Mitch Kapor founded Lotus Development with
Jonathan Sachs in 1982. Their spreadsheet
software, Lotus 1-2-3, quickly surpassed
VisiCalc to become the new industry standard.
VisiCalc had been the original "killer app"
for personal computers. Kapor was a VisiCalc
product manager at Personal Software when
he wrote VisiPlot and VisiTrend, companion
products to VisiCalc. He left to found Lotus
just as legal conflicts were distracting
VisiCalc's developers, and the arrival of the
IBM PC opened a window of opportunity for a
better spreadsheet. Lotus 1-2-3 could handle
larger spreadsheets and added integrated charting, plotting, and database
capabilities. It became the killer app killer.
Lotus went public in 1983. Kapor served as president and CEO from 1982 to
1986, and as a director until 1987. IBM acquired Lotus in 1995 for $3.5 billion.
Kapor cofounded the Electronic Frontier Foundation (EFF) in 1990 and
now leads the Open Source Applications Foundation, a nonprofit that promotes
the development and adoption of open source software.
Livingston: How did Lotus get started?
Kapor: I bought an Apple II in the summer of 1978 because I had become
obsessed with personal computers and just had to have one. I didn't know what
I wanted to do. I very quickly and fortunately started generating some consulting
income, writing programs for individuals who had bought them, like an
ophthalmologist who wanted to use it in his practice and an investment analyst
who wanted to look at stock market data. And I met other people in those days
that had Apple IIs, because it was very much a hobby phenomenon. Several of
us started an Apple II user group called New England Apple Tree.
89
Mitchell Kapor
Cofounder, Lotus Development
6
CHAPTER
One of those people was Eric Rosenfeld, who was a graduate student in
finance at MIT. As a favor to him, and because it was kind of a challenge, I
helped write a statistics routine that ran on the Apple II that he could use to
analyze data in his dissertation. It took me a weekend. He actually had to
explain the math to me; once he explained it, I understood the math.
Afterwards, we kind of realized, hey, this might actually be useful to other
people if we built a statistics and graphics product on the Apple II. It was called
Tiny Troll after something called TROLL, which was a time-sharing thing
at MIT.
At the same time, Dan Bricklin and Bob Frankston were developing
VisiCalc, also in Cambridge, and when it came out, it set the world on its ear. It
was far and away the most useful piece of software ever done for a personal
computer. It was incredibly innovative. It started generating sales of Apple IIs,
and it was a cut above everything else.
The authors of VisiCalc were Software Arts. The publishers were Personal
Software, which then changed its name to VisiCorp somewhere along the way.
I knew the VisiCalc authors because they came to the meetings of the Apple II
user group that I had cofounded, and that's where I first saw VisiCalc in probably
1979.
They introduced their publisher to me--this is Dan Fylstra and Peter
Jennings--and they said, "We would like you to take Tiny Troll and rewrite it
and clean it up so that we can bring it out as a companion product to VisiCalc."
They wanted to have more offerings since they had such a hot product. And I
agreed to do that. I still had a partner, but I think he was probably beginning to
teach at Harvard--anyway, he was otherwise engaged. I was at business school;
I decided, when this happened in November 1979, that I needed to learn about
business because that's where the market was going to be.
I thought I was just going to clean up this little product over Christmas
break so I could finish my education. I would make some money and that would
be that. And I only thought that because I was totally ignorant about how long
things took. I had no background in computer science. I was self-taught--I was
still writing in Basic. I had no management experience; I was in business school
at the time. In fact, I had spent my years after college as a radio disk jockey on
a progressive rock station. I was a transcendental meditation teacher, and a
mental health counselor at a psychiatric unit of a community hospital. That was
OK, because there wasn't really a personal computer software industry. It was
still kind of a hobby thing becoming a business, and nobody really took this stuff
seriously, so I wasn't ludicrously unqualified by the standards of the day.
But I was wrong about how long it was going to take to do this thing. I was
inspired to want to do a really great job by VisiCalc, which was so much better
than anything I could ever write. But I said, "I want to try to do something that
could stand up well." And I faced a difficult decision because school was starting
again. I took a leave of absence from school to finish the product.
