DEC & the Alpha: The Human Cost of Bucking the Innovator's Dilemma
November 4th, 2022

One of the insights Clay Christiansen made in Innovator's Dilemma is that companies don't become stupid so much as the business model changes under their feet and every incentive pushes them to ignore it until they fail.

He mentions hard disk drive manufacturers and steel makers but the example applies to computers in many ways. Microcomputers (e.g. modern PCs) replaced minicomputers which replaced mainframes.

But there is also a human cost associated with changing and this often provides friction. An interview with the Computer History Museum touched on this recently. It was with Robert Supnik. He was a member of the team that pushed for the DEC alpha line of processors.

A brief recap: Digital Equipment Corporation, DEC for short, invented the minicomputer circa 1960. These were very popular and thus very lucrative. Imitators like Data General and Interdata sprung up to get a slice of the action. Come the 1980s and workstations from companies like Sun & Prime Computer were starting to catch up. Minicomputer companies were in descent and many went bankrupt. After a time, you could buy a workstation with a Motorola 68000 to do almost as much as a VaX. One cost $9,000 and the other $150,000.

So far its a classic innovator's dilemma. Sun had a new business model that let them sell at a lower profit margin (~50%) than DEC, the incumbent (~75%).

But unlike many Innovator's Dilemma stories, DEC did somewhat bounce back. They built the Alpha processor which was competitively priced to a Sun workstation. It would also use Unix, sort of a differentiator between workstations and PCs. In addition, it had a somewhat better pricing structure because it could use Windows NT which was charged by the machine instead of by the core. DEC Alpha workstations became popular with the dot com boom because of their 64-bit RISC design which meant you could run a very fat machine to answer thousands of simultaneous connections. Amazon bought them to handle multiple Christmas surges in the late '90s.

Still, the introduction of the DEC Alpha didn't go so well for the engineers involved. Though they played a major part in saving the company, some bridges were burned. DEC moving to the Alpha meant many departments were redundant. Board layouts, power supplies, connectors were all standardized to use off-the-shelf components in order to be compatible with the PC. These managers lost the confidence of their subordinates. Nobody wanted to work for them. Lacking clout at the company, they moved on.

This human side of the Innovator's Dilemma rarely gets talked about. It seems particularly pertinent nowadays. One can talk about how organizations need to be adept at how they can be undercut and that they're better off doing the hard thing now before they're destroyed. But what of the managers? The employees? A company is composed of them and those people might still want to get invited to parties and BBQs. Is it worth throwing all those away too?

If you want to watch the full interview, which I highly recommend (all six hours!), you can do so below: