When both Channel 4 News on UK-wide television and the British Broadcasting Corp’s Financial World Tonight on Radio Five Live carry items on the bug in Intel Corp’s Pentium chip, it is quite clear that the analysts that said the whole affair would be forgotten in a week and everything would be back to normal were whistling in the dark. On the contrary, the affair is likely to become a business school textbook study on how to mismanage a damage limitation exercise. The seeds of Intel’s apparently cavalier attitude to the bug lie in the fact that while the publicity side of the company has spent the past 18 months spending a fortune assiduously wooing the end-user and the consumer market in general with copious television and print advertising, with the commercials for the Vacancy upgrade slots being followed by the Intel Inside campaign and the big consumer push behind the Pentium, the company as a whole still does not actually regard the end-user as its own customer.
Worst offence
Intel sells to the likes of Compaq Computer Corp, Dell Computer Corp and AST Research Inc: they are the organisations it sees as its customers. So the first lesson is – get the company’s internal mindset right before you embark on enormously expensive advertising campaigns that sell a totally different view of your place in the world from the one that you yourself still hold. The company’s worst offence was not to announce that there was a problem as soon as it became aware of it: that mistake leaves it wide open to a whole string of actions seeking redress. Instead, it waited to go public until a user was severely inconvenienced or worse by the bug. The enormous gap between the company’s publicity and its self-view is highlighted again by the fact that it has been promoting the Pentium as suitable for applications for which people have hitherto needed a RISC-based workstation, and yet its internal mindset saw its microprocessors so narrowly used for word processing and simple spreadsheets that it did not think anyone was likely to be inconvenienced by the bug, its justification for simply putting the bug on its to-be-fixed list and not thinking any more about it until the wires started humming with angry traffic overwhelmingly critical of Intel. Andy Grove was right to go public and apologise for the bug, but by then, so many people were so angry that his explanation just got added to the list of the company’s offences and the traffic on the newsgroups switched to querying whether Grove had really had anything to do with the official statement that went out over his name.
Rocket scientists
Moreover the tone of the message was wrong: it was still professional to professional where what was needed was not simply an apology but a grovelling apology. It’s no good moving your products from a pure mass market into a market so broad that it encompasses realms occupied only by rocket scientists and nuclear power station designers and then lamely say that the Pentium has fewer bugs in it than any predecessor part: if you want to sell it for arithmetic-intensive computer-aided design applications, you can’t get away with making it only as good as the best late model parts aimed purely at the word processing and desktop database market. Those new users you are trying to win have got to be damn’ sure it really is much closer to being perfect if they are to go away content. A striking feature of the unhappy incident is that we have not seen any crowing by competitors over Intel’s misfortune: all chipmakers know only too well that there but for the grace of God go they. The issue is generating so much heat and controversy not because the company’s flagship chip turned out to have a bug too far in it: it is entirely because of the way Intel has mishandled the damage limitation. What should the company do now? It’s another occasion where the helpless response is if I wanted to get to there, I wouldn’t start from here, because the worst mistakes have already been made. As Compaq Computer Corp is already doing, it should encourage users to disable the mat
hs co-processor. The cost of a full recall would be horrendous, but if it took back all the affected parts, it could resell them as Pentium equivalents of the 80486SX with the co-processor disabled. And there are people out there that want a good chip simply because they feel they might want to do arithmetic-intensive work on their machines in the future: Intel’s come back to us for a replacement when you do answer simply isn’t good enough, and in the minds of such people simply compounds the offence. The company still doesn’t seem to grasp the enormity of the damage that has been done. It may all be to do with perception and to have very little to do with reality, but it is the world of perception that Intel plunged into and embraced as soon as it embarked on its vast advertising campaigns aimed directly at processor end-users.
Beethoven’s Ninth
Apart from the relative handful of people for whom the bug is a devastating problem, it is the people for whom the bug in reality matters least who are now the ones that feel most negatively about the Pentium and about Intel. People like to feel that the expensive durables that they buy are perfect, and just as the 99 people in 100 that would never notice that a manufacturing fault in a compact disk of Beethoven’s Ninth meant that four bars were missing from the second movement would still feel that the disk was fatally flawed for them, so there is an army of low-end business users and home computer buyers that is never going to feel entirely happy about using a Pentium machine when they know that there is something wrong in there. That’s a key difference between the professional and the consumer market, and one that Intel needed to understand before it put so much money and effort into putting its name and its products before consumers that would previously never known a Pentium from a petunia. The Pentium name itself is now probably so tarnished in so many minds that the company needs to give a completely new name to the the new generation without the bug. Meantime the company suddenly has a very good justification for its strategy of getting the successor P6 to market as quickly as possible and obsoleting the Pentium.
