At the start of the 1980s four men controlled the US IT industry – Ed deCastro (Data General), An Wang (Wang), Ken Fisher (Prime Computer) and Ken Olsen, of what was officially the Digital Equipment Corporation but a company that was always known with affection by its many customers as DEC. All MIT-Havard educated engineers, with powerful egos and total technical self- confidence, they and their influence are as much past history as the Spanish Empire, falling prey to the terminal shift of power in American computer economics from Route 128 around Boston to Silicon Valley over the course of the late 1980s.

By Gary Flood

Now Olsen’s work is truly finished, as DEC becomes a wholly owned subsidiary of Compaq Computer Corp. Olsen, who was described in 1986 by Fortune magazine as arguably the most successful entrepreneur in the history of American business, co-founded Digital Equipment Corporation in 1957 with a business plan that didn’t even mention computers, so unfashionable were they then. The target was to sell $95,700 worth of modules; Olsen did a little better than that, with DEC becoming the 27th largest US corporation at its peak in 1990. It’s now languishing in the lower ’60s, as its only second CEO in its 40-year history leads the battered remnants off into the safe arms of Compaq. Olsen did that mainly off the back of one extraordinary computer, of which only 49 were ever built, the 18-bit PDP-1, which was so compelling that its first customer, fellow legendary, and also now non-independent, New Englander Bolt, Beranek and Newman bought it right off the trade show it was first shown at on the first day for, $150,000 in 1960 dollars. The PDP-1 launched a whole new industry, which came to be called departmental computing, giving information processing opportunities to all those unable, or unwilling, to deal with the mega-corporate IBM mainframe style approach. Olsen’s DEC went on to launch the PDP- 5, the world’s first true minicomputer with the first time sharing OS in 1963, the PDP-11 in 1970, which had shipped over 600,000 by 1989, all of which contributed to it becoming a $1bn operation in 1977, the year it launched the highly successful VAX (Virtual Address eXtension) architecture, which along with its equally successful MicroVAX successor enabled DEC to seem infallible. (Well, apart from the disastrous proprietary micro desktop fiasco of 1982.) Up until mid 1980s DEC could do no wrong; the decade 1979-1989 it was the second most successful company in America in terms of annual compound growth rate, at 21.7%. By the end of 1989 DEC was a $12.9bn company with 124,000 staff worldwide. In the 1980s DEC regularly reported $1bn profits for its full years and could boast of 50% margins and 20% growth rates. Unquestionably the world’s number 2 after IBM, in terms of cultural importance if not actual dollars (remember the Japanse giants), it had a reputation for honesty, engineering excellence, and solidity.

What went wrong?

Then it all went haywire. Much has been written and wrangled over about why Olsen was finally asked to step down by his board in July of 1992, an event as shocking in its way as when Imperial German chancellor Otto von Bismarck was dismissed by the new Kaiser in 1890, an event many see as setting Germany on the disastrous course of launching the Great War. In that fiscal year DEC was so strung out as to be losing $3m a day, or $2.79bn overall. What went wrong? It seems in hindsight that DEC’s whole success was based on what was in its time a revolutionary idea, that of a portable architecture that could grow with the customer’s needs (PDP then VAX) – an idea later copied line by line by Microsoft, which had the fillip of the actual writer of the VAX operating system VMS, Dave Cutler, joining it in October 1988 to essentially do it all over again with NT (as the bitter joke in DEC has it, Why is NT so good? Because it looks just like VMS!). But that revolutionary idea was already crumbling before NT under the guns of both the PC and open systems assault

waves. (Who now, incidentally, would call Olsen a fool for referring to Unix – ironically enough of course, created on one of his own computers, the PDP-7, in 1969 – as snake oil in 1989 or thereabouts? For given the wretched lack of unity in that whole area, which meant it couldn’t and wouldn’t ever really be open, didn’t he have a point, albeit he made it too honestly and openly?) Added to the fact that the ground was changing under DEC’s feet, there was the whole cultural problem. DEC was – and is – an engineering company, in the way that Oracle Corp has always been a marketing-led company. Thus the overemphasis internally on engineering macho and abilities rather than business savvy. Given the fact that DEC engineers sold to end user engineers, this tended to create a closed loop, which for 30 years meant that DEC enjoyed fantastic market pull, never really had to bother with marketing (given the level of trust at that level), which in turn meant DEC sales engineers were always more order takers than need creators. (How astonishing it is in this day and age to recall that its sales force didn’t even get commission for most of its history!)