It then came to be the spring of 1980, and I thought I was done, and I wasn't
done. I didn't know what done was with software. I had, roughly speaking, an
alpha version of the product--it had some demonstrable features. I decided
90 Founders at Work
that I needed to totally rewrite Tiny Troll to be much better, and give it a totally
new user interface, and so on. Based on that misunderstanding of what done
was, I said to the publisher, "What I want to do is to come out to California"--
which was where they were--"and you should hire me to be your new product
manager. I can finish this thing in my spare time. It's almost done."
Now why did I want to go out there and be the product manager for the
publisher? There were a couple of reasons. The main one was that I had come
to understand that the big economic opportunity was to get stock in a startup,
and this was a way of doing it. I had a royalty contract--like a book contract--
and they said "fine." So I moved out to California without my program having
been completed.
So now I had gone from writing and rewriting Tiny Troll, which eventually
was called VisiPlot, to being product manager for several versions of VisiCalc--
not the flagship Apple II version, but the other versions. I worked for the publisher,
for Personal Software, with the Software Arts people. And a number of
things transpired. I was in California for 6 months and had no time to work on
getting my own products finished. But I found it incredibly fascinating to be in
Silicon Valley and learned a lot.
Personal Software had brought in venture capital just before I arrived, and
while I was there they brought in more management. The VCs brought in more
senior management from places like Intel, and I was moved aside. I could see
that my power and my access were being marginalized, which I didn't like, and
I didn't feel that the business was being conducted with the degree of integrity
that met my standards. And we had actually never consummated this swapping
royalties for stock. So I said, "You know what, I'm going to go finish the product
that I promised you. Let's unwind this." And I moved back to Boston and then
I finally finished the product. It took another 6 months.
They brought it out in the early part of 1981. And it started generating a
huge amount in royalties right away--a huge amount relative to what it was. It
generated about $100K a month in royalties, but I had essentially no expenses,
so that's a lot of money.
Now, all of a sudden, I had options about what to do next. In the course of
developing VisiPlot, I had come to certain conclusions. And there was one
other factor: somewhere while all this was happening, I had worked in assisting
the VisiCalc guys in devising a way to exchange data between VisiCalc and
VisiPlot. That was important because it provided a way to actually make graphs
out of spreadsheet data, which was an obvious piece of functionality.
Bob Frankston had developed something called the data interchange format,
and VisiPlot was one of the first other software applications to support it.
I'd worked with Bob on that--he played the lead role, far and away. But while
there was a way of moving data between these two programs, it was really
cumbersome. There were no hard drives in those days. Everything was on
floppy disks, which had limited capacity. And furthermore, VisiCalc had a
copy-protected floppy disk to prevent piracy. So if you wanted to make a graph,
you had to boot up VisiCalc, you would make your spreadsheet, and then you
would save a file in this special data interchange format to the second floppy
Mitchell Kapor 91
drive--you had to have a second drive because you couldn't save it on the first
drive. Then you had to quit out of VisiCalc completely and then start up
VisiPlot and then read in the file and then you could see the graph. If you
wanted to look at another graph and you hadn't saved the data, you had to
repeat the whole process.
I could think of several ways to make this process less cumbersome, one of
which would be to put both programs on a single disk. I raised the issue with
the guys at Software Arts who did VisiCalc and they weren't interested in it at
all. In fact, at various times I raised a number of ideas with the publisher about
combining the programs and they weren't interested at all.
Why weren't they interested? The people who did VisiCalc had serious
technical backgrounds and a bunch of computer science training. They knew
what they were doing and they had the hot product. I had no credentials or
background, at best sort of minor league success. So I don't think they really
saw me as an equal. And the publisher was even worse in my view, in that they
were firmly convinced, between the venture capitalists and the people they
brought in, that they knew how to build this thing into a big business. And they
saw me, when I was there as a product manager, as an annoyance--as a marginal
person without experience or credentials who was kind of a pest. And I
suppose I was kind of a pest.
So they had no interest in doing more stuff with me. They were trying to
figure out how to technically get rid of me. And I took advantage of that fact.
I didn't like it, but I took advantage of it. The royalty rate that VisiCalc was getting
and that I was getting was very high. My royalty rate was 33 percent of their
gross margin. And VisiCalc's was higher--they got 35.7 percent. At the time the
contracts were done, the economics of the business, which was a new business
of packaged software for PCs, was not well enough understood to know that
that was obviously an insupportable thing to do. But it quickly became apparent,
because huge sums were flowing back to the authors, but the publisher was
the one that was incurring very significant expenses for support--which was
their responsibility--and all the marketing and sales expenses. Anyone who is
familiar with the business understands that royalty rates adjusted downward
pretty quickly.
