Why did Digital Equipment Corporation Fail?

Digital Equipment Corp Logo

I answered that question on Quora, but thought i’d put a copy of my (long) answer here. Just one ex-employees perception from prisoner badge #50734!

I managed the UK Software Products Group in my last two years at DEC and had a 17 year term there (1976-1993). There are a wide variety of components that contributed to the final state, though failing to understand industry trends across the company was not among them. The below is a personal view, and happy for any ex colleagues to chip in with their own perspectives.

The original growth 1957 through to Jan 1983 has based on discrete, industry based product lines. In combination, they placed demand on central engineering, manufacturing and sales to support their own business objectives, and generally ran the show to dominate their own industry segments. For example, Laboratory Data Products, Graphic Arts (Newspapers!), Commercial OEM, Tech OEM, MDC (Manufacturing, Distribution and Control), ESG (Engineering Systems Group), Medical Systems and so forth.

The finance function ran as a separate reporting entity with management controls that went top to bottom with little ability for product lines to unduly behave in any way except for the overall corporate good.

The whole lot started to develop into a chaotic internal economy, albeit one that meshed together well and covered any cracks. The product lines were removed in Jan 1983, followed by a first ever stall in a hitherto unblemished Stock Market Performance. The company became much more centrally planned, and one built around a one company, one strategy, one architecture focus (the cynics in the company paraphrased this as one company, one egg, one basket). All focus on getting VAX widely deployed, given it’s then unique ability to run exactly the same binaries from board products all the way up to high end, close to mainframe class processors.

The most senior leadership started to go past it’s sell by date in the late 1980s. While the semiconductor teams were, as expected, pumping out impressive $300 VAX silicon, the elements of the leadership became fixated on the date on which DEC would finally overtake IBM in size. They made some poor technology investment calls in trying to extend VAX into IBM 3090 scale territory, seemingly oblivious to being nibbled from underneath on the low end.

Areas of the company were being frustrated at the high end focus and the inability for the Executive to give clear guidance on the next generation processor requirements. They kept on being flip-flopped between 32-bit and 64-bit designs, and brought out MIPS based workstations in a vain attempt to at least stay competitive performance wise until the new architecture was ready to ship.

Bob Palmer came to prominence by stopping the flip-flopping and had the semiconductor team ship a 64 bit design that completely outperformed every other chip in the industry by 50-100% (which ended up being the case for nearly 10 years). He then got put in charge of worldwide manufacturing, increasing the productivity by 4x in 2 years.

The company needed to increase its volume radically or reduce headcount to align competitively with the market as it went horizontally orientated (previously, IBM, DEC and the BUNCH – Burroughs, Univac, NCR, CDC and Honeywell were all vertically orientated).

Ken Olsen got deposed by the board around 1992, and they took the rather unusual step of reaching out to Bob Palmer, who was then a direct report of SVP Jack Smith, to lead the company.

While there was some early promise, the company focussed around a small number of areas (PCs, Components & Peripherals, Systems, Consumer Process Manufacturing, Discrete Manufacturing and Defence, and I think Health). The operating practice was that any leaders who missed their targets two quarters in a row were fired, and the salesforce given commission targets for the first time.

The whole thing degenerated from there, such that the company made equivalent losses in Palmers reign greater than the retained profits made under Olsen 1957-1982. He sued Intel for patent infringement, which ended up with Intel settling including the purchase of semiconductor operations in Hudson. He likewise sued Microsoft, which ended up with Microsoft lending DEC money to get its field force trained up on Microsoft products (impressive ju-jitsu on their part). Then sold the whole company to Compaq.

Text in some of the books about DEC include some comments by C Gordon Bell (technical god in DECs great years) which will not endear him to a place on Bob Palmers Christmas Card List (but words which many of us would agree with).

There was also a spoof Harvard Business Review article, written by George Van Treeck (a widely respected employee on the Marketing Notes conference maintained on the company network), which outlined the death of Digital – written in 1989 or so. Brilliant writing (I still have a copy in my files here), and he guessed the stages with impressive accuracy way back then. His words are probably a better summary than this, but until then, I trust this will give one persons perception.

It is still, without doubt, the finest company i’ve ever had the privilege to work for.

Rest in Peace, Ken Olsen. You did a great job, and the world is much better for your lifes work.