Arthur van Hoff was part of the Java development team at Sun Microsystems
when he left in 1996 to found Marimba, a software distribution company.
Joining him as cofounders were two fellow developers from the Java team, Sami
Shaio and Jonathan Payne, and Kim Polese, Java's product manager.
Marimba received lots of attention from the press and venture capitalists
early on. The company grew from a 4-person startup to a company with more
than 300 employees at the time of its IPO in 1999. van Hoff left the company in
2002 to start another startup, Strangeberry. Marimba was acquired by BMC
Software in 2004.
Livingston: At what point did the four of you start talking about leaving Sun
and starting your own company?
van Hoff: Jonathan had left Sun, and, when I tried to convince him to come
back, he said, "Well I don't know if I'll ever come back to Sun, but I'll do a
startup with you."
So we decided to do a startup, though we literally had no idea what we were
going to make. The first thing that we did was drive around and find office
space, which was getting pretty hard at the time. We found a little office above
a flower shop on California Avenue in Palo Alto, and we went to a second-hand
office furniture store and bought these heavy metal desks--$25 apiece. They
weighed a million pounds, but we somehow carried them up the stairs.
Livingston: How did you fund your company at first?
van Hoff: Initially we all put in a little bit of money, I think $25,000 each. If you
don't take a salary, that can last you a long time. Because starting a company is
free, right, if you have a friend who is a lawyer. The law firm that we used,
Gunderson Dettmer, will basically not take payment until you get funding.
Silicon Valley is that way; everything is geared toward getting you started, and
then you pay.
153
Arthur van Hoff
Cofounder, Marimba
11
CHAPTER
We spent about $1,400 to furnish the entire office, including equipment
like a fax machine and printer. We all used cell phones at first, and we had no
Internet access for the first couple of weeks, just the whiteboard.
Livingston: You took a pretty big risk to decide to start a company without an
idea. You must have known that the four of you were pretty compatible?
van Hoff: You know, in a hot market like that, you saw a lot of people with crazy
business ideas that were never going to work but they were getting funded. We
were coming out of the Java project and felt that it was a pretty safe bet that if
you are part of a core team and you leave together, getting an idea is not that
hard. Anybody can have good ideas.
Over the years, I've learned that the first idea you have is irrelevant. It's just
a catalyst for you to get started. Then you figure out what's wrong with it and
you go through phases of denial, panic, regret. And then you finally have a
better idea and the second idea is always the important one.
After Marimba, when I started Strangeberry with Jonathan, we had no plan
whatsoever. We just put in some money and decided to spend a year brainstorming.
We built all sorts of things, and everything we did turned out to be
very relevant, because you're in the right area and you are giving yourself time
to investigate. Eventually, you run into an interesting idea and you execute on
that. People are really the key.
Livingston: When you left Sun, did they try to stop you?
van Hoff: Kim and I did a very dramatic thing. We arranged to have a meeting
with Scott McNealy. He asked what we were there to talk about. When we told
him that we were leaving to do a startup, he said, "Well, I can't wish you good
luck, because everybody would go and do this. But I'll tell you one thing: don't
fuck with me."
One of the things that we wanted to build was a user interface builder. Java
was an interesting model, but there weren't any tools for it. So we spent the first
few months working on a user interface, and then these guys from a small
startup visited us and showed us their product, and it was pretty much what we
were doing. They were acquired by Netscape like the next week, and they
turned into the IFC (Internet Foundation Class). It was ironic because that
eventually turned into the JFC, or Swing, the Java toolkit.
Livingston: Were you devastated that another company was doing the exact
same thing?
van Hoff: Not really. Once they were acquired, we sort of threw in the towel
because Netscape was so popular and there was really no way we could compete
with that. We hadn't spent a lot of time on it yet. We had some prototypes
and it was working quite well, but we moved on really quickly.
It was very surrealistic at the time because we had a lot attention from the
press. There was a full-page photograph in Wired with no information at all. We
weren't telling anyone what we were doing--mostly because we had absolutely
no clue and we didn't want to let on.