So here's the way the world looked to me at the time: I have a hit product--
not a number one product, but it's making money. And it has an insupportably
high royalty rate. I no longer work for the publisher, but I know how they think
and I'm uncomfortable with them. And I know they don't want to work with
me. So I felt what I should do is to have them buy me out. They would get control
of the code, close out the royalty stream, and I would go do whatever I was
going to go do next. I saw that was in everybody's interest. And that's in fact
what happened. They bought us out for $1,200,000, so I made a whole bunch of
money. I had never made more than $14,000 a year--I told you what kinds of
jobs I had. We had taxes to pay and I had a partner to take care of, but I wound
up with $600,000, which I divided into two piles. (I'll talk about that in a
minute.)
92 Founders at Work
The non-compete was the hinge issue. I'd been thinking about what I
wanted to do next and in fact had hired Jonathan Sachs, who was the person
who architected and implemented the original version of 1-2-3. We had the
basic concept in mind, which was an integrated spreadsheet and graphing program
with other stuff. They bought me out 6 months after we started, which
was in November '81, and Sachs had started in the summer of '81. We didn't
have any code. We were considering a bunch of different ideas. It was still very,
very early, but I knew I wanted the ability to go do this thing.
I also knew the publisher wasn't going to do the buyout if they didn't have a
really strong non-compete. But remember, I had done a graphics and statistics
program, not a spreadsheet, and I proposed that they carve out an exception in
the buyout to do this integrated graphing calculator program, betting that they
would be sufficiently motivated to get the deal done that they would look at this
thing and go, "This is a very big ambitious thing. We don't really think he has
the ability to pull this off. This gets us what we need, and for the sake of getting
the deal done, we'll sign off on it." So basically, I told them what I was going to
do, taking advantage of the fact that I didn't think they would take me seriously,
because I know they didn't take me seriously. And that's what actually
happened.
It just goes to show you shouldn't underestimate people. You shouldn't
judge from appearances like that.
Livingston: So now that you were free and clear, what were the first things that
you did?
Kapor: Jon had implemented spreadsheets previously; he was one of the few
people. And that's how I knew him. But he had made the mistake of being in a
business with technical people and no business people. He had been at Data
General, and the first spreadsheet that they implemented was for the Data
General minicomputer. Well, there was no market for that.
And then Sachs and his partner were sort of going, "What do we do now?
This didn't work." I forget how I ran into Sachs, but I convinced him to come
workor my fledgling little thing. Remember, I had the royalties. He had some
ideas; I had some ideas; we succeeded in spite of ourselves.
I was so convinced that VisiCalc had a lock on the market that I had to convince
myself that we were going to do something that wasn't fundamentally a
spreadsheet. Of course, what we did was fundamentally a spreadsheet, but the
self-deception I engaged in wasn't sufficiently damaging to be fatal. But there
was a big push to call it integrated software, to add other capabilities, to wrap
other things in it.
The galvanizing event was when IBM announced the IBM PC in August
1981. It was very important in the history of PCs because it legitimized the
whole field--because of IBM's imprimatur. Until then, the personal computer
hardware companies were Apple, Tandy, and Commodore. IBM was the first
"real" computer company to come out with a PC, legitimizing it for the business
marketplace. And that was not lost on me.
Mitchell Kapor 93
So we decided to bet on doing something for the IBM PC, which proved to
be one of the reasons why we were successful. They had decided to outsource a
lot of the key elements of what they were doing, right to the distribution.
Rather than selling it just through their own sales force, they were selling it
through retail stores like Computer-Land and Sears, which at the time was a
very radical idea. They had gone to Intel for the microprocessor; they had gone
to Microsoft for the principal operating system. And I said, "They're smart.
They realize they don't understand this business, so they'll go to the best
people. They're not going to have a lot of ego, and this is the way things are
going to work." Also, they had put a 16-bit chip in the machine with greater
memory capacity. And memory capacity was an enormous issue.
The Apple II had 64K--not megabytes, kilobytes--of memory. It was tiny.
