthy parents; with a socially-prominent, and doting, mother whose interventions played a significant role in many of the key developments in his life. He reportedly[ii] talked to her as many as three times a day even through later life. Unlike almost all other commentators, however, John Heilemann[iii] suggests it was his lawyer father who was most influential; and indeed being steeped in the law may have later caused his son to later behave unwisely at Microsoft's anti-trust trial. As a teenager, though, encouraged by his mother and possibly living up to the example of his great-grandfather who had founded Seattle's National City Bank, he undoubtedly showed a precocious intellect at an early age - not least in mathematics - and even as a teenager was seen  as a bona fide computer nerd (or geek, which he still is!). Perhaps not unexpectedly, one of his heroes as a teenager was Holden Caulfield from 'Catcher in the Rye'[iv], passages from which he was still reciting to girlfriends when he was well into his thirties[v]; though in later life his great hero seems to have become Leonardo da Vinci - perhaps because he identifies with him! On the other hand, he was a most unusual teenager - even for the money obsessed West Coast - in that he was - above all - preoccupied (indeed obsessed) with turning a profit on whatever he touched.

His mother's first defining influence, though, came when the 'Mother's Club' at his private school set up a fund to provide the students with time on a large (General Electric owned) mainframe computer and then at CCC (Computer Center Corporation) where the 'Mother's Club arranged for them to use its new DEC computer after hours - in return for searching out programming errors in its software. Spurred on by the irresistible opportunity this presented to him, he was soon became one of the early hackers invading various other computer systems; demonstrating early a disdain of the law which seems to have stayed with him - if the Anti-Trust proceedings are anything to go by. Recently, though, the only penalties he seems to have racked up are the speeding tickets he has earned around Seattle in his range of fast cars! Early on, however, he was caught by the CCC security officers when his hacking activities crashed the mainframe. He has said that he then "…swore off computers for a year and a half…I tried to be normal[vi]", though Janet Lowe[vii] (perhaps with more conviction) insists it was his mother who, worried that this distraction might cause him to fail his schoolwork, ordered him to stop! She might have had a point, because he later dropped out of Harvard to start the company which afterwards became Microsoft. We now move on therefore, to the part of his life story which was enmeshed with that of IBM.

IBM Bearing True Gifts

Clearly the most important turning point, in this context, was the initial decision which ultimately made Bill Gates a billionaire; personally worth, on paper at least, something like $6 billion by the beginning of the 1990s. More important, for the tale of IBM's woes, it handed the control of the PC operating system to him! As we have seen elsewhere, there was a degree of inevitability that PC group would look outside for an operating system. As a small IBU, with very limited resources (to match its limited targets) and impossible timescales, it had to 'buy-in' these things; and in any case the philosophy (enshrined in the then popular management theory as 'make or buy') was very much in vogue with IBM senior management. PC group was to be congratulated by them, at the time, for their initiative; and many other parts of IBM were encouraged to follow suit - though, fortunately perhaps, few did.

What may seem less understandable, even with hindsight, is why IBM chose Microsoft. The obvious choice was DR (Digital Research), led by Gary Kildall, whose existing operating system CP/M  (Control Program/Monitor) was the most popular one then on the market; and, adopting an unusually gentlemanly approach, even Bill Gates (who had an agreement with Kildall to keep out of his market[viii]) had pointed IBM in this direction. Its president, one of Silicon Valley's 'accidental' entrepreneurs (who, typically for the time, had originally called the company 'Intergalactic Digital Research[ix]), Unfortunately though, according to legend, Kildall blew his chance of becoming a billionaire when he played so hard to get that he wasn't even there when IBM first called (though the managers from PC Group had an appointment). According to some reports, in a tale which has grown with every retelling but in this case supposedly originating with Bill Gates[x] himself, having had a (very expensive as it turned out) contretemps with his wife, he was out for a spin in his plane. In fact, he was actually away on a previously booked business trip. In his absence, his wife - the who actually ran the business end of things - initially refused to sign the standard (but legally complex) non-disclosure agreement IBM managers placed in front of her; which - weighed down by IBM's legal baggage - they had to get signed before any substantial negotiations could take place. Despite the apocryphal stories, though, she did eventually sign[xi], albeit perhaps not fast enough for IBM. Some commentators say, again incorrectly, that her husband wasn't even contactable later or that he was just as offended by the non-disclosure agreement.

More important, most say he wouldn't change the CP/M operating system to handle the 16 bits the IBM PC needed to use- or perhaps more believably wouldn't do this fast enough to meet IBM's near impossible deadlines. Whatever the reason, Gary Kildall ruled himself out of the main race - though (after threatening legal action over the similarities between DOS and CP/M) he was later allowed to put CP/M on the new IBM machine; but foolishly chose to price it at $240 per copy against Microsoft's $40[xii] - and that was the final nail in its coffin! So, in one of the lucky chances which have so characterized Gates' career, Kildall handed the resulting billions to Bill Gates - who by then no longer felt he had to live up to his previous agreement with Kildall!

Gross Failures…whilst few individuals, or organizations, achieve outstanding results, the one way you can be sure to beat the odds, in terms of achieving anything other than an average performance, is to fail; badly! In particular, the organizations which stand out from the crowd are typically those which have taken such bad decisions that they have in effect destroyed themselves. Much is suggested about how to succeed in business, but little is said about avoiding failure.

Even so, Bill Gates, opportunistic bright guy that he was (and still is), had an almost non-existent organisation, with a very limited track record; and did not even have an operating system to offer! It has been suggested by some commentators that the key to this was that John Opel sat on the same board, of the charity United Way, as Bill Gates' mother, Mary. I personally doubt this, since Opel wasn't involved at the time, but Janet Lowe[xiii] insists this was the case. The regular face to face meetings with his mother must, though, have later made it even more difficult for John Akers, who had then taken over his predecessor's place on the board, to act against a supplier (her son, Bill Gates) - and in the process against IBM's own philosophies - in any really aggressive way. I cannot believe, however, that this would have had any impact earlier, at the time the project started; especially with a group which was about as remote from Armonk as anyone could get.

Based on my own experiences of similar situations within IBM, my own suspicions are that the IBM personnel found Bill Gates a useful front man for taking actions which the IBM system would not have let them take. Had they tried to write their own software it would have been bogged down in quality control procedures for years. Had they tried to buy outright some software, and it then became IBM's, it too would have been bogged down with IBM's Business Practices staff and in risk-control procedures for months. Bill Gates took all these problems away from the group; and, precisely because his was such a small organisation, the IBMers probably thought they still had some degree of control (though, of course as events were to prove, they didn't!). I myself from time to time used similar outsiders to front work for IBM which would have been impossible to get through the system otherwise.

Regulating Bootleggers…organizations often come to depend upon 'illicit' development operations in order to bypass their bureaucracies. Such approaches may be essential to future development. They can, though, subvert the whole organisation - so senior management needs to develop the special skills (in particular those of balancing freedom against risk control) to manage such operations. Apart from the work of Peters & Waterman, which refers to 'skunkworks' management, theory pays little heed to those who operate on the borderline of legitimacy (and none to control over their potential subversion).

