Oral-History:Mitchell Kapor

From ETHW

About Mitchell Kapor

Photo courtesy Seth Resnick

Mitchell Kapor is the founder of Lotus Development Corporation and the designer of Lotus 1-2-3, as well as an entrepreneur, philanthropist, social activist, and investor. He founded Lotus in 1982, served as President (later Chairman) and Chief Executive Officer from 1982 to 1986, and as Director until 1987. He also served as Chairman and Chief Executive Officer of ON Technology from 1987 to 1990.

The interview begins with a discussion of Kapor’s background and the problems and issues inherent in start-up versus mature industries. He then details his decision, along with Jonathan Sachs, to launch Lotus 1-2-3. He discusses their early competition, the new rules related to new industry, Lotus’ pricing strategy and site licenses, and the use of Lotus’s brand name throughout the 1980s. Kapor also explains the importance of understanding the capabilities of a platform, the challenges of a hypergrowth environment, and the importance of building the business. He describes his own fanaticism regarding product design and quality, his attitudes toward product support and enhancement, and his decision to leave Lotus. Kapor also discusses Lotus’ product diversification and the launch of his second company, ON Technology. The interview concludes with his evaluation of classic management problems of starting a second company and the role of research and development.

About the Interview

MITCHELL KAPOR: An Interview Conducted by Andy Goldstein, Center for the History of Electrical Engineering, 20 May 1993

Interview # 157 for the Engineers as Executives Oral History Project, Sponsored by the Center for the History of Electrical Engineering, The Institute of Electrical and Electronics Engineers, Inc.

Copyright Statement

This manuscript is being made available for research purposes only. All literary rights in the manuscript, including the right to publish, are reserved to the IEEE History Center. No part of the manuscript may be quoted for publication without the written permission of the Director of IEEE History Center.

Request for permission to quote for publication should be addressed to the IEEE History Center Oral History Program, IEEE History Center, 445 Hoes Lane, Piscataway, NJ 08854 USA or ieee-history@ieee.org. It should include identification of the specific passages to be quoted, anticipated use of the passages, and identification of the user.

It is recommended that this oral history be cited as follows:

Mitchell Kapor, an oral history conducted in 1993 by Andy Goldstein, IEEE History Center, Piscataway, NJ, USA.

Interview

Interview: Mitchell Kapor

Interviewer: Andy Goldstein

Place: Mr. Kapor's office at EFF, Cambridge, MA

Date: 20 May 1993

Educational and employment background

Goldstein:

Were there any specific experiences in your background that you found invaluable in preparing you for senior management?

Kapor:

Let me start by saying I was not very well prepared for the management challenges that resulted from the very rapid growth of Lotus. This is managing in a start-up situation of hypergrowth. In our first year of operations we were a $50-million company, and in the next year we were a $150-million company. Nothing really prepared me for that.

But what was helpful--if anything was helpful--was that I had had a wide range of life experiences before this happened, as a counselor in the psychiatric unit of a hospital, as a disc jockey, and as a meditation teacher. For instance, I have a master's degree in counseling psychology. Some of the communications and group skills that I developed in the course of my graduate school and work experience in psychology turned out to be very transferable to the business environment. Some of my sense about how to motivate and work with small groups also came out of that experience. So that was useful. I also went to business school and dropped out of it. That was mildly useful.

Lotus

Development and management roles

Goldstein:

Were you aware that your role in Lotus was shifting? When you were developing the product, were you conscious of a change in your role to more management-oriented responsibilities?

Kapor:

I was conscious of an enormous change in my role due to the changing circumstances of the company. The nature of the management challenge shifted a lot from a very small, tightly focused group of people to a much larger, multi-level, and more complex organization. I consider both of those activities to be management. Some people might say that management only starts when you have a big firm, and that whatever it is you do when you're running things with only three people isn't management. But that seems to me to be an arbitrary distinction.

Goldstein:

Was your technical familiarity with the products influential in your managing?

Kapor:

Oh, completely. I think managers who don't have a firsthand knowledge of the substance of the products that they're making are at an enormous and usually fatal disadvantage. Having firsthand knowledge in the software business means having a technical background and fluency with software. In the start-up mode, if you don't know it yourself, who can you count on? Even in later stages of growth, if you are the CEO and you don't really have the technical background, you are forced to rely on other people. It's just very difficult if you don't know which questions to ask. Mistakes get made that could be really crucial.

Goldstein:

Where did you find that expertise particularly valuable?

Kapor:

Obviously in development it's completely critical. I would even argue that in marketing it's very critical. One of the things you have to take into account when you look at this issue of a technologist as manager is the phase in the industry life cycle of the participants. In the personal computer software industry at the time Lotus was starting and when I was running it, it was a very young and immature industry. The basic ground rules regarding how you went about doing development and how you went about marketing were not well established. In fact, they were being invented. There was a lot of trial and error.

