Building A Better Mousetrap is Not Enough
If you observe rapidly changing fields of endeavor, like the computer and electronics industries, over time you eventually come to the paradoxical conclusion that superior technology often loses out to its inferiors. The 68000 lost out to the 8086. Beta lost out to VHS. Multics lost out to lesser operating systems. And now Alpha loses out to IA64. The best explanation why this often happens is that backers of inferior technology recognize its deficiency and simply try harder to sell their product. They strenuously try to make up for shortcomings through non-technical means that often prove decisive if the gap isn’t too large. These means include high powered marketing, skillful exploitation of political and regulatory factors, and getting to market first to take advantage of the so-called network effect. Conversely, superior technologies are often created within organizations dominated by an engineering culture. Techies often expect everyone to think like techies and falsely assume customers will readily appreciate the mathematical beauty or abstract elegance of a superior technology, and will buy it for its own sake. As a result they seriously undervalue the need to actively market their superior technology. A good example of this principle is the Motorola 68000’s failure to dominate the 8086 in the face of Intel’s “operation crush” in the late 1970’s .
The question of Alpha’s marketing viability is a sore point for many Alpha users and observers of the technical scene. Although the marketing ineptitude of the late great Digital Equipment Corporation is still legendary among veteran computer users, the passing of the Alpha to Compaq was supposed to change all that. Compaq became very successful by selling other people’s technology. But if anything, Alpha marketing got worse, possibly because of rivalry between the newcomer and existing x86 product divisions which may have felt threatened or jealous. Many technical and academic computer users report eerily similar anecdotes of expending great effort to obtain an Alpha evaluation unit or even purchase multiple systems from Compaq or its distributors only to give up in exasperation. I had one of those experiences myself several years ago. This contrasts sharply with the generally swift reaction that the slightest hint of interest will invoke from a well-oiled selling machine like Sun Microsystems.
Alpha’s economic viability simply comes down to market share and margin. The total market for high-end 64-bit MPUs is perhaps on the order of a million parts per year, and these drive system sales of over $40 billion per year. At best, Alpha had a 10 to 15% market share. So it would be at a serious disadvantage in economies of scale when MIPS, PA-RISC, and even Xeon to some extent, are replaced by IA64, and commoditization of 64-bit hardware reduces margins throughout the industry. There is little doubt that Alpha, generally and widely regarded as the best architecture and processor technology in existence, could have achieved much higher sales and market share over its lifetime if competently marketed and supported by a stable, forward looking corporation. Compaq was likely quite truthful when it asserted that it couldn’t keep Alpha going as a viable business. Samsung and API Networks can’t escape their share of the blame either, as they seemed far more interested in price gouging existing customers than adding value or expanding the overall market. It is rumored that at various times in its short life Alpha technology was offered to Intel, Apple, and Sun, and was rejected due to a combination of valid business concerns and not invented here (NIH) syndrome. It is interesting to hypothesis what might have happened if events had taken a different course.
Discuss (16 comments)