Digital Equipment Corporation - A Case Study
August 2007 (The New Business Road Test)
DEC founders Ken Olsen and Harlan Anderson set out in the late 1950s to provide functionality similar to large mainframe computers – mostly IBM’s, in those days – but in a smaller, more bare-bones machine. In 1959, the company came out with its first computer – the Programmed Data Processor (PDP-1). Olsen described this computer as a console ‘with all the instruments and lights, very much like you see in a power plant’. The PDP-1 cost the customer $125,000–150,000. By 1965, DEC had sales of $6.5 million, with profits of $807,000.
In 1966, DEC started selling the PDP-8. While considerably less expensive than its predecessor, each of these machines still sold for $18,000. DEC marketed the PDP-8, with its high-quality video display terminal, to businesses, universities, newspaper offices and publishers. It was also a particularly attractive computer for third parties who bought the PDP-8 machine from DEC, customized the hardware and software to meet the needs of their customers, and sold the enhanced computer as their own product. DEC’s third-party business soon accounted for 50 per cent of its sales. By 1970, DEC was the most successful minicomputer manufacturer in the world.
Through the early 1970s, DEC remained a leader in the minicomputer industry. Olsen said, ‘For many years we made the same two computers, the PDP-8 and the PDP-11. We kept that design consistently so that software the customers wrote would continue to work on newer models and the software we wrote would continue to work and get more and more robust’.
Quite deliberately, rather than join the competition for the PC market as it emerged in the late 1970s, DEC avoided it and concentrated on networking issues. ‘We made some PCs designed to be part of the networking but the general PC market was not for us. There were too many people in it . . . You could build them in your basement. That was not for us’. The VAX was DEC’s product line that connected several minicomputers in a LAN. One of the company’s most popular networking products was the VAX 8600, a product that allowed a system of mini computers to function like a mainframe. But targeting the mainframe market, with its sales trend heading south, flew in the face of the rapid growth in the capability of PCs.
Finally, in 1980, DEC did begin developing personal computers, but Olsen insisted that the new machine be called an ‘application terminal and small system’ rather than a PC. ‘We believe in PCs. We encourage them. We network them. We use them in large numbers. But we still believe that most people in an organization want terminals. Terminals you don’t have to worry about data management, you don’t have to worry about floppy disks. You just sit down and it does the work for you automatically’.
DEC’s late decision to enter the PC market, and to enter with three different product lines (Rainbow, Pro and Decmate), proved both confusing and damaging. In 1984, the effects began to show. In the third quarter of that year, earnings were down 72 per cent from the previous year. And that was only the beginning.
In 1988, Sun Microsystems introduced computers that ran the UNIX operating system. Hewlett-Packard soon followed with its own UNIX-based Apollo computers. All these systems had far more computing power than DEC’s minicomputers and were much less expensive. Moreover, they ran on UNIX, which was rapidly becoming the de facto standard in operating systems, thereby encouraging third parties to write innovative software that would run on these platforms. Ironically, much of the UNIX software was developed on DEC machines. DEC, however, had been doing so well with its proprietary VMS operating system that it gave its UNIX offering little support. As UNIX took hold, no longer were DEC’s minicomputers, with their proprietary operating system, the best alternative. They were no longer in the race.
By the dawn of the 1990s, DEC found itself in dire trouble. Tens of thousand of employees’ jobs were lost. By 1994 DEC, once 126,000 people strong, was a company of only 63,000. Finally in 1998, DEC, by then no longer a computer maker, was sold to Compaq.
What did DEC miss?
There were many things that DEC did right in its heyday. It fared well for a quarter of a century – a veritable eternity in the high-tech industry. And, while it did post some impressive financial results along the way, it was plagued repeatedly by an inability to stay in front of the technology curve, missing the mark on some sweeping trends.
