How to lose share steadily in the printer market: pretend that you still own it
IBM Corp’s mainframe printers may produce mountains of output for banks, utility companies and manufacturers. But their importance to users of large computers is dwarfed by the rest of the printer market. IBM’s printer business will probably generate a billion dollars in sales worldwide this year. Hewlett-Packard Co, the largest of the many vendors serving the low end of the non-impact printer market, will sell $6,000m in printers this year, according to Wall Street experts, and about half this money will be for laser printers, the rest a mix of colour printers and inkjet printers. The total world market in non-impact printers, including Xerox Corp, NEC Corp, Siemens AG, Hitachi Ltd and lots of other vendors, almost certainly exceeds $10,000m a year and on top of that there is a maintenance business of comparable size plus a huge market in toner and other consumables. By our reckoning, then, IBM has approximately 10% of the markets for printer sales, service and supplies. Despite this modest market share, IBM acts as if it had five times that much clout. As a result, users of IBM mainframes (and AS/400s too) live in a world of 240 by 240 or 480 by 480 dot printing, a unique printer command language and a suite of programming conventions that are very tightly bound to the past. Just about every other kind of computer, the ones that account for 90% of the printer market, lives in a 300 by 300 or 600 by 600 world and sends messages to printers in PostScript or HP-PCL. Hewlett’s small and mid-range laser printers speak PCL, of course, and they all offer PostScript fluency as an option. They do not speak IBM’s AFP language. Hewlett would prefer an all-PCL world, of course, but it has bent to the will of its customers and in return they have helped it prosper. That is one reason Hewlett is the market volume leader, while IBM failed so miserably in low-end printers that it sold that business. All the desktop systems used to design documents talk to their local printers using one of the popular schemes; none use IBM’s methods.
No good reason
Even when desktop systems use IBM software for document design, software aimed at creating documents that will ultimately pour out of IBM mainframes, the draft copies of the documents are almost invariably created using a PostScript or PCL printer. There is no good reason why IBM has failed to enhance its big printers so they can speak PostScript or PCL. The company is just plain stubborn. IBM’s printers could easily provide the slightly higher resolutions favoured by the rest of the world. IBM should have no difficulty whatsoever supporting PostScript or PCL directly while maintaining compatibility with its old standards. All IBM has to do is listen to the market, which has voted for new standards with 90% of its budget funds. We expect that there would be a rush of buyers the moment IBM finally offers optional PostScript or PCL features on its heavy duty page printers. There would be indirect benefits, too. Accepting the popular printer standards would help IBM in its effort to persuade large enterprises that mainframes are the corporate servers of the future. IBM might even be able to market its big printers to users of large Unix systems, something it cannot do very well now. Who knows? IBM might even become successful enough in the printer business to once again choose the rules. From the August 1994 edition of Infoperspectives International, published by Technology News Ltd, (C) 1994 Technology News Ltd. All rights reserved.
IBM’s personal computers are in the doldrums after 1993’s dead-cat bounce
IBM Corp’s situation in personal computers looks more precarious than it has in a while. During 1993, IBM rebounded after a serious fall, increasing its sales by adding numerous AT-architecture machines to its product line, reducing prices and increasing its emphasis on retail distribution channels. This year, IBM tried the same tactics but was outmanoeuvred by Compaq Computer Corp, Apple Computer Inc and
, in the United States, a formerly small company called Packard Bell Electronics Inc. Who is Packard Bell? Nobody except a computer vendor with strictly retail sales that has shipments exceeding IBM’s retail plus corporate plus direct marketing volume. On Wall Street, the recovery of IBM’s personal computer business during 1993 would be called a dead cat bounce. We don’t know precisely where the phrase actually started, but we believe it initially described a temporary upturn in a stock (or maybe the whole market) that had fallen sharply for cause. Even a dead cat will bounce… once, may be the original phrase that entered what Wall Street calls a vocabulary. IBM’s personal computer sales have dipped because customers were a lot more savvy than IBM figured.
Friends and neighbours
Their education came in 1992 and 1993, when tens of thousands of naive, trusting home and small business buyers believed there was a high likelihood that IBM cared enough and knew enough to sell machines that would be useful and remain so. In fact, IBM’s offerings were too skinny for the software that buyers wanted to run – often by an order of magnitude. So, when asked by their friends and neighbours about buying an IBM personal computer, a lot of IBM’s 1992 and 1993 customers advised, Watch out! Such a warning is very difficult to ignore, despite the seductive power of IBM’s advertising. IBM says its mainframes are now selling so well that it is out of machines until next year. The company has not disclosed how much manufacturing capacity it has left, but there is no doubt that demand is greater than IBM or anyone else expected. We wouldn’t be surprised if IBM began to forecast a lasting rebound. But we think this is just another dead cat bounce, the result of customers finally executing purchase decisions that have been deferred for a long time. If IBM wants a sustained recovery in large systems and a happy customer base, it must work very hard to perfect the CMOS hardware it announced and fit it with attractive versions of all its systems programs. In the meantime, IBM must improve the value as well as the price-performance of 9021s and make the systems more attractive in other ways, too. It must also improve the bridges linking its current hardware to the CMOS generation. These bridges must be open so users with IBM-compatible machines can consider the IBM CMOS alternative. But bridges can be crossed in both directions. So those CMOS machines had better be pretty good. They should, in fact, be an order of magnitude better than 9021s. – Hesh Wiener(C) 1994 Technology News Ltd. From the September 1994 edition of Infoperspectives International.