And not all of that was available. Actually, if you wrote programs on the
Apple II, you started with a 48K memory space. So the programs were tiny and
the user data was tiny and people were building spreadsheets that exceeded
memory. It was a fundamental limit of the Apple II, because it was an 8-bit
microprocessor. IBM used a 16-bit microprocessor and I said, "Ah, this will
permit people to build bigger spreadsheets." The memory space of the IBM PC
when it came out was 640K, 10 times the size. So I said, "16-bit, faster processor,
more memory says IBM. We should target it. We should build a product
that is optimized for it."
Now, the IBM PC came out day one, August '81, with a version of VisiCalc,
and with a version of MultiPlan, which was Microsoft's spreadsheet, but neither
of them took advantage of the full capabilities of the IBM PC. In particular,
because they had been put under a lot of pressure to get a product out, they had
taken the code for the 8080/Z80 Intel/Zilog processors--8-bit code--and
tweaked it a little bit. The point is that VisiCalc on an IBM PC still ran in 64K
of memory. You had 640K available, but you couldn't address it in a spreadsheet,
so it was as if it wasn't there. And I said, "This is really an opportunity
here."
Plus another factor: because I knew all of the individuals, I knew that
Software Arts and Personal Software were fighting with each other over the
royalty rate. And I knew that they were essentially distracted and they were not
working together, and I knew that Personal Software was hiring its own developers.
I felt guilt-ridden about coming out with a product that was going to be
competitive with VisiCalc, so I did my best to pretend to myself that it wasn't
going to be competitive. I ultimately said to myself that the fact of the matter is
that I didn't create this opportunity, they did. If they had been on the job, I
would have gone and done something else because the opportunity wouldn't
have been there. But I saw a gap in the marketplace and I said, "We should do
something that lets you do bigger spreadsheets, that's faster, that takes full
advantage of the IBM PC, that integrates the graphing, so you could hit one
button to get a graph"--because I knew people wanted that--"and have a
better user interface for non-expert users"--which we did--"and allow user
customization and user programming"--which we did in the macro language.
So there was a set of ideas that gave 1-2-3 its character, that really made it a
94 Founders at Work
second-generation product, that had sufficient differentiation that was immediately
visible when you demoed it, and that was what gave it its market entree.
Being at the right place at the right time also helped. The business world
was poised to adopt personal computers. They were reasonably priced and they
did something useful, which turned out to be Lotus 1-2-3. So the market just
expanded dramatically, far faster than anything any of us in the company would
have imagined.
Livingston: When you demoed it, were there parts where you knew people
were going to go "wow"?
Kapor: Yes, I think the one-button graphing in particular, and the speed of the
calculation. VisiCalc users loved VisiCalc; they just wanted it to do more. And it
didn't. And when we showed that this did it right out of the box, they went, "I
get it." I used to get applause doing demos all the time.
This was all so new then, in a way that was recapitulated in the early days of
Netscape, the first time people saw a web browser, web content; the first time
people looked at Amazon. So we had our version of that in the '80s.
Livingston: I read you spent 10 months programming it. Did you program it?
Kapor: No, Sachs did. He wrote virtually all of the code of the original version.
We came out with it in January '83. He started working on that code base probably
in October of '81, so that would be 14 to 15 months. All written in assembly
language, for speed. This was the fifth time he'd implemented a
spreadsheet, so he was pretty good at it at this point.
Livingston: Wasn't VisiCalc written in assembly language too? Why was Lotus
faster?
Kapor: Because they were writing for an 8-bit machine, and they didn't take
advantage of the 16-bit architecture in a whole variety of different respects. We
just had more optimized code. And we had a different recalculation algorithm.
We were the first spreadsheet to do something called "natural order of recalculation."
If your spreadsheet had forward references in it, VisiCalc would take
multiple passes over the whole thing to calculate stuff, but we did one pass
through the entire formula chain, and as long as there weren't circular references,
it would calculate properly. So it was much faster for certain cases.
Livingston: Was the code tuned to the IBM machine?
Kapor: It was tuned to the Intel 808X 16-bit architecture. And Sachs was also
very, very good. He was just an artist at high performance with limited resources.
I didn't know how good he was; I got lucky. I knew he was good, but he was a
genius at this sort of stuff. The two of us together was essentially 1 + 1 = 3,
because I had a vision about the product and very strong ideas about the feature
set and the user interface, and he was generally willing to let me drive
things at that level. He had the responsibility for the technical architecture and
implementation, but I'm actually quite technical, so I was able to talk with him
to fully understand a number of the issues and limitations and modify the
design in a way that was consistent with what we could actually do. So we had a
critical mass of knowledge between the two of us that neither of us had alone.