154 Founders at Work
But we then focused on software distribution, because the system that we
helped build at Sun was not really scaling very well for real applications. We
came up with the idea for subscription-based software where, rather than buying
software, you subscribe to it and you get updates automatically.
That was an interesting idea, but it's only now that it's really popular. These
days, a Windows computer updates automatically and so everybody expects
that--but at the time this was a very new concept.
By the time we announced that we were doing software distribution,
PointCast had come out. PointCast did push technology, which had some similarities
to what we were doing, but we were immediately filed under "push."
And that became a real problem, because for years we had to explain to people
why we weren't a push company.
Livingston: Did all the publicity help or hurt your cause?
van Hoff: Well, all press is good press. It definitely helps. Whenever we wanted
a meeting with an executive at a big company, we'd get it because we were very
well known. Nobody had a clue what we were doing. So the mystique around
Marimba gave us a lot of inroads to companies, which is incredibly helpful to
get deals done.
In the end, it can work to your disadvantage because you always have to
reeducate the market--you have to keep explaining what you really do. And
you never have anyone coming to you saying, "I want what you have," because
they don't know what you have. So it can work both ways.
There were always reporters talking to Kim because she was a female CEO
of a technology company. I don't know if that was a good thing. There was so
much focus on her and so little focus on the company. I'd go to parties and
people would ask where I worked and when I'd tell them they'd say, "Marimba?
Oh yeah, Kim Polese works there, right?" And I'd say, "Do you know what we
do?" And they'd say, "No, I have no idea."
So if we were selling Kim Polese, we did really well. But that's not what
you're there for. You're there for the product. So I think all the media hype did
not work to our advantage. I think that Kim fell into that trap early on, and it
was hard to get out of.
I remember one particularly bad article by Fortune magazine. This reporter
came and visited the company for two days while we were on a company outing.
We really opened up the kimono and spent hours with her, telling and showing
her everything. Then the article that came out was an expose on Kim and it was
made even worse because they'd taken these photographs of her that were real
extreme close-ups. It was terrible; it just made us look very bad since it was all
about her. And all this time we spent with the reporter on our technology had
been a complete waste of time, which was incredibly unfair.
But that's the problem: it's so much easier to write an article about Kim than
it is to write an article about the company. It's not very interesting to write about
mediocrity. You have to write about the extreme, because that is what people
want to hear about. So when companies are all about selling product, traveling
and working hard, it's all really boring stuff.
Arthur van Hoff 155
Livingston: Did you ever suggest to Kim that she stop talking to the press?
van Hoff: Yes, we did, but she ignored it. Kim was a good CEO in the startup
phase, but as startups grow it can get more difficult. Marimba went from 0 to
40 people in the first year and grew to 300 people during the IPO. Anybody can
run a company up to 100 people. You just have to be intelligent and have good
intuition. There's a lot of tedious work you need to do, but it's not that hard. But
there's a point when the company gets bigger that it just becomes a management
problem; it becomes something that you have to have experience in.
Managing people and motivating teams requires a very different skill; it's
not something that you can do by the seat of the pants. So the lack of experience
eventually begins to show if you don't have somebody who can make decisions,
for example.
We had this really funky power balance in our company where we had a
really strong VP of sales and a really strong CFO and a really inexperienced
CEO. And whenever there was a decision to be made, she couldn't break the
tie. And what do you do? Once Kim got replaced by John Olsen, he was completely
different. John had run big companies and it was really easy for him to
make decisions that were very hard for us to make. And that tells you that as a
founder, you have the skills to start companies from scratch, but it doesn't necessarily
mean that you have the skill to grow it till they're larger.
Livingston: Did you have a plan in the beginning to get that big and take it
public?
van Hoff: Well, every business plan has an exit strategy and ours was IPO. That
was the right choice at that time. Right now, you'd aim for an acquisition.
Everybody that joins a startup hopes to get rich. They also do it because it's fun,
but you're taking a bet on winning a lot of money. But the odds are skewed
against you because not a lot of startups actually succeed in fulfilling that bet.