Today it's very different, in that the basic ground rules are extremely clear to everybody because anybody who hasn't followed those ground rules is out of business. There is a different kind of expertise that is needed in leading this kind of company. Back then, you really needed to be very creative in the sense of discovering how to make products, what to put in them and what to leave out, and how to market them. That's pretty well understood today. But there is a fiercely competitive battle for the hearts and minds of customers. The product categories are stable, but the competition is much, much stronger than it used to be.

There is a different kind of expertise that's required to lead that. There is also far more complexity if you are running a company. Today Lotus is on the order of a billion dollars a year. The complexity of coordinating all that with these huge product lines and international operations and thousands of people in the organization is daunting. There is a whole set of management issues in the coordination of complexity that simply didn't exist then, or just existed in a very embryonic form. You need to be a specialist in doing that. Being a technologist isn't particularly helpful. I don't think it's necessarily unhelpful, but it's not an expertise I have in managing this kind of complexity. In fact, I don't like it because, for the most part, it is not hands-on.

Goldstein:

Do you think it's more traditional business school type work?

Kapor:

Yes, sort of. It's the sort of thing that business schools purport to teach people how to do. I don't think they actually do. The point is that my experience at Lotus was in the early days, at which time being a technologist was very important--in fact, critical--because of this need to figure out the ground rules. Without having the necessary technical knowledge, it's inconceivable that you could do that in the start-up phase of an industry.

Goldstein:

The management issues are different in a mature industry?

Kapor:

In a start-up versus a mature industry, yes, they are different. Or in a mature firm in a mature industry. You have two continua of maturity. One is the firm itself, and the other is the industry. You could have, for instance, a young firm in an old industry. The one thing you can't have is an old firm in a young industry. Actually you could if the firm was in some other industry, and they were getting into something new. So it's what I would describe as a two-by-two matrix of the age of the firm and the age of the industry. The interaction between being a technologist and being a manager is really contextually dependent on which quadrant you are in.

Performance; spreadsheet development for IBM PC

Goldstein:

Do you have examples? Is there a disadvantage to not being a technologist in these first quadrants?

Kapor:

Yes. There are a number of key decisions that we had to make, for instance, about the development environment. One place where many firms went wrong was not so much that they had the wrong product idea, but they went about developing it in the wrong way. They developed it in Pascal when they should have developed it in assembly language. Or they developed it on mini-computers and cross-assembled it, whereas they should have developed it more directly on the target platform.

There was a certain wisdom then about good software development and methodology that would have led one to adopt the wrong strategy. In fact, a VP of development might well have suggested exactly the wrong strategy. So when the CEO is a technologist and also understands the business objectives and needs very clearly, then there is a better likelihood of making the proper decision, which in our case was to develop in assembly language right on the target machine for the IBM PC. What was at a premium was performance. That was the way to get the highest degree of performance. It was more difficult to develop that way than in higher-level language, but the performance sacrifices would just not have been acceptable to the marketplace.

A pure technologist might not have understood that it wouldn't be acceptable to the marketplace, and a pure manager might not have understood the magnitude of the trade-off. Whereas if it's one and the same person making that decision, you integrate those two facts or those two constraints together, and say, "well, we've got to develop in assembly language."

Goldstein:

But what is it that led you to that successful perspective?

Kapor:

I knew that performance was paramount in terms of the delivered benefit to users, and I knew enough to know that the way that you would get the maximum performance is by developing in assembly language. In fact, we had been developing the prototype in C and we could just see that the performance wasn't there. This decision incorporated my beliefs, my values, and my observations about what users wanted.

Goldstein:

Were there any special circumstances critical to the launching of 1-2-3 versus the introduction of the PC?

Kapor:


Audio File
MP3 Audio
(157 - kapor - clip 1.mp3)

Yes. I should explain the basic story: Jonathan Sachs and I were playing around with different ideas for an advanced spreadsheet, a graphic spreadsheet that incorporated other functions and had programmability to it. What crystallized our thinking was the announcement of the IBM PC. We had started before it was announced, and in fact we didn't even know it was coming.

It was apparent to me that that was the proper target machine for our efforts because of two factors: One was a marketing factor that IBM was going to be a big deal for business. Two is that it had certain desirable characteristics. The first is that they were going to distribute in Sears and Computerland computer stores, not through the IBM sales force. They had gone to the outside for their processor, operating system, and applications, which I thought was a very smart move. I feel they were violating their own conventional wisdom about how to do products. So I said there must be some very smart people at IBM because they already understand you have to take a different approach to succeed in this market. That, combined with the fact that it was IBM, which would make it very acceptable to business, and combined with the fact that it was a 16-bit processor, whereas the competitors such as the Apple II and the Tandy machine were 8-bit machines, made a big difference. I could see that users wanted to have more memory for their spread sheets because they were running out of room on the Apple II. They wanted things to be faster so that they wouldn't have to sit there and recalculate. It was clear it was the right platform.

