MPM: Ken, in the mid to late 60s you knew my company, Applied Data Research, and you knew my face because I came to Maynard many times on ADR business. In the early 70s, after I joined the company, you came to know my name as well, and you had a rough idea of what I was doing in the Larry Portner organization, Portner having been your VP of Software Development at that time You and I never exchanged more than a few words, it's simply that we were in many meetings together from time to time over a four year period. Now, regrettably, you're four years dead so there's no hope of my actually interviewing you for this site. I have my own take on your management philosophy and I'm going to have you speak to me from that viewpoint, okay? With me couching your imaginary remarks in my writing style, okay?
KO: That's fine, Mike. Just remember that I'm not here to defend myself.
MPM: Okay. Now ... How shall I put this? ... ... I don't think you had a radical management style at all, not if looked at from the correct viewpoint. I believe that you were simply playing venture capitalist within DEC. The product lines were the companies you were funding, and service groups such as engineering and software development were the customers of those companies, paid by the product lines, not by you. You didn't much care what the products were. You were more interested in creating an environment in which other people could have good product ideas, develop them, and then take them to market to see whether the outside world wanted to buy the products in sufficient quantity to make those ideas profitable.
KO: Go on.
MPM: The mythical matrix management system then becomes the normal relationship between a consulting organization, its employees and its clients. The consultants are employed by the consulting companies, but in a very real sense they work for the clients, who pay the bills and then call the shots regarding budgets, schedules and specs. This is especially true if the consulting assignments span years rather than days. Inevitably the consultant has two bosses -- one who provides the funding and therefore calls the shots, and the other who supervises how the work gets done -- the account manager, if you will.
KO: That's exactly right, Mike. The product lines and matrix management are two sides of the same coin. Now, I want to go back to something. Don't forget that when I founded DEC my fundamental orientation was the engineering of digital modules and digital computers. I don't think I'd have done the KO equivalent of Ben & Jerry's Ice Cream, not because it wouldn't have succeeded but rather because it would have bored me to death. I wouldn't have wanted my name on anything that wasn't somehow related to electrical engineering, and it wouldn't have been as much fun for the kinds of people I enjoyed working with. Anyway, go on ...
MPM: So you like electrical engineers and wanted to work with them?
KO: Of course. I'm an MIT-degreed engineer myself and have never forgotten my roots. The world may see me as a big time industrialist, and some even called me the ultimate entrepreneur, but on the occupation line of my personal tax returns I always put "engineer". That's all, just "engineer". Without interruption electrical engineering has been the fastest-paced area of technology development since Michael Faraday invented the electric motor in 1821. If you want to change the world for the better and have fun doing it, become an electrical engineer.
MPM: I have to note, Ken, that finance is simply the engineering design of profit-generating machines that run on money but are made of people, goods and services. It's a different branch of engineering, but one that applies to all businesses in a capitalist economy, and the skills of the discipline are portable across industries. Adam Smith noticed this back in the 1750s, that once you rise to a certain level in a company it no longer much matters what line of business that company is in because you have become a financial manager who, in principle, could run any company. Smith also spoke of "projectors" -- people who undertake projects -- and today we would call them venture capitalists. Anyway, I will speculate that your VC orientation is why, in the 60s, Jim Murphy remarked to me that at the beginning of every quarter you would say to the product lines, "I'm laying out all this money for you, paying your bills ahead of time. By the end of the quarter I need you to have paid me back, with a profit."
KO: You've got it, Mike. I was in fact playing venture capitalist. The product lines were substantially independent businesses. I required them to make a quarterly profit, but as long as they were able to do that I kept my distance. I wanted to know what was going on but I didn't want to interfere -- not unless people weren't making their numbers. The product lines were free to fund the in-house service groups, but they were also free to buy things on the outside.
MPM: So you tried to set up free market competitions even within the company?