Mitchell Kapor 95
Livingston: What went wrong?
Kapor: A number of things went wrong or almost went wrong. I almost ran out
of money. Lotus 1-2-3 wasn't the only idea that we had. I had done this thing
with some other people called Executive Briefing System for the Apple II that
was like a precursor to PowerPoint. We did some other projects; I had hired
another group of people and basically had spent down the $300,000 that I'd
allocated. It was almost gone and we were nowhere near product, because of
doing all these other things and not having done this before.
I had $600,000 after taxes and paying my partner, and I divided it into two
piles. I took half and said I'm going to buy a house. It was $89,000, the least
expensive house in Cambridge--this was in 1981. I said that I could live on
$40,000 for at least 5 years. So I had the other $300,000 that was my own seed
money, but I almost ran out.
I got lucky in that Ben Rosen at Sevin Rosen decided to invest. He was the
only VC that I pitched (I didn't understand anything about venture capital).
And that was fortunate, because without him, I don't know what we would have
done.
Most of my mistakes came after we launched the product, not before--after
we started shipping in January of '83. I had no significant experience in building
an organization or building a management team. And I intuitively did well
when I was leading the whole team, but once we got past 25 people, you can't
do that. And so I made a series of classic mistakes in hiring. And not building a
good middle management structure. And not recruiting a board that could help
me build the company. Big mistakes in picking a successor, big mistakes in having
an undisciplined product strategy--I was much more interested in having
distinctive, innovative products and thinking about what would make sense for
a product line for our business overall--and big mistakes in expanding too fast
and not having discipline about what we were doing. So I give myself a C or C-
on all that stuff.
Livingston: You guys grew to 1,000 employees before you went public. Did you
know you were going to go public when you started?
Kapor: I didn't know when, but this was what I'd learned from my time in
Silicon Valley. To be honest, here's what I was driven by: I wanted to do really a
great product. Almost from day one I understood that I was passionate about
the applications themselves, that they'd be integrated, easier to use and be
powerful. They'd help make people more productive and I cared a lot about
that. The other thing I wanted was financial independence. I had an enormous
desire not to be dependent on other people, or to have to have a job. I wanted
to dictate the terms. So I knew if you had an IPO, then you had a liquid currency
and you had the ability to cash in and get that.
So I actually pushed for an early IPO, which we did successfully. But that
brought all the usual problems. The main problem we had as a very young public
company was that people did not understand the industry or its dynamics
and therefore they consistently misvalued the stock and misunderstood what it
was about. Because it was new and it was different. Eventually, people figured
it out, but I was very impatient.
96 Founders at Work
I did not set out to build a big company. I actually wanted to be a software
designer. I saw having a company not exactly as being a necessary evil, but there
wasn't a good alternative. My experience had convinced me that being a program
author and having somebody else publish it wouldn't give me enough control
over the process. In Hollywood, the very successful directors like Steven
Spielberg quickly understood that they also needed to be producers and have
their own studio in order to retain control. It was a similar thing.
There were some other funny things about it. In the '60s, when I came of
age, business was not a cool thing. We were all counterculture people with long
hair and sex, drugs, and rock 'n' roll. It was the '60s; I have the pictures to prove
it. I don't remember any of it, but as someone said, if you can remember the
'60s, then you weren't there. But it turned out I also have some entrepreneurial
talent. It's not surprising--my father was a small businessman, my grandfather
was a small businessman, it kind of runs in the family. But I think I had cultural
biases against seeing it or valuing it that took a while to get over. So while Lotus
was getting started, I just saw it as a vehicle to doing great product. I never
wanted to have a big company.
Livingston: The word "creative" comes up a lot when you do a search for Lotus.
Did you make a conscious effort to have a creative atmosphere at the time
when programmers were considered dull and nerdy?
Kapor: Yeah. I was interested in really cool products, so I guess that's where
that came in. I had a very unconventional background and really no interest in
building a button-down business culture. And I'm not an engineering geek,
either. These types of companies tend to reflect the personalities and interests
of their founders. Microsoft is very much cast in Bill Gates's image; and Apple,
Steve Jobs; Borland, Philippe Kahn. And so we tended to have more creativity
and innovation.
