Pretty nifty, those new AS/400 Server Models, aren’t they. We are not likely to knock them anyway, because we’ve been saying for some time now that IBM Corp should come out with application-specific versions of the AS/400 starting with a dedicated server version, and what the company has announced is precisely what we had in mind. One of their biggest attractions from IBM’s point of view is that since they are no use as stand-alone AS/400s, they do not compete with the existing machines, so that their pricing is not tied to standard AS/400 pricing: IBM can price them solely to maximise sales and profitability without worrying that they will cannibalise standard AS/400 sales. Existing AS/400 users may well add some of the new servers to their installation, but in general, these would be incremental sales rather than substitute sales. The concern now is that although IBM has launched the right product, it doesn’t seem to recognise how – and to whom – it should be marketing it. It talks airily about extending the enhancements that make the new models exceptional database servers to the entire AS/400 family, which implies that it does see existing AS/400 users as the primary market, where surely the whole point of the new servers should be to win IBM new business from presently 100% non-IBM shops.

Stigma

The stigma of proprietaryness is digging the grave of the mainframe business, and the AS/400 is in mortal danger of inheriting the same stigma. IBM should be pulling out all the stops to persuade users of Unix compute servers and NetWare networks that all their database access problems could be solved if they added an AS/400 Server to their configuration, that yes, their Hewlett-Packard Co or Sun Microsystems Inc machines are making an excellent fist of doing their sums super-fast, but that there is no way that a Sybase Inc or Oracle Corp database on the Unix machine can answer database queries anything like as fast or as effectively as IBM’s database, the AS/400 Server. Get enough of the things sitting quietly in the corner on the Unix or NetWare network, unobtrusively getting on with what they are paid to do and causing nobody any trouble, and the proprietary sigma will melt away. IBM will be able to point to such installations to reassure those users tempted to assign all their data processing to a full AS/400 but nervous about its proprietary nature that they really have nothing to worry about – because all those open systems users are happily hammering away at the AS/400 Server with all their database queries without a care in the world. After all, there is no such thing as an open database management system: all are proprietary to the companies that developed them and it’s no easier to convert from Oracle to Ingres than it is to convert from VMS to VME.

Job-creation scheme

Why does all this matter? After all, the AS/400 is only a $13,000m business and IBM is a $60,000m company, so an extra 10% of AS/400 business that IBM might forego by not marketing the new machines in the most appropriate manner is not going to make a great deal of difference to the big picture. The answer is that the AS/400 is the only blue chip business IBM has left. The fact that anyone has to ask whether the personal computer business is profitable makes it clear that that business is well-nigh useless to IBM except as a job-creation scheme: if IBM does succeed in knocking the Gateway 2000 Incs and the Zeos International Ltds, even the Zenith Data Corps out of the market, there is no way that it is going to knock Compaq Computer Corp, AST Research Inc and Dell Computer Corp out too, and if it were to succeed in hardening personal computer margins to the point where they really covered traditional IBM-style costs, a host of new players would be born to force prices back down and eliminate those margins again. Much better for IBM shareholders would be for the company to leave the mass personal computer market to the companies that can make acceptable profits on present margins and – like Rover Group Plc concentrate on the higher-margin premiu

m products which with the ThinkPads, IBM – or IBM Japan Ltd – has finally demonstrated it is capable of doing. As to IBM’s other businesses, the RS/6000 is an attractive enough little operation,but it is tiny: it is very unlikely to reach $3,000m this year, but even if it did, it would account for only 5% of IBM’s turnover – perhaps 6% or 7% if IBM starts to shrink more rapidly, and far too small to make a real difference to IBM’s overall business. After that IBM is really left with the chip business, which is finding selling anything on the merchant market painfully hard work, its disk drive business, which has failed to get anywhere near even $1,000m in OEM sales, and the morass of the mainframe business, a hole that is only going to get blacker. The best brains in the analytical side of the industry are focussed on trying to find a strategy that will enable IBM to save the mainframe, but we can’t see one, because anything IBM does will rack up big up-front costs for lower returns and lower margins. Our counsel of despair is that IBM should cease almost all development that doesn’t have clear benefits to growing parts of the business, particularly the AS/400, put the mainframe base on a care and maintenance basis, keep the current generation in production indefinitely to meet whatever demand remains, and plan for slow attrition – while handing on to customers, particularly in the area of software licences, as large a part of the benefit of reduced spending as it can, so that while they won’t be IBM mainframe customers for too much longer, they won’t have any strong desire to shake IBM’s dust off their shoes altogether might indeed consider moving at least part of their work to the AS/400. Which is why it’s so frustrating that IBM never seems to be able to grow the biggest AS/400s as fast as its biggest customers need the extra power.

Laciviously

Oh, and there is the wonderful new world of services that IBM like every other manufacturer – is plunging into with such desperate hope. No-one doubts that IBM can win big, big business in facilities management, even in systems implementation, but few are confident that the company can make any money out of it: IBM has absolutely no aptitude for that kind of activity, has demonstrated time and again how totally hopeless it is at managing big software projects, and there is a growing army of companies growing up in the shadow of Electronic Data Systems Corp and Computer Sciences Corp, ready to pick up the pieces of IBM’s failed projects and relate the story lasciviously in fascinating and graphic detail to every future prospect where IBM is bidding. Unfortunately for IBM, if its future is not the AS/400, IBM has very little future, and one would feel a lot happier about the company if it showed more sign that it really was aware of the fact.