In terms of why Armonk did not see the risks, Paul Carroll (in Big Blues) says, of the key meeting of the Management Committee (MC) which approved the setting up of the group, that"...the Management Committee continued to see this review as just one of the numerous proposals that came in front of them each week. Lowe and his team thought of the PC as a home computer and continually referred to it that way, so it hardly seemed as if would be of any interest to IBM's corporate customers." In other words the MC approved a home computer not the Personal (Business) Computer it later became.  He goes on to say of the other key issue; "There was little discussion about whether using outsider's parts would mean IBM couldn't control the direction the market would take, the way that IBM controlled everything about the lucrative mainframe market. But nobody in the room had the foresight to realise how important the personal-computer market would become or how little control IBM would, in fact, have. The MC members worried more about the possibility that outsiders would be allowed to sell the PC, because IBM wasn't sure it could make then live up to its blue-suited white-shirted standard. Even that concern didn't really heat up until later." As it was a home computer, aimed at a small market (and with only a small investment at stake) did it matter? You bet it did, but - given the information in front of them - there was no way they could have known that. A much better question is why didn't they do something later when they did find out what was happening? Des Dearlove[xiv] says of the deal "The effect was like introducing a fat and sleepy buffalo to a piranha."

Internal Weak Signals…the problem of detecting weak signals in the external environment is well understood - though no solution is generally available - but that of recognizing the equivalent signs of impending 'fractures' within the organization is not. These may represent an even greater a threat to the future of the organisation; and the solution is available - all it needs is a management information system which reports on them! Most management theory does not, however, even allow for such internal fractures.

At the time Microsoft's business was based on porting mainframe languages to minis, and its main product was its BASIC interpreter for these. This was reportedly[xv] developed (setting another precedent) on the back of hundred of hours of illicit time spent on the Harvard computer when he was studying there (and against the rules allowing an outsider to work on the same commercial development, for which - again demonstrating his unbelievable luck - he only got a slap on the wrist). So when Bill Gates and his team were also visited by IBM - they assumed that this was what they wanted. Unlike Digital Research, they happily signed the non-disclosure agreement and, uncharacteristically, were very deferential to IBM - a posture they adopted with IBM (but nobody else!) until the deal finally fell apart! So it is reasonable to imagine Akers agonizing over how he could 'do the dirty' on Mrs. Gates' nice young son when, paradoxically, Bill Gates himself would never had any qualms had the situation been reversed!. Indeed, when they first met him the IBM team thought Gates was the office boy (an image he seemingly cultivated since his 'adversaries, to their disadvantage, then consistently underrated him) -

Perhaps setting the precedent for the business practices to come, the BASIC work had already been the subject of a lawsuit about ownership with Pertec; though in that case, for once, the arbitrator admonished Pertec for its business practices not Microsoft.

Opportunism…Bill Gates' greatest skill is that of applied opportunism; of quickly recognizing potential opportunities and then getting behind them. This ability is in part a gift, but it can be developed by opening the mind to the whole range of opportunities. Dignified by the umbrella term of  'creative management', a range of techniques have been developed by some academics.

The IBM team were, in the meantime, reporting that despite his lack of resources (but presumably impressed by his deferential demeanor) they felt they could work with Bill Gates; an off-the-cuff judgement which later came back to haunt IBM.

QDOS

Thus, when the IBM team returned, they offered Gates the BASIC software business but - much more important - they mentioned, almost as an aside, their need for an operating system. Gates, the supreme opportunist, jumped at the chance to bid; where, to give him his due, most of us would not have had the courage to do so. There are a number of semi-apocryphal stories about what happened next. One suggests that Gates was already friendly with Rod Brook, the owner of Seattle Computer Products, who was developing a 16-bit operating system called SCP-DOS (Seattle Computer Products Disk Operating System) or, according to a number of other sources, called QDOS (rather ominously standing for 'Quick and Dirty Operating System'!).Whatever the real story, Bill Gates was left to find an operating system and eventually located one.

The most important aspect though, according to Paul Carroll, was that Bill Gates' contact in IBM - Jack Sams - told him to purchase the operating system; again as I probably would have done, to get round IBM's internal systems. Gates paid $75,000 for it - $25,000 initially, for the licensing rights, and then $50,000 for the exclusive rights, in his own name - and nobody can deny that, in terms of the billions he subsequently made from it, it was a bargain! Soon afterwards (again according to Carroll) Sams, who - perhaps understandably - may have wanted Microsoft to remain viable so his own operation within IBM would not fail, persuaded Gates to take a tougher negotiating line with IBM. It was a lesson Gates learned well, and later followed to IBM's great disadvantage! Arguably, as you follow the story of Microsoft, Bill Gates strength was indeed that he was an outstanding negotiator; and it was a skill first taught him (at least in part) by IBM. He was also lucky. QDOS borrowed somewhat from CP/M, and just before the launch of the PC DR threatened a lawsuit on this basis; but IBM bought them off by also putting DP/M on the machine alongside (IBM) DOS (Disk Operating System) as Microsoft's product was then called (the version which later powered the clones was, for obvious reasons named rather differently - as PC DOS). It was DR's poor judgment, and again Microsoft's good luck, that they botched their whole marketing effort (and priced their system at six times as much as DOS).

Confident Negotiation…probably the main 'skill' needed in negotiation is confidence. On the one hand, if you have nothing to lose or, on the other, if your experience leads you to expect a good outcome, your confidence will usually make you an excellent negotiator. The nervous negotiator, who is uncertain of the outcome, will almost always lose. Negotiating skills are, rather questionably, usually the preserve of specialist - very commercial - trainers.

He was even luckier that IBM did not insist on an exclusive licence; though, paradoxically, Gates had been - in 1976 - one of the first to vehemently object to the pirating of software; claiming that his BASIC had taken up $40,000 of machine time - which may have been true, but he failed to mention it was from Harvard that he illicitly obtained this[xvi]! At this time, courtesy of IBM, he was set free to sell this new software to other hardware suppliers. At the time, when IBM held such a dominant position, this probably did not seem relevant; but later, when the position was reversed and the clones dominated the market - using Microsoft's PC DOS - it clearly was seen to be a historic mistake. Thus, though he was only paid a one-off fee of $200,000 by IBM for its own PC DOS, he was almost literally given a licence to print money with his own version; MS DOS.

IBM Loses Its Bottle

Above all, though, it was Bill Gates' abilities as a negotiator, and the lack of ability of his IBM opposites which determined the day. According to Paul Carroll, at the first set of key meetings, in 1985, when IBM might have re-negotiated the deal - since IBM was still in an immensely strong position (and indeed was eventually offered - in an unusually conciliatory gesture in 1986 - a 10 per cent share of Microsoft, which it turned down!) -  it was William Lowe (who had by then taken over, again, from Don Estridge) and his boss Mike Armstrong who were indecisive. Carroll goes on to say "That indecisive quality can leave IBM vulnerable - as it did repeatedly in Lowe's and Armstrong's dealings with Gates." Whatever the reason, at this critical time, and over the next few years, Bill Gates regularly negotiated himself out of having to make a less favorable deal with IBM - and even from having to face competition from it. For a decade, somehow or other, he kept IBM off his back - and set the basis for his billions! Incredibly, considering the constant in-fighting, the relationship lasted until 1992; though over most of that time Gates already had a team planning what Microsoft needed to do if IBM ditched it!

Indecision…almost worse than taking the wrong decision may be not taking the decision at all! Management theory does insist, perhaps at times with rather too much machismo, that decisions must be taken promptly.