KO: Of course. Market forces are much more effective at running things than dictatorial management can ever hope to be. I was simply doing inside DEC what General Doriot had been doing from the outside with his American Research & Development venture capital firm -- creating a portfolio of businesses that stood a good chance of being highly successful in financial terms. A few might fail individually, perhaps even most of them, but there still would be a few outstanding winners that would carry the whole organization.
MPM: You mean like the PDP-8.
KO: And the PDP-11. And the PDP-10. These were all ground-breaking winners, don't you agree? And they were all designed by shirtsleeves engineers, not a corporate staff.
MPM: Absolutely. But let me ask you this: If bottom up entrepreneurship is so important, how do you explain the success of IBM, which was and is as regimented as companies come, exactly the opposite of DEC?
KO: I don't have an explanation for IBM. If you have one, I'd like to hear it.
MPM: I do have an explanation but no, you're not going to want to hear it. My position is that your product lines ideas-flow-from-the-bottom-up business model works well up to a certain company size. Beyond that point the business must become top-down, and when it does you need product planning done by people who make a career of it. I'm talking about your nemesis, Marketing, as it's taught in the business schools. IBM had and has top-down planning and execution. DEC lost control of its business when VAX took over the company beginning in the late 70s because its planning was still bottom-up when what was needed was top-down.
KO: Hm-m-m-m. I'm going to have to think about that. However, I will say that Gordon Bell's insistence that the product lines be disbanded and that all company eggs be put in the VAX basket made me very nervous. But isn't that what you're advocating, that past a certain growth level you need a central vision imposed from the top? Wasn't that what Gordon was doing? Why was DEC unable bring that off?
MPM: As I see it DEC eventually failed because Gordon left the bottom-up Old Boys Club design culture in place while telling the Operations Committee that his stable vision of a simple 11/78 cleanup of the 11/70 was being implemented with a sense of urgency. Since technology never stops evolving, under Gordon there was never any such thing as a product concept freeze, only an endless round of one-pluses with nifty new features dreamed up by the hardware and software people. There was endless gold-plating with wonderful new technology wrinkles, and therefore with cost targets that were missed by a mile, and with schedules that always slipped badly. He said as much in a 2006 interview with New Yorker Magazine -- that he always goes where technology takes him, as far as he can, as fast as he can, period end of discussion.
KO: You know what? I finally noticed that myself. The early VAXes (I refuse to say VAXen) had an operating system but not many fast new peripherals to go with their increased CPU horsepower. And when new mass storage peripherals finally were ready, the processor was out of date -- too costly and too slow. By then we had an all-VAX strategy but everything was late, overpriced and under-performing, the complete opposite of what Old DEC had stood for. The VAX business takeoff was a slow motion horror show for years, with expensive systems unable to ship because certain key peripherals had not been released to manufacturing. I kept pressing Gordon to Do The Right Thing for the customers but he never did figure it out. I guess that's why he had a heart attack and then left in 83. He could no longer take the pressure. He had no idea how to regain control of the company's engineering process, which had become unable to deliver products either on time or on spec, much less on budget..
MPM: IBM would have approved the basic 11/78 idea that the Operations Committee was told about, then frozen the concept and begun detailed design, then frozen the resulting hardware and software designs, then implemented the committed written designs, and then had a painless manufacturing and sales rollout of the CPU, the peripherals and the software. All the while the marketing folks -- the product planners -- would have been cross-checking the stable product definition against their ongoing market research and ongoing talks with customers. Only if there were major divergence between market reality and what was planned would there be any significant design changes much less product definition changes.
KO: I don't like to structure things that way. That's not me. That's not DEC.
MPM: No, but what you did like simply did not work past the point where the product lines were dissolved. Old DEC -- the DEC I joined -- worked. New DEC -- the VAX DEC -- the DEC I left -- was badly and irretrievably broken even well before VAX sales peaked in the early 90s.
KO: So are you saying DEC should have been more like IBM?
