It’s good news week as IBM begins to face up to reality and accepts it may shrink…

Some very good – if surprising – news is beginning to seep out of IBM Corp, although whether it will go down too well with the knee-jerk reactors on Wall Street is another matter. We have commented many times that now IBM has seriously embarked on a programme of cost-cutting, it is discovering that it is not a once-for-all fix, but is apt to turn into the policy Chairman Mao Tse Tung advocated in one of his loonier phases, that of Permanent Revolution. Another Chinese proverb springs to mind he that rides the tiger finds it difficult to get off. The good news is that IBM seems to have accepted that there will have to be still more reductions in staff levels year by year for the foreseeable future so that where one manager was assigned to each of all but the smallest agents in the UK, it now has just two for the whole agent programme. IBM now seems to be determined to become the low-cost manufacturer in the industry and it realises that it must offload more and more of its marketing on its resellers. Similarly, it is trying to drive its manufacturing plants harder by becoming a major player in the OEM market: again others bear all the nitty-gritty marketing costs – only one or two IBM salespeople are needed to close those OEM contracts worth several tens of millions of dollars apiece. But what all this means is that the target of $100,000m in annual sales will become an ever more receding and distant memory: indeed it implies that IBM will see $50,000m a year in turnover again sooner than it sees $100,000m. Because it is certainly feasible – and attractive – to have your marketing partners bear a far higher proportion of the marketing costs than hitherto, but to do so, those partners have to be allowed to keep a far higher proportion of the sale. The trends towards third party and OEM sales means that IBM’s business should become a lot more profitable – but all the trends in the industry, downsizing, moving to open systems, means that IBM will find it hard to keep its turnover at present levels, let alone grow it further in the near term.

Of tack that seems to be too good to be true

There is a hint of another piece of very good news for IBM Corp’s fans blowing in the wind, although it is still a very long way from being a certainty. For at least 15 years now, DOS-to-OS conversion has been regarded as a cornerstone of IBM’s mainframe strategy. Those VSE users, now stripped of their original indentity, were earmarked as future cash cows whose fate and destiny was to be led to the MVS slaughter. But despite all IBM’s blandishments there remains an awful lot of them, stubbornly sticking where they are and refusing to submit to their preordained fate. IBM has done all it can to make them feel unwelcome in their present abode, saying no they can’t stay where they are and expect to be welcomed into the Systems Application Architecture fold, and then, well yes, they might, and now back again to no, they remain outcasts, albeit with a shiny new bell to ring to signal their uncleanness labelled VSE/ESA. But now, ADM Consulting Inc suggests that in a near-heretical change of policy, IBM won’t even raise an eyebrow if they say they are considering migrating instead to the AS/400. If true, this is wonderful news. It means that at last the dead hand of the IBM hierachy, which put the heirs and successors to the long-deceased Data Processing Division first, last and always when any conflict of interest or demand for development funds arose, is beginning to crumble. The long-standing policy and its concomitant culture have been a dead hand on progress, innovation and liberation for 15 years now, most notably starving for nearly a decade the System/38 of the development funds it needed, and condemning today’s interactive users to a totally inappropriate hardware architecture born in the early 1960s. If ADM is right, within five years or so, the AS/400 will represent a bigger business for IBM than ES/9000.

Looking for the strategy behind the parallel

IBM-Apple IBM-Intel alliances

What is IBM Corp’s strategy for the desktop? It seems hard to believe that the company, having gone to enormous lengths to try to ameliorate the problems caused to itself and in users’ minds by its 1970s and 1980s proliferation of incompatible mid-range systems, intends to recreate those problems for itself and its users on the desktop. Yet what appears to be a 10-year commitment to Intel Corp’s iAPX-86 architecture alongside the IBM-Apple-Motorola PowerPC project promises to do just that. In the past, it was a maxim that giant corporations were inherently stupid, but that IBM was much more intelligent than most. IBM gave that nostrum a knock by failing completely to forecast how successful an open personal computer architecture with an IBM logo on it would be, thereby laying the seeds for all its current woes and unleashing all the forces that are devastating the computer industry today. The Personal Computer was emphatically a Good Thing for users, but it has turned into an unmitigated disaster for IBM. The new agreement with Intel will enable IBM to improve margins in its PS/2 business, but it is unlikely to buy the company enough to turn the PS/2 from a cash drain to a cash cow – the margins it will be able to command are still unlikely to match those that Apple still enjoys by virtue of its proprietary architecture. Moreover the agreement with Intel looks to be so anti-competitive that it is bound to raise hostility from the rest of the industry: is the intention simply to make a putative Personal Systems Inc a more valuable property? What would be the motive behind that? Suppose IBM were to reach agreement for a friendly acquisition of Apple – a share exchange merger. It would almost certainly be told by the anti-trust division of the Justice Department that it could go ahead only if it agreed to divest Personal Systems Inc. The fact that it was able to get financing arranged for the divestment of Lexmark International Inc demonstrates that even in these sour times, funding can still be raised for buyouts that are solid prospects – as Personal Systems would certainly be when released from the IBM shackles and constraints – and overheads. IBM would be able to commit to minimum purchase levels of PS/2s for five years or so, the agreement with Intel would have to go with the company, but contract use of IBM’s process technology could still be available to the company – and IBM would be shot of a problem child yet be in a position to commiserate with users seduced down the OS/2 path by saying hey, we didn’t want to get rid of it – the government made us do it. So we return to the question posed at the start – does IBM have a grand strategy up its sleeve, or is it just plain stupid?