What happened is that there were two spreadsheets available for the IBM PC from the beginning. But both of them were basically warmed-over 8-bit versions. They only supported a total memory size of 64K, whereas with the IBM PC you could have 640K. I happen to know that they weren't optimized for the IBM PC. So I had to place the bet that if you came in and did a program that was optimized for the IBM PC--to take full advantage of its capabilities--and if you added the graphics to the spread sheet and put in the user programmability and improved the interfacing, that's a winning combination. That was the hypothesis that we set out to test. In fact, that's actually what we did, and it worked.

Marketing and implementation, Lotus 1-2-3

Kapor:

We also raised what for that time was a significant amount of extra capital--initially a million dollars, and then later some more money--in order to do a big marketing campaign and introduction ads in The Wall Street Journal. That seemed to me to be common sense. If you want to meet the business market, don't advertise in Byte Magazine, instead advertise in The Wall Street Journal. In fact, it was common sense. We just happened to be among the first people to figure out what was sensible and go do a good job executing it.

Goldstein:

If that was common sense, were there any early business decisions that you made that were less certain?

Kapor:

Well, of course there were a lot of people who thought the whole PC industry was a big shot in the dark.

Goldstein:

Did everything seem crystal clear to you?

Kapor:

No. The thing that seemed to be really frightening was that VisiCalc had an iron lock on the spreadsheet market on the Apple II, and there they were on the IBM PC, and it was selling. MultiPlan was also in this. Both of those products were distributed by IBM, as well as by their vendors, VisiCalc and Microsoft. So the idea of having yet another entry in that category seemed to me quite frightening. I said, "people are not going to switch." At least I was afraid that they were not going to switch. We went to a lot of trouble to try to position 1-2-3 as integrated software, not as a spreadsheet. That lasted about a month after we went out with it, because it was clear that it was an enhanced or second-generation spreadsheet. So we let the market reposition us.

One of the peculiar characteristics of the whole Lotus story was how quickly things happened. There wasn't a lot of time in which there was much uncertainty. In other words, I spent about $300,000 of my own money from royalties for a product that I developed for the Apple II. I was running out of money, and I went to try to raise venture capital. I was able to raise it almost immediately. In the context of doing that I wrote a business plan which was ludicrously far off in the revenue aspects. We thought we would do three to four million dollars the first year. But it said we are going to deliver this product, and we are going to do professional marketing, and we are going to run with it. And that's what we did. We actually got the money in April 1982, and we announced the product in either September or October. We went to Comdex in November, and we wrote a million dollars in orders on the floor of Comdex the first time we had gone out. We knew that our estimate of three million for the whole year was ludicrously low, so that's when we started raising the forecasts. We shipped the product on time in January of 1983 as promised, and then we had this $53-million-dollar year. The major challenge, actually, was in implementation. The one thing that we did very well was that we executed. We met all of our commitments and our deadlines, and we shipped. We built a factory that was able to meet demand. I don't know how many units we sold the first year. But I can figure it out. What is 50 million divided by 250?

Goldstein:

Two hundred thousand.

Kapor:

Yes. Two hundred thousand. We also had manuals to print. The biggest challenge was the implementation challenge of having to scale up. We had to build everything from scratch, including whole financial organization capable of registering the revenue and paying out. We hired hundreds of people so there was a human resources function. Then we had our marketing. We were training dealers; we had road shows where people would go out and train dealers. We had customer support organizations. We had everything that you find in a modern personal computer software company. We built it all on the fly during that first year.

Goldstein:

It sounds like it takes a lot of savvy to know where to get the venture capital, to go deal with Comdex, and build a financial organization.

Kapor:

I had had four years of very intensive internship in the personal computer software business. I had been a consultant. I had run a cottage industry software company, published a lot of stuff, and sold it mail order and through distribution. I worked for a Silicon Valley start-up. I had started out as a product manager, which is where I learned a little bit about venture capital. I had as much experience as anybody in the business side of it. There were a lot of people who had worked for Digital for ten years, or Hewlett Packard. But it turned out that most of what they knew was wrong. It was unhelpful in this new industry with new rules.

Goldstein:

When you say wrong, do you mean inapplicable to the new industry?

Kapor:

Yes, inapplicable. I think I got a little lucky with the venture capital. I knew Ben Rosen slightly, and I went to him first. I said, "I'm thinking about doing this thing." So he said, "write a business plan." I wrote a business plan, and they decided to fund it. I had a track record. I had written two products myself: Visiplot and Visitrend, both personal software.

Goldstein:

That's where you earned the three hundred thousand?