The other thing that I cared about a lot right from the very beginning was
creating a workplace that treated people well. At Software Arts, they felt I had
attitude problems. I didn't respect authority. I basically thought, "The people
that are running this are stupid and they don't listen to me and I don't like being
here, being told what to do." It was a mixture of keen insight and adolescent
emotions that I carried for a very long time. So when I unexpectedly found
myself running this high-growth successful software company, the thought of
making it be the kind of place that I would want to work at and different from
all those other places was incredibly appealing.
There were some other key people there who shared that feeling and I
think I probably hired them. And so we did all sorts of very progressive things
with the corporate culture. We invested in the human resources function extensively.
We surveyed all the employees annually on quality of work-life issues,
and took what we heard very seriously. We had a corporate values statement
that wasn't just on a piece of paper. We actually at one point tied a portion of
the managers' bonuses to how well their direct reports viewed them exemplifying
the corporate values. I made every single manager get on the support lines
and listen to customers, no matter what function they were in.
Mitchell Kapor 97
When I was running Lotus, we never had a single employment discrimination
lawsuit; we had a whole bunch of different alternative dispute resolution
conflict management approaches, through the employee relations function.
And then we had a diversity committee that had out gays and lesbians on it--
this was in 1984. We were the first corporate sponsor of an AIDS walk. We had
a corporate philanthropy committee in which the employees actually made
decisions about where the money went, not the pet projects of senior management.
So for many people what was memorable and important about Lotus was
that it was the best place they ever worked.
The other thing to say is that because I lost control of the company--I felt
overwhelmed by what I had created, did not know how to step up to it, put
enough brakes on, hire the right people and be collaborative--I wound up
jumping ship and leaving pretty early, in 1987. And my successor, a very poor
choice on my part, did not share the same vision or values and he wound up disassembling
most of what we put in place. So it was a bittersweet sort of thing.
It was ultimately not sustained. Learn from that, too.
Livingston: Can you remember anything else that surprised you?
Kapor: Oh, almost everything. I didn't expect to find myself in this situation. I
really didn't. Being successful surprised me enormously, shocked me, especially
the magnitude of it. VisiPlot was a success and I had made some money, but I
didn't understand how big the industry was going to get; how big we were going
to get.
Our original business plan called for $3 to $4 million in sales. Ultimately, in
1983 it was $53 million. So it was a 1,700 percent forecasting error. And then it
tripled the next year to $150 million. I was totally unprepared for the magnitude
of the success and the rate of growth. It would have been psychotic to say
it was going to get that big that fast. Or you'd have to be prescient, but I'm not.
So mainly what I was thinking in those days was "We'd better make sure
that we don't blow it, having gotten here." And worrying that it could all fall
apart as quickly as it came about. So I was terrified! Inwardly. And excited. And
unprepared. I became a minor league celebrity in Boston, being recognized in
restaurants, and that was weird. And people started to act differently around
me, because when people are seen as having power or they're seen as having
some special resources, people get weird because they project their fantasies
onto the person or they start telling you what they think you want to hear. If you
watch people around Sergey Brin and Larry Page from Google, it's very amusing,
but, to be the recipient of that... I wasn't particularly prepared for, nor did
I want most of that. I mean, I liked attention, but it's a lot to get used to and a
lot of it made me profoundly uncomfortable.
And there was a series of values challenges that came up with running a
business that I was unprepared for that were very painful.
Livingston: Can you describe one?
Kapor: Lotus as a company wound up suing some other companies that
were copying our look and feel. Now, that was not done on my watch. I was
98 Founders at Work
transitioning out. But I was actually still on the board at the point when we
voted to bring the first two suits, and I voted in favor of the suit out of company
loyalty--a decision that I regretted the next day and have regretted since then,
because I felt that it was an inappropriate use of copyright law to try to prevent
someone from making a product that looks and works the same if they develop
it independently.
I was really torn about how to handle this, and all my net worth was tied up
in Lotus, so it was kind of a mess. There was too much, too soon, and not a lot
of time to grow up in and not a lot of mentoring. There weren't elders or people
to learn from who had been through it whose values I shared.
Livingston: Who were your mentors?
Kapor: Ben Rosen for a while in some respects, but then he made his money,
cashed out, stepped off the board and went on to other things. And plus, he was
not a business guy, he was an investment analyst. So there were some people
that were somewhat helpful, but nothing like what I would have liked or what
exists today.