In the late 1960s, a group of DEC engineers led by Ed de Castro was assigned the task of designing a 16-bit computer that would replace the then-current 8-bit technology. Their final plan contained a basic 16-bit system that could be grown to 32 bits as well as a series of compatible products that would allow users to upgrade their existing machines rather than replace them. But what de Castro’s group was suggesting amounted to scrapping the entire DEC product line and replacing it with the new 16-bit machines. DEC’s management soundly rejected it. So in April 1968, de Castro and two other engineers left DEC, raised their own venture capital, and started their own company, Data General Corporation, to produce 32-bit computers. By 1969, Data General was one of the hottest new companies in minicomputer manufacturing, tapping a market that could have been DEC’s.
Then in 1972, a DEC team working on the PDP-11 recommended that DEC develop a product that combined a computer (the PDP-11/20) with a terminal and a printer. According to the PDP-11 group, this ‘Datacenter’ would appeal to a broad market of individual users, including scientists, technicians and others in administrative positions. DEC’s leadership rejected this individual computer idea. Had DEC pursued the datacenter, could it have been the PC pioneer? We’ll never know.
By 1980, with Apple and other personal computers beginning to make waves, and a year before IBM’s PC introduction, DEC’s product managers, those individuals that face the customer, suggested that DEC begin to play in the personal computer space. Olsen and his team refused. The rest is history.
Why did DEC repeatedly miss key changes in its marketplace? It’s difficult to know with certainty without having been in DEC’s meetings or inside Ken Olson’s head. The contrast with Virata, however, is striking. Virata had extensive connections up, down and across its value chain. When Virata got new information, it fanned out its other connections to help it interpret what it had heard. DEC, too, may have had some of these connections, but if it did, its top management wasn’t very good at listening to them or lever- aging other connections to take advantage of the information those connections provided. DEC leadership, like the ostrich, had its head in the sand.
The vibrancy of DEC’s connections was perhaps encumbered by DEC’s focus on and belief in its own technology and its faith that its solutions were superior to others. ‘They believed [their] operating system was simply the best and would remain so into the new millennium’, said Jean Micol, a former DEC marketing executive. If this is the case, why bother to develop connections for keeping track of external developments. Call it corporate arrogance or simply naïveté. DEC missed 16-bit computing. It missed PCs – not once, but twice! It missed UNIX. And now DEC is gone.
Markets and industries do change – especially high-tech ones. Success in changing markets requires well-developed connections to keep abreast of the changes, and it requires a top management team that’s open-minded enough to consider changing course when conditions so indicate. Had Olsen spent time talking to and building a wider set of informational relationships – with DEC’s sales channel and distributors, with its OEM manufacturers, even with its own marketing department – then he would have heard the resounding push towards PCs in the early 1980s and towards UNIX in the late 1980s. Hindsight suggests that the DEC team simply wasn’t up to this task. Are you?
Lessons learned from DEC
As we have seen, DEC failed to adapt to trend after trend in the computing industry: 16-bit computing, the rise of PCs, and UNIX. The problem for DEC was not that they had no connections or that no one in DEC saw these things happening. Indeed, some did. Of the three outcomes listed above that the right kind of connections can deliver for an entrepreneurial firm, DEC’s difficulties seemed to be with the third issue, i.e. obtaining a broadbased assessment of these developments from a variety of perspectives outside the firm. As a result, DEC’s decisions to not pursue these developments in a timely and aggressive manner appeared to have been based on DEC’s blind faith in its own products and solutions – arrogantly and naïvely, some would say – rather than on the basis of the marketplace evidence that was there to be seen and understood.
An inward looking culture, especially in a rapidly changing industry like computing, adds additional risk to what we’ve seen is an always-risky game of entrepreneurship. As Andy Grove, long-time CEO of Intel, wrote, ‘only the paranoid survive’. The same is true for those assessing new opportunities. Yes, this means you. Being inward looking, focused on your idea rather on the market and industry where it might take root, and focused on building the right team to help you achieve your dreams, is a pathway to impending disaster. Having a broad set of the right kind of connections – who you and your team know – does matter, not only in running your business once it starts, but much earlier as well, in assessing and shaping your opportunity and developing your business plan. Don’t go too far without them. Entrepreneurial success is not just about what you know. It’s about who you know and your ability to use your network productively.