MPM: No, I'm saying that if you're going to play in IBM's arena then you have to have formal product planning. If you can't do that, or if you don't want to do that, then you should get out of IBM's kitchen. By the way, I didn't care for IBM either. I had the opportunity to go to work for them directly out of college. I had no interest in that and instead became employee number twelve at ADR. Yet I knew and respected their capabilities, which were formidable, a story for another day.
KO: Well, if your much-loved product planning is so important, why didn't IBM succeed in the PC business? Wasn't the IBM PC a planned product?
MPM: In effect IBM started the PC business, legitimized it, and yes it was a planned product. However, they did it out of a Florida-based skunk works that was only loosely connected to IBM headquarters. The PC division was enormously successful for a while, just as DEC succeeded in the VAX business, for a while. Yet both companies were eaten alive by their competitors, who pursued a relentless product price minimization strategy that was of no interest to IBM and no longer of interest to DEC. And when Don Estridge, the IBM PC skunk works division manager, died in a plane crash, the corporation took the opportunity to crush this rebellion against IBM culture once and for all, leaving the mainframers and their obsolescing products firmly and eventually almost fatally in charge..
KO: I think I see what you're getting at. The PC marketplace cared about IBM compatibility, but it had no particular loyalty to IBM as a vendor. If a personal computer was IBM compatible at one third the price then it was perceived as being better than the IBM offering, and the business flowed away from IBM to the clone makers.
MPM: Yes. Precisely. Now what happened with DEC and the VAX?
KO: Following your train of thought, a competitor who wanted to knock off the VAX would have to build a VAX clone. Well ... No ... Not quite ... In fact, the ideal product would be a clone of the VAX software running on a clone of the IBM PC, at a rock bottom price.
MPM: As you said earlier, please go on.
KO: And that's what Dave Cutler did when he went to Microsoft -- he reverse engineered his own VMS operating system, porting it to the Intel hardware base, which eventually allowed Bill Gates to graft the Windows shell onto the result. No wonder DEC couldn't compete. But this was later. Windows NT didn't really get going till the year 2000 and by then DEC was in serious trouble. WinNT was simply the coup de grace. What was the fatal shot?
MPM: Go on.
KO: If the PC marketplace was driven as much by price reductions as anything else, then the DEC software licensing fees for VMS would have stood in the way of DEC having a real presence in that marketplace. Ditto for University of California at Berkely's version of UNIX and DEC's version, ULTRIX. In all these cases -- VMS, UNIX, ULTRIX -- the vendors of this software all priced themselves out of the PC market. They were fighting yesterday's war of the operating systems, not recognizing that the battleground had shifted to system price minimization.
MPM: Go on.
KO: Hence Linux. It had just enough real time capability to allow internet routers and servers to be based on x386 platforms, absolutely gutting the VAX-based server/router business. Hm-m-m-m. I guess when we told the world that VAX was "culturally compatible" with the PDP-11, we shot ourselves in the foot. Whatever DEC brand loyalty there had been going into VAX, the new incompatible instruction set killed it off. Yes, a whole new generation of customers arose, but they weren't loyal to DEC, they were loyal to decent performance at low system prices. When DEC ceased to be the price/performance leader, they took their business elsewhere.
KO: Even so, you can't tell me that VAX was a bad idea. For goodness sake, Mike, there are still VAX clusters in operation that were built twenty to thirty years ago, and they're still doing mission critical work in banking, in stock exchanges, and in certain military and national security installations. To this day nobody has matched the failsafe characteristics of a properly configured VAX cluster. Isn't that capability an important societal contribution all by itself?
MPM: That's true, Ken. But the computer marketplace punished DEC anyway, not because VAX was a technical failure -- it wasn't, VAX on Alpha was a technical masterpiece -- but rather because, in the end, VAX was a financial disaster. Look. I'm not going to convince you, but you're not going to convince me either. Let's talk about something else.
KO: Okay. For example, do you want to talk about the financial disaster?