I try to give back now and help other people try to sort through stuff. I'm
also much clearer about my values, and have been for quite a while now. I think
business at all costs is just wrong. I think there are certain things that you just
don't do, and that acting with integrity and decency in business to me is just a
given. I simply don't compromise on those things. When a person has those
types of values, you have to be careful what type of project you undertake,
because as soon as you undertake a project and you have those values, you're
just going to be so conflicted you won't know what to do.
Livingston: What advice do you give to people who want to start startups?
Kapor:It depends on what type of advice they want. You can't tell people what
they don't want to hear because they won't care and it's just a waste of breath.
And everyone comes in with some kind of agenda.
I like working with entrepreneurs who have a compatible set of values and
are inspired by a vision and are passionate about some piece of disruptive technology--who
are going to create something that actually has value for people in
a way that can be a game changer. That's sort of my sweet spot. But every project
is different, so the specific advice needs to be customized.
The most important thing for me is, I don't want to work with someone who
says, "Just help me make the business be more successful." I want to work with
entrepreneurs who are personally passionate, committed, and believe in what
they're doing. Not all entrepreneurs are like that. Some people may be just as
happy selling canned tuna--"Just show me an opportunity where I can make
money and I'm going to go do that." You think Mark Cuban really cared about
what they were doing at Broadcast.com? This is not to criticize him as a businessman--I'm
just observing--but I don't think he had a fundamental passion
about that business. There was an opportunity, he saw it, he built something, he
sold, and he cashed out at the right time.
Mitchell Kapor 99
Livingston: When you were developing Lotus 1-2-3, you always had working
code. Wasn't that type of incremental development very much ahead of its
time?
Kapor: Yes. I think Sachs and I saw it the same way. We figured out a number
of things very early on that became conventional wisdom. I'm not sure that we
were that much smarter than anyone else. But we had the requisite smarts and
we were in the right place at the right time. So this developing close to the target
machine and having code running, it seemed obvious once you looked at it
that there were enormous benefits to it. The reason it wasn't obvious is that the
machines were barely powerful enough to do development on.
Typically, the systems you develop on could usefully consume a lot more
computational power than what you might actually need in the delivery platform,
and so that's the argument for doing the development not on the target
platform. But there were corresponding disadvantages because it tended to
produce bloated--not optimized--code if you're using a cross-compiler or
cross-assembler. And optimization of limited resources was the name of the
game when you were talking about a 64K machine. So what worked in the minicomputer
world--the techniques and best practices--just didn't work for
microcomputers. I shouldn't say that, because Sachs did come from the minicomputer
world, so that's not right. But a lot of the conventional wisdom was
just wrong and that's what saddled a lot of people. They just didn't think things
through from first principles.
And always having something running was a Sachs thing, because it was just
his experience it was a good thing, and I saw it and said, "Yes indeedy, we
should do this," long before extreme programming.
Livingston: Was there ever a point when you wanted to quit?
Kapor: After we shipped and the business felt like it was getting out of control,
yes. The most fun parts were from time-equals-zero till 1984. I was terrified
about stuff--how's this going to come out, what's going to happen?
I did almost walk out. We raised a second round of venture capital, which I
think, if I had been more sophisticated about business, we wouldn't have
needed to do. We could have just borrowed the money. We turned cash flow
positive so quickly. If I had been a little bit less risk-averse... but that's another
story.
We got to the closing for the second round and they had a very sharp lawyer
on their side--our lawyer wasn't so good--and all these things were happening
at the last minute, all these onerous terms, and I got up and said, "I'm not going
to do this. I don't have to do this today. We don't have to close here and I'm just
not going to agree to this. I'm gone." And they backed down completely on
their onerous terms.
I was just pissed off about this for a long time. These were supposed to be
our investors, they were supposed to be on the same side, but they were highly
adversarial and totally willing to take advantage of us. And saw absolutely nothing
wrong with that. I don't really like conflict, was a conflict avoider; it takes a
lot for me to get up. And I really was going to get up and go home and we really
weren't going to close.
100 Founders at Work
Livingston: You weren't bluffing?
Kapor: No, I wasn't bluffing. I was prepared to take whatever--run out of
money or find financing elsewhere. My attitude was: that's the wrong way to do
business. I don't care that that's the way the world works, it's wrong. That is the
way most of the business world works, but sometimes you just have to stand up
and say, "Not on my watch, not here, not this way."