Windows

The crucial development was, however, that of the Windows - which was eventually launched as genuinely workable operating system (as version 3.0) in 1990. Although IBM apparently conceived the idea of its equivalent OS/2 in 1984, Gates - as always ahead of the game - reportedly had become interested when, at the November 1982 COMDEX, he had seen the VisionOn graphical interface launched by VisiCorp; then one of his main rivals in applications software, with revenues on the back of its VisiCalc spreadsheet which were then double those of Microsoft. In typical style, Gates reacted to this - as described by Edstrom & Eller, the latter having been a key member of Gates' programming team for thirteen years [xvii] - like "Saul on the road to Damascus". He saw both the threat, VisiCorp could set the standard (as it seemingly had already decided to do), and the matching opportunity. He immediately set his best programmers working on it, two years before IBM even started to consider anything like it; and eight years before Microsoft's finally successful version appeared; so the other lesson perhaps is that even Microsoft's best programmer's can't always deliver next week!

In another typical move he had purchased a $15,000 Xerox PARC Star system; so his team could 'reverse engineer' (copy the workings of its software, though not the exact code - which is the part that is always copyrighted) its most innovative features - especially its use of icons, which is now such a familiar part of Windows. Again using whatever (even dubious) means at his disposal, Gates eventually hired Scott McGregor, who had written Xerox's windowing system[xviii], to lead his own graphics group.

With Dan McCabe duplicating VisiOn's windows and Rao Ramala working on the icons, a (demonstration) prototype was ready by April 1983; though this only demonstrated the concepts and was a long way (eight years) from a working system. Even so, the Interface Manager Group of just ten people - whose only strategic advice from Gates (once more typically) had been to 'squash VisiOn' - was in chaos until new management, including McGregor, was put in place. The key to success, however, was not writing  the software but persuading the software and hardware vendors (playing each off against the other) to write the drivers which allowed their products to interface with Windows - an investment by them which has since become a barrier-to-entry protecting Microsoft's monopoly position.

According to Martin Eller[xix], Gates was fixated upon IBM; even though it was the only major PC vendor not to support Windows. His near obsession was, though, perhaps understandable when its support had been so important in the creation of his fortune - and, at that time, it was of course still the giant which dominated the industry (whose ire it was not wise to earn!).

But meeting IBM's needs frequently seemed to get in the way of the rather chaotic development of Windows. Thus, when IBM announced an, ultimately unsuccessful, upgrade to DOS but also a competitor to Windows - TopView - Gates wanted to support this; despite the fact that meant support for DOS, which the developers were having difficulty maintaining in the new operating system. Eller makes a quite fundamental observation, though an inevitably biased one, when he suggests[xx] that at the time "Gates had never been involved in any of the architectural design of Windows, nor had he ever been personally involved in writing such large amounts of code…and he seemed totally oblivious to the fact that every feature change radically screwed up Window's stability, testing, and ship date." In fact even now, despite more than 25 years working in software Gates has held just one patent. The result was a lack of strategic direction for a year. When, in 1984, the delivery date still was a year over the horizon, and the size of the software meant that it would only run on a PC with a hard disc, McGregor left and with him went much of what technical leadership there might have been.

Even when, in November 1985, Windows 1.0 did finally make it out of the door, it didn't support either Mac or TopView applications and, once more according to Eller[xxi], its DOS compatibility was a joke. He summarizes the position as "In short, the product was essentially useless"; so maybe that is why Microsoft seemed to want to forget about it! Gates left only a three man team in place keep the project alive in theory, but it had clearly gone to the bottom of his priority list where (surprisingly in view of its phenomenal success later) it remained until the early 1990s; when the sale of a million copies of Windows 3.0 eventually awakened Gates to its potential. In the meantime Gates once more focused on IBM, and OS/2.

The main problem emerging from the joint work with the IBM development team was not just the difference in dress and culture but, much more importantly in programming standards. Having worked with the IBM PC development labs at Hursley myself, I am aware of just how much of their time is spent on debugging their programmes - and, in particular testing them - so that when they go out the door they are near perfect. Microsoft's programmers, pushed by Bill Gates, were more concerned that their work went out fast; and testing (or even thorough debugging) was a luxury they would happily dispense with. Indeed, Gates is obsessed with the perishability of software, as reported by Randall Stross' book[xxii] 'The Microsoft Way',  he has said "…there's not a single line of code here today that will have value, say, in four or five years time". That is why their later discovery of beta-testing, whereby the testing could be passed to customers who would even pay for the privilege, came as a nicely cost-effective development. But, needless to say, their cavalier attitude did not impress the IBM programmers; even though the resulting speed of development was why IBM was so keen to use them to get round its own bureaucratic delays! IBM needed them, and they still needed IBM's money and prestige.

Beta Testing…using customers to test, and even develop, new 'products' is very productive for the vendor, and well received by the customers (as it leads to an interactive dialogue). In one variation on the approach, the Japanese vendors launch many incrementally new products and wait to see which is the fittest - and survives. This was very different to the traditional Western approach which looks for radical developments; though the high-tech industries have now widely adopted beta-testing (though perhaps without really understanding what is involved!).

Because IBM wouldn't build support for Windows into the TopView software they were also developing, the Microsoft team were looking for alternatives. The one they found was a TopView clone called Mondrian, developed (yet again) by a small company, Dynamical Systems Research (DSR) owned by a group of Princeton Physicists. Microsoft bought the company for a stock swap (not even cash) valued at around $1.5 million[xxiii]. In the process, in a typical move, gates also 'bought' Nathan Myhrvold, who became his most influential technical advisor in the early 1990s.

Eventually, in 1986, to the delight of Gates IBM did indicate that it might support a Windows type interface; though it still did not go as far as doing anything about it! As a result of Gates betting his future on OS/2, by the end of 1986 Windows was seen - within Microsoft at least - as an orphaned bit of code, though the application developers continued to use it as a useful model so they would be ready to port their work to OS/2  (which Gates, perhaps rather foolishly, thought would be compatible) when it came. By then, however, the IBM OS/2 team had now been split, to include IBM's Hursley lab in the UK - with whom, by coincidence, I was working at the time. As described by the various commentators, the eight-strong Microsoft team may have been ensconced at Hursley for three weeks in the luxurious mansion there - annoying the local IBMers by playing Frisbee on the extensive lawns - but IBM's own dozens-strong programming team was (perhaps more significantly) holed up on the top floor of an ordinary office building in Basingstoke new town - ten miles a way.

Indeed, IBM and Microsoft did eventually launch, jointly, OS/2 Presentation Manager - two years after the IBM statement - at the end of October 1988. Significantly, signaling his enthusiastic support at the time for OS/2 rather than Windows, Gates is quoted[xxiv] as saying "This is the milestone that changes the rules for everybody…OS/2 Presentation Manager will be the environment for office computing in the 1990s." I guess that this is one time IBM now wish his proud boast had come true!

His partner, Steve Ballmer, was even more explicit "This is it, after this we're not going to have any more Windows. It's all OS/2" Indeed, the next version of Windows, version 2.01 which was full of bugs (and hence rapidly led to 2.02 and 2.03), was specifically launched only to support the new Compaq 386. A couple of programmers were left to maintain it, but that was to have been - with it becoming nothing more than an interim development platform for OS/2 - to all intents and purposes the end of Windows.