Kapor:

Yes. That's right. They were companion products to VisiCalc. In addition to having gone to work for that publisher as a product manager, I had also had this other relationship as an author. They had the publishing contract. It was a tiny industry and I knew all of the principals in it. Anyway, Ben Rosen and his partner L.J. Sevin decided to take a chance on me. In conventional terms, if you looked at my resume, I had no management experience to speak of. Another thing is I had no technical experience. I don't have a degree in computer science. I had a couple of courses. I am a very mediocre programmer. The programs that I wrote and that made money were in Basic because I can program in Basic. I can't program in assembly language well enough to do a commercial product. So viewed that way, why would anybody invest? On the other hand, I had a lot of experience, as much experience as anybody, and am self-taught in terms of being handy around computer work. They decided to take a chance. More power to them.

The rest of what we did, in terms of the launch and the strategy and so on, was just applied common sense. I sat and thought about it. There was a team of people who contributed to it, and we had to make a series of decisions. I don't remember us agonizing over this. In fact, it seemed kind of obvious. It was scary because we didn't know if it was going to work. It was a lot of money. But we just had the foresight and the good fortune to be in the right place at the right time.

Over the years literally hundreds of people have come to me and said, "I have an idea for the next Lotus 1-2-3." I find this pretty amusing because I try to explain to them that we didn't know we were going to be the next Lotus 1-2-3. The idea didn't exist. We had something much more modest in mind. What we did in that kind of hypergrowth the circumstances have to be very special and appropriate to it. You have to be in at the very beginning. It has to be far enough along that the market has potential for explosive growth, but not so far along that somebody else has gotten it. You are either there or you are not there. It's one thing to say, "we were smart; we recognized there was an opportunity." That's true. But I was fortunate to be at a point in my life where I was able to start a company right near the beginning. Not everybody can do that.

Lotus 1-2-3 pricing, distribution, and competitors

Goldstein:

One thing you didn't mention before was pricing strategy. How did Lotus's price compare to VisiCalc?

Kapor:

It was a lot higher. I forget what VisiCalc was, whether it was $195 list or $295. We went out at $495.

Goldstein:

Was there much precedence for software that expensive?

Kapor:

Well, D-BASE had started at $795 and might have dropped by that point. I basically wanted to go with the highest price that we thought was supportable. Believe me, we didn't run focus groups even on this. $495 just felt like the right price. It turned out that it was okay. There's another thing. A fully-loaded machine, the most expensive machine that you could buy, was about $5,000. A minimal machine was about $2500. So at $495 it was ten percent of the fully-loaded machine or twenty percent of the lesser machine. That seemed to make sense as a ratio.

Goldstein:

Were there other decisions about site licenses?

Kapor:

We didn't have any site licenses at the start. That came later. International distribution also came later. We decided to go with retail distribution. We did decide to start a corporate sales force to call on corporate accounts directly, to stimulate volume purchases. That just made sense. That's where the business was. That's what the customers wanted. But we did fulfillment through the retail channel for a long time before we figured out how to sell direct.

Branding

Goldstein:

You alluded to the success of IBM. One factor was the IBM name. That suggests that you believe that brand is an important asset in this industry.

Kapor:

Yes.

Goldstein:

How did you use Lotus's brand name once that became established?

Kapor:

I think our conscious efforts to exploit it were much less successful than the momentum created by the success of the product. People identified the use of the entire machine with the use of Lotus as a spreadsheet. They saw the machine just as a vehicle for doing spreadsheets, which benefited us enormously.

Goldstein:

Yes. The old PC's that have the Lotus part burnt into them and that's all.

Kapor:

Yes, that's right. The market ordered us because they liked the product. The product was extraordinarily useful. It empowered a whole class of people in business who were nontechnical professional people. It gave them a productivity tool for this extraordinarily wide range of uses, anything involving calculations with numbers--not just financial ones. It was a Godsend. The product fully exploited the capabilities of the original IBM PC. We did such a good job with that. You can argue that this feature maybe was a little clumsy or we left something out, but nobody could do a better product because we had just hit the spot. Now, of course, Windows is another deal. But between 1983 and the late 1980s you just couldn't do any better. Between the momentum from being the category leader and the product excellence, it was a self-sustaining phenomenon. When we marketed other products, even though they said Lotus, if they weren't right, they didn't sell.

Symphony and Jazz

Goldstein:

I'm curious about that expansion with Jazz and Symphony. Did those seem like further exploitations of the enhanceabilities of PC's?

Kapor:


Audio File
MP3 Audio
(157 - kapor - clip 2.mp3)

That was the idea with Symphony, but it was the wrong approach. It was based on miscalculation, a misunderstanding. We understood that there was a market for different productivity tools, spreadsheet, word processor, database and a market for integrated software. But the integrated software market turned out to be at the low end for people who don't want full functionality. They want a little of this and a little of that in a single, simple package. We didn't do that. Symphony was an integrated package that tried to be all things to everybody, and it failed by that standard. It failed to create a huge amount of momentum. Over the years it has generated hundreds of millions in revenue, but it didn't have this kind of momentum to propel the company. We misunderstood what it was that the market wanted. This was not surprising, given that it wasn't clear to anybody.

