The announcement on December 31 that IBM and Merrill Lynch & Co had decided to dismantle their International MarketNet share price data joint venture marked the end of almost the last surviving of a whole string of joint ventures and diversifications entered into by IBM in the US in the first half of the 1980s. Most of the ventures are now no more than a dismal litter of torn up agreements and high hopes dashed. Typical is the belief of International MarketNet that it could be generating a billion dollars of business for the two partners by the turn of the decade. Still alive, just, is the Trintex viewdata venture with Sears Roebuck, but the third partner, CBS, has pulled out of the triumvirate, and there is as yet very little to show for the investment made. IBM also hoped to become a major player in optical disk systems for the educational market through its DiscoVision joint venture with MCA and Pioneer of Japan, but that one fell apart some four years ago. Even closer to home, the agbeement that Texas Instruments would do the chips for the Token Ring local area network began to look rocky after IBM’s first series of Token Ring products bypassed the Texas chips for lower performance parts from elsewhere.
About the only joint ventures that still look viable are the ones with Merrill Lynch on computer lease finance – something that IBM really should know all about. Eyebrows have been raised over the remarkably attractive lease terms IBM has been offering on that front, but the comfortable assumption has been that IBM must and did know something its competitors didn’t about planned life cycles of machines like the 4381. But given the company’s inept performance over the same period in so many other arenas, it is no longer safe to assume that IBM won’t find that it has to take a bath on its own residual values as well. Perhaps the most ambitious venture of all was the formation of Satellite Business Systems with Comsat and Aetna Life & Casualty in 1976. Even when IBM took control of the venture in 1984 and sold it to MCI Communications in 1985 in exchange for an initial 16.6% of that company, the moves could be read as part of a grand strategy on IBM’s part to become a major player in telecommunications by stealth. The thesis was that for anti-trust and regulatory reasons, IBM was happy enough to see SBS keep a low profile – until such time as the regulatory climate changed in its favour and it was permitted to sell communications services in tandem with its computers. The MCI deal looked superficially like a setback until one recalled that IBM had agreed not to take its holding in Rolm Corp above 30% yet still ended up buying Rolm outright. The assumption was that IBM had pulled a neat trick, getting rid of the loss-making headache of Satellite Business Systems, but putting itself in a position where if all went well, MCI, with IBM’s backing and support, would become a painfully successful competitor for AT&T – but would at some stage run critically short of the cash needed to maintain the pressure. At that point, IBM would open its moneybags and make MCI an offer it couldn’t refuse, in return for a controlling interest. IBM would then have made itself a head-to-head competitor with AT&T on the cheap.
Instead, MCI is preparing to report heavy losses this year, customers don’t seem to be dashing to buy MCI data services with their IBM computers, and, worst of all, Satellite Business Systems, which was on the brink of turning a first profit when IBM sold it, has plunged back into the red to the point where MCI is going to have to take substantial write-offs on satellite equipment. The Rolm acquisition is beginning to look like a deep disappointment as customers prove slow to take up the concept of building their office automation systems around a central telephone switch. The investment in Intel did the latter much more good than it did IBM, so much so that IBM has made arrangements to divest its holding in the chipmaker. IBM’s agreements with Microsoft Corp haven’t given it the operating system it needs ev
en for the 80286 chip, let alone the 80386. Even within its own businesses, IBM’s attempts at diversification have largely failed. The Merrill disappointment has been closely followed by the announcement last week that the IBM Instruments Division was to be disbanded. The overwhelming penetration of the Personal Computer masks the fact that the machine has done incomparably more for companies that feed off the market that IBM has created – the Lotuses, Digital Communications Associates, Compaqs, ComputerLands, Olivettis – than it has done for IBM itself. Within the Personal Computer world, IBM has attempted to diversify into home computers with PCjr and failed, tried to become a major retailer of its own products and failed, ending up selling its chain of Product Center stores to Nynex Corp. And although IBM is believed to be preparing to re-enter the home computer end of the business, probably with an 8086 machine built for it by Matsushita, it is unlikely to be much more successful with that than it was with PCjr.
More likely the machine will simply keep the MS-DOS standard alive in American schools, while Apple continues to take about 50% of the total market for itself, and IBM has to share the other 50% with a whole string of clonemakers that will ride the IBM offering into the schools market with similar machines and software that are both better and cheaper than anything IBM is prepared to offer. Faced with such a dismal catalogue of failed diversifications, IBM increasingly gives the impression of being assailed with the very characteristics that it so assiduously seeks to instil in its competitors and its users – the dreaded FUD, fear, uncertainty and doubt. It needs to make really bold moves to recapture the initiative from the competition, but having seen what its name did to a product that was intrinsically by no means particularly bold, the Personal Computer, it is scared rigid of the impact that a fiercely-competitive 80386 box would have, not simply on the market, but on its own other lacklustre low-end products. Superficially, the 9370s appear to represent a bold move, and are certain to sell in enormous numbers – by the standards of a Hewlett-Packard or a Data General. But are they attractive enough, and above all cheap and easy enough to use to sell in enormous numbers by the high standards of the old, super-confident IBM of only three short years ago?