We exceeded our own expectations in the end. But a business plan is a tool
that you use to sell the idea to VCs. The VCs look at it and say, "There're no
spelling mistakes and the math seems right. But I like the people so let's invest."
A lot of the decision-making is very emotional. There's no formula that identifies
good business plans versus bad business plans. So I think it's not really a fair
question to ask "Did you execute on your business plan?" because every business
plan is just a wild guess, right? You could easily add a couple of zeros
everywhere and sell the same thing to people. Instead of 10 percent market
growth you make 20 percent market growth, and suddenly you make $200 million
more in the fifth year, but so what? They're marketing tools.
Livingston: What big turning points occurred in the first year?
van Hoff: We did a first release of the software, which was a really important
thing. We hired some executives and lots of great people. We hired some really
bad people too--we had to fire somebody in the first year. We had our first lawsuit
filed.
Livingston: Was there any time when you wanted to quit?
156 Founders at Work
van Hoff: Marimba is an unfair case because we were willed on like crazy by
the investors. We really had an unfair opportunity because when we got funding,
the VCs were calling us. They all wanted to invest because they had heard
about us and wanted to find out what we were doing.
So we got a really good first round of funding--$4 million from Kleiner
Perkins. Though I thought they wired the money in these situations, they actually
gave us a check. So we had two checks--from the Kleiner fund and the Java
fund--and Sami goes, "Let's go to Kinko's and make copies!" So he takes the
checks to Kinko's and comes back with the photocopies, and he forgot to take
the checks out of the copy machine! Luckily they were still there.
Another story I remember from our first round of funding was when they
gave us the checks--the lawyers were there, Kleiner was there, and I said, "Oh
great, now I can buy that espresso machine!" and they all jumped me and said,
"No, you're not going to buy an espresso machine with this money. This is to
start the company."
And it became a sticking point. We were very frugal and we didn't spend
money on frills, but after the IPO there was a really bad time for Marimba
when it was very difficult to hire people, and all the early people that had been
there 3 to 4 years were starting to leave. Morale was very low, and so I went to
the CFO and said, "Look, I want to buy an espresso machine." And he said,
"No, we can't do that, it's too expensive."
A few weeks later, when another senior engineer quit, I said, "Screw it, let's
go buy an espresso machine." So Jonathan and I went online and bought this
super-duper Italian, fully automatic, $15,000 espresso machine on his credit
card and submitted the expense form. The CFO almost had a baby. It was
unbelievable.
This was a beautiful piece of work, and they came and installed the espresso
machine and it was the best money we ever spent. Every morning, people
would meet and crowd around it. This thing was just it, the bee's knees, people
loved it, they couldn't stop talking about it. A month later, the CFO came and
said, "I'm sorry, we should have done this years ago." And it tells you something
about where you spend your money and what you spend your money on. It's not
just business-related expenses. You also have to create an environment that you
like so that people are happy and feel they are valued.
Livingston: Did you get along with your VCs?
van Hoff: VCs are an interesting bunch; you can't live with them, you can't live
without them. They are instrumental in your success because they give you
money and a really strong endorsement. They have this mafia-like network of
connections and they help you with deals and find the right executives. They
are really working your case.
In my experience, it rarely happens that they turn against you, because
you're a team and if the team isn't working, the company will likely fail.
Occasionally, when you're a screw-up, they'll have to make a tough decision and
fire someone, but that's rare in my opinion. Because they wouldn't invest in
your company if they didn't believe in you and your team. So I've always had a
good experience working with VCs.
Arthur van Hoff 157
Livingston: Was there anything about your technology that people misunderstood?
van Hoff: We spent a lot of time talking about subscription-based software distribution
and that took a long time for people to get. In hindsight, we were
probably a little early with that. Now it's a very well-understood thing. The
Microsoft operating system updates automatically. Updates to virus programs
come over automatically. In the beginning, a lot of people we talked to said, "It's
too early. Do I really want to do this?"