Technically speaking, Jazz just did not work well enough. It did not do enough. Its spreadsheet and the other features in it were not good enough. Jazz was Symphony for the Mac. It had the same problems of product concept, and it had additional problems of performance reliability. Microsoft took the 1-2-3 lesson. When they did Excel, they did an exceptional spreadsheet for the Mac that was tuned to the Mac as a platform. They understood that that was what they needed to do. When they did that and got out there, they owned that category. And they still own it on the Mac.

Influence of hardware on software design

Goldstein:

So you are saying that an understanding of the hardware is indispensable.

Kapor:

An understanding of the capabilities of the platform. Not so much just the bits and bytes of the hardware, but what you can do with it. That is absolutely critical. Also understanding what the market wants. People just want the most kick-ass spreadsheet they can get for the money that takes best advantage of whatever platform it is they are buying. So the Mac had a graph phase, and the PC didn't.

Goldstein:

Would you play to the hardware in appreciation of the power of the hardware above the quality of the software?

Kapor:

No, I wouldn't. Now we're talking about design. We're saying, "How do you design a successful product?" Appreciation of the platform is clearly one thing. Having to write a specific feature set is clearly critical.

Goldstein:

How about on the execution? You know, the code itself?

Kapor:

That has to be first rate also. You can have good ideas on the right platform, and if you do a lousy job in implementation--meaning that it's too slow or it's too buggy or it takes too long to come to market--you will get killed.

Goldstein:

So it all has to be there.

Kapor:

It all has to be there.

Stress of fast corporate growth

Goldstein:

You were talking about the value of the technical background for managers. I was wondering if you brought anyone up in Lotus who didn't have that, who hadn't put in time in the technical area, who was a product of a business school or management experience. Was there ever a time in Lotus's history where you needed people like that?

Kapor:

I have to characterize what happened a little bit differently. We were in something like a war zone. Usually you think of tragedies, like natural disasters, shipwrecks, or things like that--things that produce great unhappiness, as being the source of huge amounts of stress for the people that are involved. But it's also true that hypergrowth and enormous success produce exactly the same types of stress, without being life-threatening, of course. We had the problem of building an organization and building a management team from scratch in a big hurry. I didn't know anything about how to do those things. I had a very mixed record. I built a management team of people that worked pretty well up to a hundred million dollars a year. But we were that size for only eighteen months. The job outgrew everybody. My job probably outgrew me, too.

Finding people to do finance and manufacturing actually was not so much of a problem. While the challenges of a hypergrowth environment were real, the financial functions were more or less recognizably the same as a financial function at some other part of high tech, but not PC softwares. You could take somebody who was good, and they could probably adapt. But marketing and development, key functions, were very difficult. We had a real lack of success in finding people, who could do that well, hiring them and moving them along through the ranks. We tried, but it just didn't work.

Management philosophy

Goldstein:

Would the results speak to the quality of the person? Or was it more a question of whether their common sense coincided with your own sensibilities about these matters?

Kapor:

More the latter. Part of the problem is that if I were doing things again, I might do them quite differently based on what I've learned. My management philosophy, for better or worse, suggests there needs to be a coherence of vision among the senior management. The CEO sets the tone and provides the overall leadership. If it is the wrong leadership, then the board has to fire the CEO. So I would look not for uniformity of opinion, but for the sharing of a sensibility in which there is a lot of individual differences of opinion as people approach subjects, both from their disciplinary specialty and from their own personality. I really think that that is paramount. You have to have the skills within the particular discipline. This was another problem since what it meant to do good marketing of PC software was so unclear. You had this additional problem of how to find people. How do you know what you are looking for once they come? How do you tell? We had those problems also. Now it's somewhat simpler.

But there is a question lurking in this regarding not only one's management philosophy, which I've given you, but one's philosophy of business. What's the point of being in business? People in business generally don't have either the time or, for that matter, the disposition to sit around like a bunch of Socratic philosophers and ask themselves that question. That's part of what makes people business people. In fact, it's one of the things that creates some tension when I try to fill that role because I'm naturally reflective about these sorts of things. I have an interest at heart in creating great products. The business was a kind of vehicle for bringing great products into the world. We were trying to do that. That turns out not to be a great business philosophy because you succeed in business by building the business, not by doing great products. There is a linkage between the two, but often the commitment to do great products and push the edge of the envelope actually interferes with business success in a number of ways that I can see in retrospect.