Steve Ballmer, a friend of Gates at Harvard, was persuaded to join Microsoft in 1980 - to, in effect, thereafter to become one of its founding fathers. Even so, he was seen as a weak strategist for a technology company. According to David Bank[xxv]at the time gates told his aides that that there was little chance Ballmer would ever become CEO. Of course, at the end of the 21990s Ballmer did eventually displace Gates as CEO; though who actually controls Microsoft is still an open question! In particular, at the earlier time, though, he very effectively reorganized Microsoft to get the right people in the right places. In addition, forced to become a reluctant head of sales and marketing, he was the closest of the founding partners to the needs of its customers.

Skunkworks

Unfortunately for IBM, Dave Weise - who couldn’t cope with working alongside IBM - argued his way back to the effectively defunct group. But this left him, with time on his hands, to pursue a pet project. The key to the future, for all operating systems, was to be able to run in 'protected mode' such that each application program could run alongside another but protected from it; its own code would not be corrupted by any others running on the machine. Around this time, Ballmer had wanted to buy Scroll Screen Tracer (SST) - a debugger developed by Murray Sargent, a brilliant physics professor from the University of Arizona - for use on its DOS work. Sargent, who was then on contract at Microsoft was persuaded to sell it, and spend the summer of !988 training Microsoft staff to use it and moving its own debugger into protect mode. This was when Weise met him, at a Microsoft party. Without any approval from Microsoft management, in the most celebrated form of skunkworks, the two of them - using Sargent's debugger - spent the next two weeks seeing how much of Windows they could hack into protect mode. In such unpredictable ways are corporate futures changed; without this one-man mission, the PC world would still be colored (OS/2) blue!

Skunkworks…many of the most important breakthroughs come from unexpected sources. As already suggested, the key management ability is that of recognizing their importance. Again, those running creative management courses teach helpful techniques.

The result was that, at the August 1988 planning meeting, Weise (having wisely, forgive the pun, brought Ballmer in on the secret a few days earlier) sprang on Gates - along with the then fifteen man Windows team - not just the news that he had cracked the protect mode problem but actually demonstrated a working version of Windows using it. Even though he still thought OS/2 was the going to be the industry leader, Gates gave the go ahead for the new Windows project, and the legendary version 3.0 was on its way to launch twenty-one months later.

Product Champions…the most important breakthroughs are often achieved by individuals who are so involved, indeed obsessed, that they will overcome all the odds - sometimes even going as far as to subvert the organization - to gain success for their 'baby'. Academics recognize the role, which is akin to that of 'change agent' in the wider environment, but have few suggestions as to how they can be encouraged; and traditional managers find a host of ways to actively discourage them!

Even though Gates thought it wouldn't harm the relationship with IBM, when the word finally leaked out to IBM, who quite reasonably thought they had eradicated Windows, all hell broke loose and they told Microsoft to kill the project. Accordingly, at the 1989 COMDEX Gates stood up and promised that - after the planned 1991 release of Windows 3.0 - Microsoft would cease development of it.; leaving it for the low end of the market and OS/2 being the main PC operating system.

Windows 3.0

The reality was that, launched in 1990, Windows 3.0 was an instant success; selling two million copies in the first six months. As a result it took on a life of its own; and because of the work done by Microsoft's application groups - who were using it to test the move to OS/2 - their offerings (Word, Excel etc.) also walked over their competitors! Not least, Ashton-Tate's 'Word-Perfect' word processing package, which - despite all Gates' efforts to promote his own Word (whose disastrously bug-laden first version was still remembered) - had been the brand leader through the 1980s, moved too slowly and was soon overtaken; as eventually was Lotus in the spreadsheet market. Indeed, by 1998, the 'Office' desktop applications grossed more than $6 billion, compared with rather less than $5 billion for the Windows desktop platform[xxvi]. They had become the two best selling software products in history.

In just two years Microsoft's stock-market value grew by 305% - reaching more than $20 billion by the end of 1991. According to Eller[xxvii], those inside Microsoft knew just what a fluke it had been, but the media - fed a diet by the publicists that insisted Microsoft was Gates - couldn't see it that way; and the legend was ratcheted up a notch. In turn his competitors believed it was a strategy born of his devilish cunning; all along he had planned to ditch OS/2 and give his own applications programs an unfair advantage. I tend to believe, like Eller, that - at this stage at least - it was an incredibly profitable cock-up not a well-judged conspiracy! It was only when Gates saw Windows 3.0 sell a million copies that he finally accepted - even then seeming to still want to follow IBM - it was a strategic product.

Cultural Incompatibility…alliances often demand that shared taskforces are formed. If the two cultures are as different as those of Microsoft and IBM the result may be counter-productive. Again, those running creative management courses address this problem.

Windows NT…mind you, even then he almost immediately switched his allegiance to Windows NT, which was rumored[xxviii] to have been started as a means of attracting one of DECs VMS architects; who wanted to continue working on mini-computer software. Based on the DEC architecture, unlike the mainstream Windows which still used  large amounts of Intel 16-bit assembler code, NT was to be built from the ground up; to run on the whole range of machines  (from PCs to mainframes, thus starting to deliver Gates' dream of global domination). In the new range of products, his own PC based Windows was (once more) to be maintained, OS/2 was still to be the major new operating system for the PC, and NT (the key product) was to be the portable (between machines) platform of the future. Unfortunately, when - at the beginning of 1991 - the NT team presented their work, IBM noticed that it was all windows and no OS/2; and the long engagement was over before any children had been delivered. The loss of OS/2 also damaged Microsoft's networking and language developments, which had been based on it, throwing away several years of development work in the process.

Even though it accounted for only a tenth of Microsoft's 'platform' business, Windows NT was to be Bill Gates' passport to the even more profitable business to be had from larger systems. Accordingly, throughout the 1990s, he threw resources at it. The first version, Windows NT, contained 6 million lines of code and needed a team of 250 developers at a cost of $150 million[xxix]. For later versions, the team expanded to more than 5,000 staff, producing nearly 30 million lines of code. Bill Gates eventually abandoned the 'Cairo' version, which was supposed to become his 'Windows Everywhere' and take over the world, to produce yet another version (which eventually was launched as Windows 2000) that was incompatible with the mainstream platform - but Bill Gates[xxx] proudly announced that it had  cost of more than $3 billion, more than the Boeing 747 cost to develop!

James Allchin, who was the head of the project, had been, once more in a typical manner, been recruited by Gates when he was considering buying Banyan, a specialist in networking,  but realized that it was better - and far cheaper - to simply hire Allchin away from them instead! In a comment redolent of the IBM 360 gamble, he said "We're obviously betting the company on it". At the time, in terms of Microsoft's vicious internal politics, he was fighting for supremacy over the mainstream windows which he thought was a 'nice toy'[xxxi].Time will tell if his hype was well founded, and it really is as profitable an investment as the 360 was; or more like Concorde, a price paid for the developers ego!

The Concorde Effect…the best developers are those who are product champions, dedicated to - indeed almost obsessed by - their new 'product'. With their blinkers on, they are however the worst people to decide on its value in the market-place!

Windows 95…the new stand-alone version, which was eventually called Windows 95, was to be a very different affair. As it was now to become the star product, the development team grew until it eventually ran to over 500 names, matching the size of project team IBM had been wont to deploy. Making it a 32-bit system once more created major delays, though this time they had a sound alternative (Windows 3.0, with 140 million users) already in the field. It also signed up to strict protocols on debugging and testing; though this was the point at which they discovered the real magic of beta-testing, Some 400,000 users (adding up to something like 75 million hours of testing) signed up to use the code (laden, as it turned out, with 6,000 bugs) in advance of first delivery and paid $30 a copy for the privilege!