Slash and burn cultivation

So Olsen left and an Armani-suit/jewel-encrusted ring-wearing young Texan engineer, Robert GQ Bob Palmer, stepped into the breach in October of 1992. And for five and a half years he grimly practiced the slash and burn cultivation that has left so many of us puzzling over what on Earth he was up to. DEC lost 20,000 people in those first three years, as Palmer cut $2.5bn worth of expenses from the bottom line, but the company seemed cursed never to regain consistent profitability, let alone real renewed growth. On the one hand there was the Alpha chip, twice as fast at its debut a month after Palmer became head of the company as any rival 32-bit chip, a technology DEC pushed and pushed and invested in and talked up ceaselessly; on the other the corporation sacked more and more staff and gradually sold off so many of its product lines and business units that it has ended up as Compaq prey little more than a services arm, a semiconductor design shop, with some networking and storage bits and bobs. In just the past year, we have seen Palmer alternately promising good times, then flogging off more of the company. In June it denied reports it was restructuring its services business, than paid EDS Inc $500m to look after its multi-vendor customer services unit (CI No 3,190), shifting 800 employees to EDS. The same month, it abandoned its Alta Vista Internet Software Inc IPO plans announced in June of 1996 (CI No 3,190). Then in July Palmer was optimistic about a return to growth after the four grim years, in the directions of 64-bit computing, Windows NT and the internet; then for its last full fiscal a few days later (CI No 3,211) DEC came in barely profitable, but still with revenue down 10% year on year. In November its network product division went to Cabletron Systems Inc for $430m (CI No 3,298), a move echoing its sell off of middleware to Bea Systems Inc in March. Did Palmer do all this with the eventual plan of making DEC an easy company to offload? In December he was saying that the divestitures were complete, and that having accumulated a $3.2bn cash pile the company was beginning to look for acquisitions, forging ahead into its services-centric future. And these divestitures had by year end included the astonishing surrender of a hardware manufacturer of its very ability to make hardware. How strange that game of footsie Palmer played with Andy Grove now seems. DEC sued Intel in May of last year (CI No 3,160), declaring that the Pentium, Pentim II and Pentium Pro chips were all way too close for comfort with information disclosed (under Olsen’s watch in 1990 and 1991) about the fabled Alpha. Settled in DEC’s favor in October, with it getting $700m for effectively the whole Alpha line, the deal left DEC fabless (CI No 3,277), with Palmer actually having the neck to boast that DEC was better off since it didn’t have to make large investments in semiconductor manufacturing technology any longer! So now DEC is part of a company that makes its living selling NT and Intel technology – just like he was doing? How convenient. But more likely Palmer gave in in the end because he finally saw that despite all DEC’s best efforts the Alpha had turned out not to be the next VAX, but was always going to remain an in-house DEC- only asset, and if the company wasn’t growing neither would the customer base. Take the way DEC swore vassalage to Microsoft.

VMS’ lovechild

There is a suggestion that the much vaunted DEC-Microsoft alliance, begun with so much fanfare in 1995, was actually created more out of the latter’s need to avoid a big legal bill (perhaps as high as half a billion dollars) off Palmer, who at least never seems scared of getting the lawyers on the phone, given the fact that NT really is VMS’ lovechild. Instead, MSFT gave $105m to DEC, promised to love it to death, lent it $75m to train DEC engineers in NT (resulting in the situation where DEC has more qualified NT engineers than Microsoft!), and ended up with two amazing concessions – that DEC would migrate its VMS customers to NT and that DEC would sacrifice another of its children, its office software line (remember ALL-IN-1?), in favor of Microsoft’s Exchange line. Would Olsen have agreed to any of this? No way – but Palmer felt he had to, to try and finally kick start Alpha as a mass market chip. But Gates stiffed him by getting closer to Intel, and HP, who had no proprietary chip issue and therefore whose business interests were all so mutually more inclined (basically, sell NT stuff to nuke Unix). And while all this was going on DEC was in play, despite its protestations to the contrary. We covered the Compaq sale rumor first a year ago (CI No 3,074) but got derided – then in May The Wall Street Journal revealed talks had in fact taken place. In late 1995 Palmer agreed a purchase price of (surprise) between $9bn and $10bn (DEC was then boasting a market cap of only $6.5bn), but DEC managers managed to scuttle it because didn’t want to lose historic DEC name (which under this third deal it manages to do – for a while). Then there was the second shot in the summer of 1996 – that deal would have cost Comapq only $5.6bn then, based on DEC share price, but negotiations ended in September 1996, for whatever reason. Now we’re back pretty much to scenario one, which makes one wonder why Palmer wasted the intervening two years. Eckhard Pfeiffer will now definitely achieve his ambition of making Compaq a $40bn company by end of century; he gets a large, trapped midrange base committed to some kind of mix of Windows NT and Intel chips, an excellent services arm and some funky storage and networking and (lest we forget) Alta Vista stuff thrown in for good measure. Under GQ Bob, with his nearly $4m salary, DEC has gone from a headcount of 125,000 in 1990, to under 55,000 seven years later; it has basically admitted Alpha cannot make it on its own with the Intel deal; it has gone from being a mini-IBM to a slightly bigger EDS with some product stuff; it has gone from being one of the most influential players in the data processing industry to a portly Microsoft VAR; and it has now blinked out of existence after 40 years, of which we can say only the first 33 made any sense at all (i.e. up until 1990, when Olsen started losing the plot). Bob Palmer basically spent over five years getting the company fit for selling, managing to get it to a biggest ever runrate of $14.6bn, when it had already hit $14.4bn in fiscal 1993. Compare this sorry tale with IBM, which has had its ups and downs God knows under Lou Gerstner, but which has never been seen as trying to go any other way than up – DEC turns out to have been going sideways at best. Surely Bob should now go to Apple and help Steve Jobs get that ready for the appropriately upside-generating sell-off? Now all the questions about what he was up to, and whether DEC of 1992 really was truly savable without selling off the family jewels, become irrelevant. Which was a word you could never, ever have used about Ken Olsen’s DEC. What was that line about dwarfs standing on the shoulders of giants?