When you ask what's important, there were things that were important to me then that are still important to me now, but are somewhat at odds with the answers I would give if you were to say, "What should somebody do if they really wanted to build the business?" I would say, "Do what Microsoft does. Don't worry so much about product elegance. Worry about building." The first and foremost question is: How do you build an invulnerable grip in all of the key segments of the marketplace? It may turn out that product quality has something to do with it, but that could be about third on the list. It could also be making the right alliances finding time to market independent of product quality. I just don't think like that. Or, I didn't think like that. I don't know what I would do now. I wouldn't run a company like that because I don't like facing those sorts of choices.

Goldstein:

You talked before about the importance of coherence of vision. I wonder if different cultures developed within Lotus, if you had people there from the beginning versus other people who had different management skills.

Kapor:

It's not quite that simple. But you are on to something. Since we hired so rapidly, we never had a single culture. We had fragments of many different cultures because we hired a bunch of people from Company A, and a bunch of people from Company B. I tended to hire people that I felt comfortable with, which meant that there were a higher proportion of people that were a bit off the beaten track in terms of their life experience. Those people very quickly left or became marginalized as the company grew, and that's too bad. I feel badly about that still. There wasn't a tension between the old and the new. There were quite a few people that didn't have the same inherent love for the substance of the business that I did. Some of the other people did early on, and they came to occupy more dominant positions inside the company. It turned into more of a hard-headed, aggressive type of operation.

Product quality and design

Goldstein:

You were talking about the different things that are important to a business's success. Product quality may not be number one.

Kapor:

No. By the way, I have to say something to that. When I was doing things, I had a fanatical orientation towards issues of product design and quality, end-user experience. It's not exactly fair to say that other firms that I'm implicitly criticizing don't care about product quality. I think they do, but in a more measured way of finding the right set of trade-offs between creating the user experience in the product and other business objectives. I do not want to imply that I thought that other firms besides Lotus do not care about product quality. I just have this fanaticism about it.

Product support

Goldstein:

I wonder where you would position product support in that list, and how your attitudes about product support developed over your time at Lotus?

Kapor:

You need to sell a successful user experience, and some degree of product support is critically necessary. The landscape of product support has really changed as we have moved from a totally uneducated user base to a user base in which there are a very large number of sophisticated users. This is another one of these issues where you need to know where you are in the industry life-cycle.

When we started out, there was not a lot of the infrastructure or user education and support that exists today. Today there are hundreds of books. The best books in the category are much better than the documentation supplied by the company. There are third-party training companies. Large corporations do all their own training, support, installation. We had to do a lot of things because there was nobody else to do them back then. We did them: We trained the dealers, and we would answer anybody's question about anything. At a certain point the demand for support outstripped our ability to provide it. So the phone lines would be busy. That situation was eventually ameliorated as users got more knowledgeable and people started supporting themselves, and then third-party support of various kinds kicked in.

Goldstein:

Were you ever in a position where you had to make one of those trade-offs between support and enhancement of product?

Kapor:

What happens is you build up a big installed base, and you become a prisoner of it. From a business point of view you can't add new features that enhance the power of a product just because they're great e.g. ones. Users don't want it. They say: "This is too much. We are not prepared to accept things at this pace." Or, "we don't care. We are only interested in the six basic things." So there is a lot of tension there. I didn't like that.

Lotus divisionalization

Goldstein:

I read in some interview you did in 1985 that you had just begun divisionalizing Lotus. I'm wondering what was behind that move.

Kapor:

We need to have the history on the record. In mid-1983 I hired Jim Manzi as director of corporate marketing. He became director of marketing in the fall of 1984. He became president when I was still chairman and CEO. In the spring of 1986 he became CEO. Then in the summer of 1986 I left. In the 1984-1986 period he assumed increasing amounts of management responsibility. Divisionalization was his initiative. That was the one way you got some leverage and some scale. It was by taking the same basic product and technology and serving different markets in the different divisions. I don't think that really worked out. There was an Engineering & Scientific Products Division that just did not work. The products were not right and weren't delivered. That divisionalization scheme went away. I don't even know how they have it divisionalized today. It's been through multiple reorganizations. This is a classic management problem of how to deal with complexity, and one way you do it is to divisionalize. I have to say that I'm just not very good at understanding how one ought to structure a large organization. I'm a start-up kind of person.

Goldstein:

Was it tough for you to relinquish control at various stages?

Kapor:

Let me tell you what it was like. As I said, the whole thing was like being in a war zone the whole time. It was enormously stressful, very exciting, and with a great deal of gratification and success. I was really afraid of screwing up a lot of the time. As you can gather from what I've been saying, I was also extraordinarily ambivalent about running the place. Ultimately I decided I would be happier if I wasn't. This is not a decision I have regretted at all. I tend to be a perfectionist, and I can be very control-oriented. If things are not the way I think things should be, then I get very unhappy. I became very unhappy because I was unable and unwilling to run things, and unhappy with the results of not running them. So it was a great personal dilemma.