In fact, 1990 seems to have been something of a turning point for Microsoft as a whole. Before that time its history had depended on a great deal of luck; as indeed was the unexpected success of Windows 3.0. Thereafter it turned itself into a much more conventional organization; much less dependent on chance for its success. In particular, Nathan Myhrvold was put in charge of its research and development work[xxxii]. As we saw above, he recruited literally hundreds of new developers to work on Windows, in order to rectify Microsoft's previous tendency to ship bug-laden first and second versions - only getting the product right the third time around!.

Macintosh

At that earlier time, Microsoft had also been working with Apple to port its applications to the new GUI (Graphical User Interface) machine, the Macintosh. In view of the obvious conflict of interest, Gates had been forced to create a 'Chinese Wall' between the two sets of developers; though it proved to be a very leaky one - not least because its senior managers, who knew all about the Mac, straddled the two (and were captivated by the Mac). The day the Mac was introduced, in January 1984, Gates reportedly told MacGregor to go out and buy one so it could be reverse-engineered. He did not see any conflict of interest, and did not apparently[xxxiii] want this to squash Apple, as he had done VisiCorp. It is claimed that he merely wanted to ensure the Microsoft applications running on the IBM PC were brought into line with those on the Apple Mac; for efficiencies sake It is a nice story, but it would be the first, and last, time Gates indulged in such niceties!

In any case, apparently without understanding how the Mac system worked[xxxiv] (a strange lack given his passion for software), and in particular that it was incompatible with their own windows code, Gates told the Windows team to also incorporate the Mac features; which caused yet another massive delay - said to be more than seven months.

Billionaire

Bill Gates, as we all now know, went on to become the world's wealthiest individual - becoming a billionaire as early as 1986 when he was just 31 years old - and in the process created between 1,000 and 3,000 employee millionaire employees (depending upon which estimate you take), though that number may have significantly diminished after the dot.com crash which cut Wall Street's valuation of Microsoft by a third. He has been, with a degree of self-interest, generous to his key employees in this way.

Even so, Microsoft's management style has been described as Darwinian - survival of the fittest[xxxv]. Bill Gates seems to thrive on conflict, as much within Microsoft as outside it. He is, again according to Janet Lowe[xxxvi], "…known for being confrontational, rude and condescending"; characteristics which did not help his case (or his image in the media) when he appeared before the judge in the Anti-Trust trial at the end of the 1990s. In particular, he has a low tolerance for those whose intellect he does not respect. When he throws a tantrum his favourite phrase seems to be "…this is the stupidest thing I've ever heard of…"[xxxvii] At such times he is hyperactive and literally screams at his staff, before quietening down and rocking rhythmically back and forth in his chair! On the other hand, as (reported by Janet Lowe again[xxxviii]) one developer said "…you have to understand we're all adolescents. And proud of it." But he (perhaps helped by his billions) is still seen as a charismatic leader who inspires his dedicated workers.

Much of the good news is, though, hype put out by his spin doctors. Thus, in particular, his widely reported donations to charity are often rather less than they may seem. In total up to 1997, Gates gave away just $70 million; less than a third of one per cent of his total net worth. Microsoft's subsequent 1997 claim that it had donated $73.2 million in cash and software was revealed to be 85% software - valued at retail prices (even though the cost to Microsoft was probably less than one per cent of this) - and even this (given to a school or library) was best seen as promotional support rather than charity[xxxix]! A number of the subsequent 'donations' are also best viewed as promotional activities. Gates' excuse is that "I don't have time to figure out what charities make sense"[xl] - and presumably he also doesn't have time either to appoint an adviser who would have the time!

Underdog

For years, during its battle with IBM, Microsoft was seen as the underdog It even carried this underdog image with it when it became the market leader. It was only at the end of 1997, having already been in court since 1996, when the US Justice Department's prosecution case really started, that its reputation turned around and it became the target for the near hate which now sometimes envelops it! The fact that it had brazenly defied the previous order placed on it in 1995, requiring it to offer its Internet Explorer separately, clearly did not help.

More important, for its competitors, Microsoft is determined to keep proprietary control - under the law - of all IT standards, and has done so far more effectively than IBM ever did. Its methods have sometimes been, however, questionable. For example, having signed up to adopt the Sun Microsystems Java language for use on its Windows and NT systems, the version it actually released was not compatible with the original standard; and once more fought it out in the courts. It has also been criticised for releasing substandard products just to get to market before its competitors; but, according to Des Dearlove[xli], Gates has said that "…getting to market with a good or OK product first is better than getting there second with a great product."

Control of Standards…if you can gain control of the standards which operate in a market, and especially if you can achieve something of a monopoly as IBM once did and more recently Microsoft has done, you can shape the demands of that market to favor your own offerings. Academics, especially economists, encounter difficulty in dealing with markets which experience imperfect competition - as most do!

Strategy

As - in the later 1990s - the markets for operating systems moved into new areas, especially in consumer goods such as digital televisions, Bill Gates promoted his 'vision' of 'Widows Everywhere': with compatibility across the whole range of devices. This was to be developed by Allchin's group working on Windows NT, by then the ultimately aborted Cairo project which finally become Windows 2000. This represented, in Bill Gates' eyes, the future: allowing him to enter the profitable large systems market. Ranged against them, however, were Silverberg's team working on Chicago; which in turn would become Windows 95 - consolidating Microsoft's hold on its core business with end-users. Silverberg was very market-oriented, giving users what they wanted; and his staff, largely made up of graduate recruits, saw themselves as 'pirates' dedicated to creating 'cool' new features[xlii]!

Earlier in Microsoft's history this competition between groups might have proved productive, in much the same way that IBM at its peak funded competing developments and then chose the best of them. But, in the 1990s as it matured to become a large bureaucracy, the splits between the two groups started to cause major tensions within the company. They didn't communicate well with each other, not least because Silverberg saw the job of the NT team as being 'to kill Windows 95'[xliii]; which was, indeed, its hidden agenda!

In the later 1990s an even more crucial split appeared. This was between the Windows developers in general, who were at the heart of the company, and those who wanted to respond to the new - rapidly growing - demands of the Internet. Bill Gates was, of course, amongst the former; since his 'Windows' had spent most of the last decade building the 'lock' on its users that he saw as the source of Microsoft's massive profits. The users were happy using it and - most important - they found it very difficult to give it up and switch to something else! It was the 'hard-drug' of the software industry, where persuading its users to undergo any 'cold turkey' was an almost impossible task for Microsoft's competitors. Accordingly, over the years - despite the ease of maintenance which would be the result - Microsoft had never simplified its code and had added on extra new features to strength its lock on users.

Cold Turkey…if you can build in a high level of migration costs, so that customers think it is not worth the costs of changing to achieve marginally better results and they are not too annoyed by poor service, your 'product' can build a very stable set of customers; and, if you are the brand leader, build towards something very like a monopoly!

Unfortunately, the Internet was essentially simple; and was, as such, seen by Bill Gates as a major threat to his security blanket of 'lock-ins'. But the Internet was already attracting the attention of the many 'independent' software providers on whom Microsoft relied for writing the applications which its customers used.

It was at this stage, therefore, the Bill Gates finally got interested - especially in Netscape which was rapidly becoming everyone's favourite. Unable to buy that company, his usual solution, Bill Gates response inevitably was to order his developers to reverse-engineer it and then add unique features an content  to build the lock-in again.