MPM: Yes. Why didn't you support layoffs when they became necessary? Why did you make the board of directors fire you? And that decision was unanimous, wasn't it? Why couldn't you get with the program?
KO: I'll start by asking you a question, Mike. I lost track of you around 1974. Did you get fired internally?
MPM: Yes, that very year. And that was just the first time.
KO: How many times in all were you fired internally during your stay at DEC?
MPM: Four, maybe five, depending on how you count.
KO: With any of these firings did you ever fear that you were about to be put out on the street? That when someone said "I want you off my budget in three months" you would have to leave the company?
MPM: No ... Well ... Yes and no ... It crossed my mind as a distinct possibility but I always figured that when one situation at DEC ended I would be able to find another within the company in a relatively short time. My career would change direction but in the end I was not worried about my family's welfare because of income interruption. These assumptions turned out to be true, every time.
KO: So, Mike, ask yourself this: How would you have reacted if I had forced you to leave the company in 1974?
MPM: I'd have hated it, Ken. Quite frankly I would not have known what to do right away, or even how to cope. DEC had been my spiritual home beginning in 1965, and my work home beginning in 1971. I would have found it very difficult to be kicked out of your home.
KO: And there you have it, Mike. I would not have fired you in 1974, or 77, or 78, or 81. I would never have fired you unless you had done something blatantly dishonest or criminal. Would you like to know why?
MPM: I think I know why, but the readers want to hear it from you..
KO: I would not have fired you or anyone else because employees come first in any business that effectively has my name on it. The welfare of DEC employees was more important to me than the financial well-being of the stockholders, including my own financial well-being. Therefore I would not -- could not -- be a party to any layoffs. None. Never. If I had been I would not have been able to live with myself. The board did what it had to do, but so did I and I feel good about that. I never broke the sacred unspoken trust of lifetime employment, and nobody can take that away from me. I knew we had to cut back but I urged employee buyouts, which is a totally different thing since they are voluntary.
MPM: I'm sorry to say that I could have done DEC layoffs, Ken, even though the organization you created was kind to me on several occasions. I respect your position but I was once faced with a similar situation in a business that had my name on it and I did not hesitate to act -- to fire people. May I tell you about that?
KO: Yes, but I never said that mine was the only moral or ethical position. I said that it was the only position that I, Ken Olsen, was capable of taking without violating my life principles, which I was completely unwilling to do. I think your wife shares some of my life principles though not necessarily all of them. You should ask her about this stuff.
MPM: I did consult with my wife at the time I fired people. I explained to her that we were operating at an unsustainable loss rate, that I couldn't fix the problem in time, that we had to get profitable immediately, and that to accomplish this we would have to lay off 80% of the staff right away. I explained that a two-day financial analysis of the business situation showed me that the choice was not between layoffs and no layoffs, it was between a) laying off 80% of the staff now and being a profitable though much smaller business, versus b) laying off 100% of the staff in two months and then filing for personal bankruptcy.
KO: And how did matters work out?
MPM: Cutting staff was the hardest thing I've ever had to do, yet I did it and I didn't hesitate. I made the cuts immediately, giving what little severance pay I could. I didn't rest till I had actively helped every fired employee find a new job, at least those who wanted the help, which a few people refused. Faced with the same situation today I would do the same thing because I feel that hanging on to the bitter end would simply have delayed other people from getting on with their lives while at the same time bankrupting my wife and me, which would not have helped anybody. However, let the record show that my wife and I never took a dime out of the business. I wasn't protecting my own pay because there was no pay to protect.
KO: Always do the right thing. Only you know what that is.
MPM: Yes, to thy own self be true. Ken, that's a wrap, and thanks. Thanks for everything. By the way, I haven't worn my DEC badge in almost thirty-five years but I still have it and it's a prized possession, number 14477. I would happily do it all again without changing a thing.
KO: You're very welcome. And I wouldn't change a thing either.