Decision to leave Lotus

Goldstein:

You are providing two contexts for 1986. One is the maturation of the business in the industry, a different place in the life cycle. The other is your personal position and control in Lotus. Did those both contribute to your decision to leave?

Kapor:

Yes. I felt that there was just too much responsibility I couldn't sleep at night. I think I felt overly responsible for the whole thing. That certainly didn't help. It wasn't my ambition to run a big company. I wanted to do this great product and make a big business out of it. But I didn't find the positive parts of running this big show to be very gratifying. I mostly felt that there were a lot of people who wanted a lot of things. This was whether I was a CEO or not. I was just being the chairman and the figure head and so on. They wanted things all the time, and it was just a pain. I like to be left alone to do my own thing. But instead, I was a prisoner of the spreadsheet. It just got to be this big monster. I could see that Lotus was going to be the company that was going to be dominated by the spreadsheet for many years to come. I like spreadsheets, but not to the exclusion of everything else. It's like being told you can't eat anything but Chinese food for the rest of your life. Who wants to do that?

Product diversification

Goldstein:

Did you try to diversify Lotus's product line?

Kapor:

We tried hard, but in a strategically stupid way. One thing that Lotus did that other companies did not do is systematically experiment in new types of products and bringing them to market. The only one that really looks like it's going to work is Notes, which is going to become the cornerstone of the business. But I did Agenda, which got great critical reviews and has fanatically appreciative users, but which was not a commercial success. Then there were all these other products. There was stuff we acquired from Software Arts. Metro was a pop-up TSR. And then Lotus Express; we had an early e-mail package that was a smart front-end for MCI mail. We had Symphony, we had Jazz, and we had lots of things that never made it to market. We had product after product after product. In fact, we were known for this. People would say, "great ideas for individual products; no coherence." We did Manuscript, which was an outline-oriented word processor. If we had been simpler and less ambitious in the product spacing--let's just try to do a decent word processor, let's not try to reinvent the category here--and if we'd started on that early and just kept improving it slowly we would have done much better as a business. That was the applications strategy at Microsoft. As it is, Lotus got into the word-processing business by acquiring Samna. Omni is very good and gets the best reviews, but it's competing against the entrenched competition of WordPerfect and Microsoft Word. It was strategically very weak, but creatively very inventive.

Goldstein:

Did you introduce all these products and try to diversify the way you did because of your instinct?

Kapor:

It was because of my instinct and my desire to put out good products.

Goldstein:

Or, did you conceive of it as a wise business strategy?

Kapor:

No! The former. I didn't have a clue, really.

Goldstein:

It was just this compulsion to make good stuff?

Kapor:

Right. This is not atypical. This happens.

Highlights of Lotus career

Goldstein:

I read that you like the Electronic Frontier Foundation because it is like the fun days of Lotus.

Kapor:

Right.

Goldstein:

Can you elaborate on the differences between the fun days and the days that were less fun?

Kapor:

It is just what we have been talking about. When it's small and you are running it you have this nice team feeling and you can know everything that's going on. You really feel like you are making a difference. Things happen quickly. There is a lot of novelty in it. When things get to be bigger every move you want to make has many consequences, both internally and externally, that slow down the pace of things. There are a lot of things that people tell you you can not do or should not do because it's going to have some negative effect on the business. Part of the problem was that I think I listened too much to people who didn't see the world the same way that I did. I didn't trust my own instincts.

ON Technology; challenges of starting second companies

Goldstein:

Would you characterize ON Technology as a start-up?

Kapor:

Yes, it was.

Goldstein:

How did the experience in starting it up compare with Lotus?

Kapor:

It was a struggle. I made a series of fairly classic mistakes when entrepreneurs start second companies. I was overly ambitious and did not have a sense of appropriate limits. Even though I was aware that those were risks, and even though I took those risks seriously, I still made the mistakes. You try to do too much. Once again, I wasn't willing to be pragmatic in the business sense. I had this vision of product purity in mind. We raised a lot of money on high valuation, hired a lot of smart people, and then... I'm also a pragmatist. That's the problem. I can only fool myself for so long. Some people can fool themselves forever. I realized it was not going to work anymore. I didn't have the fire in the belly to do the system software, the next great operating system. It wasn't my real expertise, and I did not want to compete with Microsoft. We tried changing the strategy to an applications strategy, but I had to fire a lot of people, and it was just a mess. I really lost my will. I shouldn't have started another company. I had done everything that I was going to do. A lot of it has been trying to overcome the bad start.

Goldstein:

Could you explain what you mean when you say the classic management problems of starting a second company?

Kapor:

Many people who were successful with a first company and start a new company want to prove to themselves and to the world that the first company wasn't a fluke. Since they were the big success the first time, they feel they have to be an ever bigger success the second time to prove it was not a fluke. But that interferes with letting things take an organic course. Starting something to prove a point is not a good strategy. Now look at NeXT. Steve Jobs started NeXT. Black is the color of revenge. He was thrown out of Apple and then went and did things that were just self-defeating. That is really what I mean.