The hidden effect, which ultimately became more important, was that, for once, Bill Gates was forced to embrace open-standards; those of the Internet. Of course, it would also extend them, to - as its competitors claimed - 'engulf and devour' them! Even the Windows desktop was to become a web-page.

It also highlighted another major split within Microsoft. By then its developers had begun to feel that their new ideas were being pushed aside in the name of 'strategy'. Bill Gates had been demanding ever tighter lock-ins - to push users towards its own applications. Bosworth voiced these frustrations[xliv]: "Before 1996 the entrepreneurs ruled Microsoft…[it] was filled with people who had to compete against entrenched winners and won…The organization was built for that". But, he added, the new bureaucratic limitations - demanded by the 'profit maximization' model Bill Gates had created - was stifling all that and demoralizing staff.

The main argument, however, then focused on whether Microsoft should be developing a (separate) new Internet platform or integrating it with Windows. Unfortunately the legal arguments in the anti-trust case meant that Bill Gates had state that integration was a necessity; and the die was cast!

Up until then the board of Microsoft had always deferred to Gates, for he had been the company's single focus of control. Now, as they saw it, this 'focus' was also becoming its single point of failure. Windows 2000 was losing its battle against complexity and other products were being delayed. According to David Bank[xlv], based on the email evidence uncovered in the anti-trust trial, Gates was losing control! In 1998, therefore, he handed business responsibility to Ballmer - who he had once said would never become CEO - and then in January finally allowed him to have the full title of CEO.

Again according to David Banks[xlvi], Bill Gates' explanation for moving to 'Chief Software Architect', that he wanted to spend more time with its product groups, caused alarm amongst these! They feared he would disrupt their efforts, where the previous advice from their functional management had been to keep Gates as far away from their work as possible.

In the event, Ballmer proved just as centrist as gates, demanding to be the focus of decision-making one every issue; a major problem where he, like Bill Gates before him, managed by controlling the detail rather than by setting strategic objectives. In any case, Bill Gates - now supposedly thinking two to five years ahead - is still the one real central focus in Microsoft, and is still seen as its 'brand image' (albeit an increasingly unattractive one) to the outside world.

The main problem for Microsoft may now be that it has lost its intense focus on Windows, which had driven it for the previous decade. As set out for employees in 2001, Ballmer stated that there were to be seven key businesses (including .NET, Office, Server Software, MSN), two priorities for staff (keeping customers and delivering the new platform) and three things the company must do well (meet priorities, cut costs and focus on its people). Microsoft was starting to look like IBM in the worst days of John Aker's rule!

Visionary?

Gates' spin-doctors plug him as being a great visionary. But he didn't invent DOS, it was IBM who brought him to that, and his most successful product Windows, the key elements of which were derived from the technological breakthroughs of others, has succeeded despite his decades long attempts to kill it. In many respects, therefore, Gates is best characterized as the luckiest man alive or at least the world's most successful opportunist.

Thus, displaying a peculiar, and continuing, blindness over the future of IT communications, he missed the development of LANs (Local Area Networks) and lost software leadership there to Novell (something that, despite his public bad-mouthing of the Novell software) he once thought was his greatest mistake[xlvii] and has twice tried to buy [xlviii].

Indeed, despite all the hype, he has relatively poor strategic vision. Although his much hyped 1995 book about the future (which talked more about his new house and his passion for encryption than what faces the rest of us as a result of his machinations), mentioned the Internet (which some of the commentators have mistakenly said it didn't), Gates predicted that this would only be a minor part of the overall 'superhighway', which was the focus of his attention at the time; "The growth of the Internet over the past few years suggests that highway applications will quickly become extremely popular and justify large investments…It's a long way from being the highway, but it's the closest approximation we have today, and will evolve into the highway"[xlix]. How embarrassingly short-sighted can a futurist be!

In fact, as usual for Gates, his interest in the whole field of electronic networking had been stimulated by the arrival of what he saw as a major new competitor: AOL. The result was the development by Microsoft, from early 1993 onwards, of an on-line system, Marvel, which soon became Microsoft Network and then just MSN. It was a closed system, however, mimicking AOL, rather than peer-to-peer as the Internet was - and as Myhrvold was wisely recommending at the time.

AOL

At this point I will indulge in another diversion, into that part of the field of Business-To-Consumer (B2C), and especially Consumer-To-Consumer (C2C) networking, now dominated by AOL. It is a diversion, in the context of this book, since it is now not directly relevant to Microsoft, or even IBM which - apart from its disastrous adventure with Prodigy (its own AOL contender) - has steadfastly stayed with Business-To-Business (B2B) markets. It is important, however, because in the longer term, albeit perhaps not for a decade or more, C2C networking is likely to become the biggest market of all.

Microsoft, despite its dependence on the mistakes of others combined with a generous measure of luck for its success, still follows many of the traditional approaches to business. On the other hand, AOL seems to have lived life in the fast-lane dependent only on luck! What is more, it seems to have succeeded despite a vision in which its main business has often been seen by it as merely peripheral!

In much the same way that the precursor of IBM was first created by Charles R Flint, the precursor of AOL was initially set up as early as 1978 by another 'serial' company builder Bill Von Meister; as a home information utility, 'The Source'. Unlike Flint, though, most of Von Meister's efforts failed spectacularly; and 'The Source', running out of funds, was soon sold to Reader's Digest; and eventually ended up as part of Compuserve (which was, in turn bought by AOL - thus closing the circle!). Only AOL achieved significant success; and then not for another decade, long after Von Meister left - to die penniless. Foreshadowing the interest in the entertainment business which - despite the reality of its customer base and revenue streams - has since seemed to represent the most important focus of AOL's own vision of its future, Von Meister's next venture was the 'Home Music Store' selling popular music down the line. Like Napster, two decades later, this was predictably destroyed by opposition from the record industry!

Von Meister then set out on another new 'entertainment' venture. This time it was in on-line video games which - spurred on by sales of millions of Atari consoles - were the flavor of the month in the early 1980s. Launched at the 1883 CES show in Las Vegas, the new venture was called the control Video Corporation, CVC, and was funded by - amongst others - Dan Case. The product, the 'Gameline Master Module' which attached to the Atari (though Atari had refused to support the operations), was initially a hit with the venture capitalists. Perhaps more important, it brought in, as a consultant and then full-time employee, Steve Case. He was Dan's brother and  was to dominate the later development of AOL, as had Thomas Watson Sr. that of IBM. Unlike others in the industry - who (like Bill Gates) started life as IT specialists - but like Thomas Watson Sr., Steve Case's career to that point to that point had been in marketing; though classically, for the age of mass marketing, as an assistant brand manager at P&G and then manager of new product development at Pizza Hut.

Surprisingly, considering his marketing background and unlike Gates, Case has adopted a low key image. He has said[l] "In my case it really is about believing that there is work to be done to build a medium that we can be proud of…It's not about so-called power; or visibility or stature. These are things I don't particularly enjoy." He has been described by The Washington Post[li] as "…the technocrat politician more than the entrepreneur.." which then added "This approach…may be betters suited to long terms success" and significantly, in contrast with Bill Gates, "He integrates a lot of views when he's making decisions…" Another reporter at The Washington Post[lii] categorized him as "…half nerd and half marketing guru…He's enough of a nerd to be really into online services but he has all the marketing skills to understand how to make it work."