I think there were these minor problems--the Blue Sky clearances hadn't
come in from some of the states--and they wanted me to personally take the
liability. The investor didn't want to take any risk. It was absurd. They only do
this because they can get away with it, because they have the money and you
need it and "fuck you." (I hope that goes in the book.)
It's just wrong, but the fact is that when the VCs do their deals and they do
the paperwork, they take advantage of entrepreneurs who haven't been through
this before. They do things on terms that favor them in a way that really can't be
justified--that take advantage of their ignorance. It's not a good way to do
business. Some of the VCs try to rationalize it, "This is just the way things are
done." Well, I'm sorry but they're wrong. Why do you think venture capital also
enjoys a reputation as "vulture capital?" It's not an accident; it doesn't have to
be that way.
Livingston: Did you try to change this when you joined Accel Partners?
Kapor: Yeah. And I think that a more nuanced version of what I was just saying
would be that there are contradictions inherent in the venture capital business
because there are significant aspects of what VCs do, including Accel, that are
collaborative with entrepreneurs, and there are other aspects that are not. I
thought that Accel was more different than I ultimately concluded they were.
But I don't think that they were worse than everyone else. There are norms and
practices that cut across individual firms that are really problematic. So I tell
people, "Know what you're getting in for. Here's the way it works."
If the VCs were more transparent and disclosed stuff so that entrepreneurs
could make a choice, that would be better. They wouldn't have to change the
terms, just disclose them and explain what they mean, and what's likely to happen.
But they don't do that. They see it as a negotiation in which having information
that the other side doesn't have gives you an advantage. It gives an
advantage in terms of that individual negotiation, but if you're trying to form a
genuine partnership where you have repeat encounters and you withhold critical
information in the first and most important one, you're undermining longterm
collaboration.
Why should they trust you? What they've demonstrated is that you are
going to act in your own self-interest at my expense if you know better than me
about something and you don't feel under any obligation to share that. That's
actually not collaborative. But it's completely standard.
You know why VCs are like this? It's not that they are bad people; it's the
limited partners. And who are the limited partners? Our great institutions--
Harvard University, Stanford University, UC Berkeley. So if you want to point
up the chain of accountability, when those people stop measuring performance
Mitchell Kapor 101
just based on the return numbers, things could change, because they'll change
the incentives.
Livingston: What would you tell an entrepreneur to understand before he/she
meets with a VC?
Kapor: I try to explain how it works. There are more choices nowadays for
people--angel money, for example. And many things are much less expensive
to do now. You can go further on your credit card than you could before. I want
entrepreneurs to make informed choices when it comes to financing. Understand
what the impacts and implications are for different financing options.
Livingston: Plus, many people don't need to have as much money to get something
started.
Kapor: You can also do some interesting things in a seed round of $100,000 to
$200,000 and it's available on very different kinds of terms.
Livingston: Did you ever do anything to seem more impressive to investors?
Kapor: I'm pretty terrible at artifice; I don't play poker for that reason. But
there's one thing I did. When we were raising money, I hadn't heard from the
VCs (Ben Rosen and L.J. Sevin) for a long time, and I was worried. So I got a
call from L.J. (he's from Texas)--"Mitch, I'm in town. Would y'all like to get
together for dinner tonight?"
So I made a reservation at the fanciest French restaurant in Boston and
raced home to change from my jeans to a suit, and we came to dinner. I ordered
a very expensive bottle of wine, and I knew he was paying for it, so I was kind of
stepping up here like, "This is serious, so I hope you're serious." I wasn't feeling
like French restaurant, three-piece suit, expensive wine. And he's making small
talk through the appetizer course. I was thinking to myself, "If he doesn't get to
the point when they have the main course, I'm going to ask him, "Are you doing
this thing or not?'" Because I knew that we were out of money. And finally at
the end of the appetizers--about 45 minutes, but it seemed like all night--he
said, "Mitch, Ben and I would like to invest in your company. How much do
y'all think it's worth?" And I dropped my fork, like a cartoon.
Livingston: How much did you tell him?
Kapor: I think I said probably $2 to $3 million. We had nothing. We had an
early-stage under-development spreadsheet, and me and Jon Sachs. So that was
the biggest number I felt I could ask for without being totally absurd