Goldstein:

You refused to learn the lesson of big companies that you had left.

Kapor:

Right. He also did something that was overly ambitious and not market-focused or pragmatic. It didn't work.

Lotus workplace environment

Educational support for employees

Goldstein:

What kind of continuing education did you provide at Lotus? Was that an issue for people who worked there?

Goldstein:

Yes. For the employees.

Kapor:

We had a tuition reimbursement plan that was pretty generous. People picked their own courses if they were job-related. Eventually the company developed a bunch of seminars and other things. But believe me, we were just scrambling to keep up with things.

Quality of work life; employee qualities

Goldstein:

What qualities did you think were important in the employees? As you expand, you have to bring on a lot of people.

Kapor:

While I was there, I was very concerned with the quality of work life, seeing that it was a good place to work and that people were treated fairly. We made a lot of investment in the human resources function and in programs to support that. What I found was that at our rate of growth, we attracted a lot of opportunistic people who were there who really were not the best for the company. We do the kind of quality control did not in terms of taking people in, and we got a culture that was a culture of entitlement to an extent.

Goldstein:

What qualities were you looking for?

Kapor:

What we really needed were people who were smart, fast, and humane.

Customer relations, focus groups

Goldstein:

How did you try to use ties to either academia or the government? Did you make alliances?

Kapor:

We really did not do much of anything with government. There was nothing formal with academia either. We were just being entrepreneurial.

Goldstein:

What were the decisions about the products that were made? How did you set yourself to keeping an ear to the ground?

Kapor:

In the earliest days my intuition and my personal knowledge was really what we used, and it worked well. We tried to build in a lot of customer contact to understand what customers wanted both informally and formally. We made site visits and had focus groups. I think that is the right thing to do. Unfortunately, a lot of what we heard was not what we wanted to hear, and this created various conflicts. We were not strictly dedicated to doing what was going to make the customers happy. That would have been a much better business strategy. We had these conflicts between a vision of doing things with new kinds of products versus what customers wanted. That was never successfully resolved.

Management on a large scale

Goldstein:

Do you have any comments just on general management lessons?

Kapor:

Managing small scale and managing large scale are two very different things. When you can know everybody, whether that is five people or a hundred people, you can lead by personal influence in a way that is very difficult to do when things are larger. I admire what Bill Gates has done at Microsoft by staying very closely aligned with the key developers--probably a couple hundred of them. He has enormous influence throughout the company, determining its product direction and its strategic direction. I think there are huge portions of the Microsoft empire which he just leaves completely untouched, that are managed through other people. I have a different product philosophy and a different set of values. You have to admire the kind of job that he has done in maintaining this effective span of control of this enormous empire. This could be focused in on empowering the developers and staying closely aligned with them, and driving the business well around that. That has worked really well. There's a real lesson in that. It's the only way I know of scaling things.

Research and development

Goldstein:

What was your attitude about research and development at Lotus?

Kapor:

I loved R&D. I would have been happy just to do R&D. A lot of products we put out were really R&D. That is where the interest in new ideas are, and that is where the smart people go.

Goldstein:

Did you try to integrate or incorporate technologies from other areas into you R&D? I'm thinking particularly of artificial intelligence.

Kapor:

Yes we did. It helped Agenda to an extent. We looked at a lot of that. Some of it went into Improv. We tried to hire people with a bit of a research background and who were current with that kind of stuff. When I was there, I always tried to get developers who were interested in producing products, not people that were interested in writing papers. We were not big enough to afford that.

Goldstein:

I only have a weak sense of what R&D is like in a software company.

Kapor:

Apple and Microsoft and their advanced technology groups, will play with things like speech recognition, and they will try to get a project to demonstrate that it's feasible to do a certain type of thing, such as having useful speech recognition for some class of applications. Then they will do a project that is a prototype to demonstrate and show them. So that is less research and more advanced development, where you take some technology, whether it's AI or speech recognition or some computer graphics technique, and you try to take it the next stage, towards pragmatism.

Goldstein:

Does that lead inevitably to a vertical integration? Or do you have to get involved with hardware?

Kapor:

Not necessarily. You might, but you might not. Often especially with the improvements in microprocessor performance, you can run many things with off-the-shelf hardware. It's just that particular algorithms and data structures that are used in AI are a thing unto themselves. If you do not have people who understand that world and what it's trying to do and how those techniques work, you can not possibly use that stuff.

Goldstein:

Would you prefer to shy away from that kind of vertical integration? Or does that seem interesting?

Kapor:

This is very interesting.

Goldstein:

Did Lotus do any of it?

Kapor:

You mean with hardware?

Goldstein:

Yes.

Kapor:

Not really.