Von Meister's bad luck held once more. The bottom dropped out of the video games market almost as soon as CVC had been launched, he was on his way out once more and Bob Cross was brought in to turn the company around. But salvation was no easy task for, in 1984, it was said that "the company had no clear product, no apparent market, no real money[liii]". The real salvation, perhaps, was the investment by Jim Kismet who had previously made his fortune by opening a chain of singles bars. It also relied on a deal with the Bell South telephone company. The new product, a modem for home use, once more failed to ignite interest; and Bell South, joining the lengthening line of punters who had lost their shirts backing Von Meister, lost its money. To avoid bankruptcy, the business was transferred to a new company with yet more new founders, Quantum Computer Services, which was to provide private (Q-Link) online services for Commodore customers. This was, a decade or more down the line, to become AOL. But growth in revenue was slow. By 1986 there were only 10,000 customers and revenue in 1987 was only $8.6 million; although for once they did make a small profit! Indeed, as Commodore's fortunes declined, they needed to find a new platform. It couldn't be IBM, since that was locked into Prodigy - and indeed was relaunching this. So, the choice was Apple, which was to receive a 10 per cent royalty on sales. Although this brought larger numbers of customers, Quantum eventually felt itself too cramped by the marketing constraints it involved. In the resulting break-up Apple also joined the lengthening line of partners which had lost money - in this case including $2.5 million to buy back the rights to its own logo!

This, in 1989, was the point at which the name of the service (but not the holding company) was finally changed to America Online, which offered much the same private online services; though now across the whole market - including IBM users. By 1991, however, AOL was to ready to become a 'legitimate' business. Previously it had lived almost as a serial 'scam'; persuading a range of individuals and organizations to bet their shirts, and then lose them, on very borderline ventures. Indeed, Kara Swisher, in her very comprehensive book about AOL (AOL.COM[liv]), summarizes this feature very 'positively' as "…throughout its volatile history, one of AOL's greatest strengths - and also the subject of later criticism - has been its ability to fund itself through a variety of innovative and controversial methods…[for] its business operations never threw off enough cash to cover the costs of its growth."

Scam Entrepreneurs…small start-ups are often more dependent on selling their ideas to investors than to consumers. They typically need 'venture capitalists' to fund them. Some of these, such as Flint at IBM and Markula at Apple, are legitimate - even if they do at times sail close to edge of legality. Some, such as Von Meister, can be more reminiscent of Mel Brooks' 'Producers'. Whichever the type, however, there is always considerable risk involved, the odds are against such small ventures, entrepreneurs - along with potential investors - should think long and hard before they jump! Even that part of management thinking devoted to SME's (Small and Medium-sized Enterprises) tends to optimistically ignore the risk of failure.

Now, though, it had something genuinely worthwhile to promote. Indeed, even Compuserve, which had been growing its consumer online service for hobbyist computer users  - based on its interactive computer bulletin boards (though its main strength became its financial and computer data) - and which had bought The Source (with its 53,000 members) to add to the half a million of its own, now wanted to get hold of Quantum's 110,000 members - and eventually offered $50 million for the company. But that was below the $60 million which its management saw as the very minimum it was worth, so the sale fell through - and, of course, AOL later returned the favor by actually buying Compuserve!

Its name, changed to AOL for the company as well as the offering, was now becoming known but sales growth was still slow. Even so, in 1992, it went public on NASDAQ; only to suffer the attentions of Paul G Allen, cofounder of Microsoft, who steadily acquired its stock. Even Bill Gates made an approach; though his characteristic opening gambit, "I can buy 20 per cent of you or I can buy all of you. Or I can go into this business myself and bury you[lv]", virtually killed that deal on the spot; though Case only won the subsequent vote on AOL's board by one vote[lvi]! The company possibly was now, in 1993, worth around $400 million - having multiplied its value sevenfold in two years. Bill Gates, in turn, went on to try and bury it with MSN; but, for once, signally failed to do so.

Having beaten of Gates, and having come to terms with Allen, Steve Case was at last allowed to emerge and make a marketing advantage out of AOL's flexibility: "From the early days, we recognized that communications - a combination of chat and e-mail - were critical building blocks…So our bias was on creating tools, empowering people, and letting then use them in any way they thought appropriate - sort of 'Let a thousand flowers bloom'[lvii]". AOL had, at long last, settled on the formula which was to make it a power in the land. Unlike the 'seriousness' of Compuserve and the IBM dominated 'family' atmosphere of Prodigy, the freedom of expression offered by the AOL service attracted subscribers; even if much of the chat was sex-oriented (and that posed major problems with the industry regulators later in the 1990s when something like half its business was still sexually oriented!). In addition, using the icon-based approach it had been forced to develop when it served Apple, it was easy to use (and install, taking only five minutes for this) - and, unlike Compuserve and Prodigy, it now could be fully used on either Apple or Windows.

Newsweek[lviii] summed it up as "By hosting hundreds of chat rooms, sponsoring the most popular instant messaging service on the globe and giving granny an e-mail address she'd never figure out how to change, AOL kept its audience and saw growth accelerated by the Network Effect (the more people who sign up, the more valuable the service)." It is this latter aspect, in particular, which we believe will be the key to future value.

Owning Members…C2C (Consumer-To-Consumer) networking will eventually become just about the most profitable ecommerce market. Ownership of its members will then be the key to value for portals; and the financial markets are already in effect putting a value per member - at the peak of the dot.com boom it was something like $15,000 for each AOL members. As yet, though ecommerce is a strange new land for business academics!

The product was now right, and the newly hired Jan Brandt took the most important, albeit also the riskiest, marketing step AOL ever made. She started the carpet-bombing of America with free AOL software disks - ultimately sending out 250 million of these! This was the ultimate in sampling, giving the millions of recipients immediate access; with free hours of usage and waiver of the first month's fee. It was a classic marketing success; in that 10 per cent of those receiving the disks used them for a free trial of the service. Over the years from 1992 to 1996, the 25 million free trials then turned into 5 million members[lix].

Sampling…the most effective device for persuading potential new customers to but your product or service is to let them sample it. Otherwise the significant step they have to make is to buy the 'product' just to see if it works for them. This trial offer is just important in B2B and services. Indeed, despite the popular image of samples plopping onto the doormat, internet services (which can, without costs to themselves or their prospective customers, offer a free trial - limited by a time-out) are likely to become the most important users of this approach. Management theorists tend to lump sampling together with other forms of promotion, most of which are usually price discounts in one form or another, which misses the point.

Even though the response generated problems with AOL's service response times, and clogged its lines, by mid 1994 AOL had a million members on line. Just as important, as Kara Swisher reports, "…its share price was red hot, and most of its purchases [of companies taken over] could be made with its stock as currency. Over the next year, AOL would use the stock market to literally print money - in the form of AOL shares."

In one of its first moves subsequently, it bought 'Bookline' to be its browser - only later to be replaced by the much bigger acquisition of Netscape. It also acquired Advanced Network & Services Inc (ANS) to build its communications network. It also started to build a stable of companies creating interactive material, buying stakes in them or entering into joint ventures; as new content creation became its vehicle for brand building. To service its new 'channels, deliberately modeled on television to make it look like a media company, it also created its own 'in-house content studio', and had potential providers make pitches to it as producers did to the film studios! The first winners included 'Motley Fool', the irreverent - but very successful - financial site.

Once again running short of funds, it worked its magic with another public offering; and at the end of 1995 its fortunes seemed to be set fair. But the world was moving to the internet and AOL's 'closed network' offering started to look stale - until, seemingly against all the odds (but eliminating Justice Department problems), Bill Gates offered to put the AOL registration icon on the Windows desktop; albeit not as prominently displayed as that of MSN. At the same time, however, AOL brought in Netscape to replace 'Bookline' as its own browser; though it didn't buy it until 1999 - for $4.2 billion -- by which time its deal with Microsoft had become much less important.

By 1996, AOL was worth $6.5 billion[lx], up 15,000 per cent in just two years, and had - once more against the odds - beaten Microsoft in the online wars of the previous year. But it was becoming the victim of 'churn' amongst its customer base, losing members almost as fast as it signed up new ones (at a cost to it of $50 each); and this also fatally damaged the AOL accounting practice of amortizing its customer acquisition costs over two years. Even worse, its competitors, including the new ISPs which were entering the business, were starting to offer unlimited usage deals - where AOL made its money from the hourly usage charges! It was forced to introduce a flat rate offering; at $19.95 for 20 hours use. The panic amongst its investors knocked nearly 70 per cent off its value.

The situation was compounded by the technical problems AOL was experiencing in servicing its more than 6 million customers. This was highlighted by a widely publicized system outage, for almost 19 hours, in August 1996; and AOL was forced to invest $250 million (eventually more than $350 million) in its communications network. Worse was to come. AOL lost $353.7 million in the first quarter of 1987, when it had to write off its deferred subscriber acquisition costs. In 1997, to improve profitability, AOL started to charge its content partners for access to its now very valuable customer base. In 1998 it, at long last, acquired Compuserve; whose AOL copycat, Wow, had failed miserably. In any case, despite its 2.6 million members, Compuserve was regularly making losses. The deal was, once more, innovative - with AOL 'buying' Compuserve from WorldCom as a swap for AOL's ANS subsidiary (which, though successful, had been getting in the way of AOL's relations with the other communications networks). AOL even got $185 million in cash! Its stock jumped 600 per cent in 1998 alone!

The icing on the cake came when at the end of the 1990s, its miraculously escalating stock market valuation - having multiplied almost a thousandfold over the decade - allowed it to buy Time-Warner. This was one of the great media corporations, with - as IBM once had - a successful track-record stretching back almost a century. This was a fairy tale come true, though whether the Time-Warner kiss will turn the frog, that AOL still is, into anything else is yet to be seen. In some respects it smacked of the 'water' that IBM's Flint had added to his deals, though Wall Street - lusting for the dot.com boom to come and very aware of the 30 million members (ten times as many as MSN) AOL now had - was a very willing partner. In many respects it took AOL back to its roots, as an online entertainment provider which it had long wanted to be; and possibly took its attention away from the online customer services base which was its real strength. More worryingly for the resulting conglomerate, it produced a corporation which was straddling two very different markets - entertainment through cable and communications through the internet - which were proving impossible to merge.

Pyramid Builders…we can all sneer at the housewives who foolishly get caught up in chain letter scams. Yet the most foolish of all are those in the financial markets who are happy to accept the inflated values set by those who create pyramids of businesses where paper profits are hidden by the takeovers themselves at one extreme and by the Byzantine mazes of globalized accounting systems at the other - all nicely wrapped up in creatively accounted packages!  Management educators do not warn their students that the biggest confidence tricks of all can be played on the financial markets.

I have also talked a great deal about customers at IBM and Microsoft, but AOL really us different because these are ongoing members; and they are much more important to it than its staff. Even so, it is worth recording that its culture is supposedly[lxi] a mix of Apple's enthusiasm and dedication to its product with Microsoft's work ethic and competitive nature. In particular, though, its dress code was definitely very casual!

The Internet Proper

Reverting to the Microsoft story, in 1993 - after his rebuff by AOL - Bill Gates started to build its own online offering to compete with it, which eventually became MSN. It, of course, had the major advantage of being given pride of place on the Windows desktop; though its use of the Windows folder approach made it much less approachable than AOL's. Furthermore, its launch slipped as did that of Windows 95, to which it was tied and by which it was also overshadowed. In the event it managed to sign up nearly 200,000 users in the first two weeks, but then - lacking heavyweight promotional support - its growth stagnated; whilst AOL at the same time was bringing 250,000 new users every month. In the event, MSN reportedly[lxii] lost $500 million in two years.

It was not until the end of 1995 that Microsoft ported MSN to the Web. Indeed, Bill Gates was very late recognizing the true potential offered by the Internet. The story[lxiii] is that suddenly in May 1995, having spent 10 hours browsing the Internet and finding almost no Microsoft standard file formats (but lots of those from competitors), he issued a directive which said "I want every product plan to try and go overboard on Internet features". Previously he had, instead, focused his teams solely on the 'grand convergence of television and computers'[lxiv] (the then media mogul's, incorrect, cable/megastream view of the future of the 'Superhighway'), even though he already had a team developing a technically similar (low-bandwidth) Remote Information Protocol (RIP) for the 'home' - code-named Homer, it was to control all the devices (from lights and heating to washing machines and television) around the home. Gates killed off this advanced consumer technology group's (low-bandwidth) work in 1992, apparently without ever realizing the coming importance of the Internet! 

Long Range Forecasting…the most powerful technique for exploring the future over the longer term, beyond ten years ahead in general (but possibly less than half that in the very fast-moving high-tech industries) is scenario forecasting.  A number of management textbooks, including my own, cover the subject of scenarios in some depth.

The end result, where his directive[lxv] also said "…with a major goal to get OEMs shipping our browser preinstalled", was all the legal trouble he brought on himself when he later tried (otherwise successfully) to recover his position by destroying Netscape, which had actually implemented something like RIP. At the time, his own browser, a separate application called Explorer 1.0, was licensed from Spyglass (the company which had bought Marc Andreessen's previous Mosaic browser code - the precursor to his later Netscape) and Microsoft then sold copies of it for $50 - even though it had often been given away free as shareware over the Internet.

Just, but only just, living up to its name Windows 95 was launched at the end of 1995. The outcome of Gates' Internet oversight was, however, that Windows 95 was overshadowed at the 1995 COMDEX by the new kid on the block: Netscape! Once again, the next version - Windows 98 - was to be the interim product just supporting new hardware; leading the way now to NT. To add something to persuade users to switch, they added the Explorer 4.0 shell to the core product[lxvi] - finally inviting the unwelcome attention of the Justice Department!

Tensions

The split between its two incompatible systems continued into the 21^st century, however, with Windows 98 moving on to become Windows Millennium Edition, still with much of its code written in Intel's 16-bit assembler[lxvii] with Windows 2000 still based on DEC architecture!

XML…Up until the late 1990s, Bill Gates had insisted that Windows was the main, indeed almost the only, offering which really mattered. At that time, however, the advent of the Internet posed a genuine - if very unwelcome - challenge to Windows domination; even within Microsoft itself.

Even as late as 1997, Bill Gates lacerated the Java team for failing to protect Windows, stating: "Why don't you just give up your [stock] options and join the Peace Corps?…Haven't you ever heard of Windows? Windows is what this company is about![lxviii]".

On the other hand, the development group under Bosworth - who had (unusually) negotiated a guarantee of independence from Bill Gates - was responding to the new demand for XML; a cross-platform development from the basic HTML language which made it possible for software applications to work together regardless of the underlying platform. On the basis of some stealth, not least by covertly persuading IBM and Oracle to also support it and boosted by the col