But we had a couple of really big successes--Morgan Stanley and Bear
Stearns. These companies that had thousands of traders all over the world
really needed to use the same software or it wouldn't work. They needed to roll
this out at 100,000 endpoints and needed to get a report and warn people that
didn't get updates. And we did that very, very well. Over time, Marimba went
from a consumer software distribution/push technology company to an enterprise
software distribution company--which is a lot more boring than in the
early days, but there was a lot more money to be made in that market.
Livingston: What would you tell someone who wanted to start their own
company?
van Hoff: If you have the energy to do it, then you should try it yourself. But
you do need to have the ability to form a team around you with good people.
Talent attracts talent.
A lot of people get stuck on the idea. They all want to invent something and
go execute on it. I think that's a fallacy. You have to have an unfair advantage in
that you have to be good at something, or you have to have a direction that
you're interested in or a market that you see an opportunity in--but you
shouldn't get stuck too much on the details, because you can't foresee your
future anyway. Because you'll go through so many changes, I don't think it pays
off to overanalyze the first business plan, for example. The first business plan is
there to make sure you can use Microsoft Word.
Eventually, you need to go to VCs and attract money, and at that point you
need to be able to put your plan in writing and sell it. That's something you
need to practice a lot. Start with your friends and your parents and eventually
go to VCs. If you get good reactions, then keep doing it. If you get bad reactions,
then stop immediately, because it's a really bad idea to sell a bad plan. You
can screw up once, but it's hard to screw up multiple times, because the VCs
won't give you the time if you come up with a few bad plans.
Another good idea is to join a startup that already has funding. That way you
can experience the startup atmosphere and all the pros and cons without really
taking all the risk yourself. Because doing a startup does mean that you have to
give up your job and your income and take the plunge. That's what holds a lot of
people back.
I'm lucky I really don't need to work anymore. If I do a startup, whether it
succeeds or fails is somewhat irrelevant--I do it because it's fun. I'd like to succeed.
When it comes to taking a salary, at Strangeberry we worked for several
years without taking a salary because we had fun doing what we were doing.
158 Founders at Work
Livingston: What do you remember as being the most frustrating things
early on?
van Hoff: The thing that was most frustrating for us very early on was that we
got a lawsuit that just kept dragging on and on, and it took so much time and
attention, and that became a real pain in the ass.
Livingston: What was it about?
van Hoff: It was a patent infringement case, without merit. Patents are pretty
frivolous overall anyway. But if you're at the receiving end of a lawsuit, it can
make things difficult.
One of the problems for the founders, after the IPO, is that you can't sell for
a certain period of time and, after that, every time you sell and the stock goes
down, you'll get personally sued--shareholder lawsuits. So every time there
was an opportunity for us to sell, our lawyers would say, "You better not because
if you lose the lawsuit then you'll get sued. You'll replace one lawsuit with the
other."
So we had to see the stock go down from $75 to almost nothing and we
weren't able to sell. We were legally able to sell, but you kind of talk yourself out
of it because you think the risk is too high. If you do a startup and the company
goes bankrupt, the shareholders lose their money, but you don't personally lose
your house. But a shareholder lawsuit is a personal lawsuit--if you lose, they
take your house, so it's a totally different ball game.
There're all sorts of crazy schemes that people use to get around this stuff.
But at the time, we were pretty naïve about these things. You don't want the
employees to focus on that, so you take the burden and deal with it.
There's a lot to be learned from doing a startup. It's much broader than you
think. Although I was the CTO, I wrote a lot of code, I did a lot of depositions,
interviewing, selling, traveling, moving furniture. That's the great thing about it;
it's not a regular job. I like that, and that's why I've done a couple more since
then.
Livingston: Was there anything you found you were better at?
van Hoff: You grow into it a little bit. We had just received the President's
Award at Sun, which is a really prestigious award that they gave out every year,
and it's a whole bunch of stock options. And we were going to walk away from
that. It's sort of ironic, because the Sun stock split three times since we left, and
if we had sold at the peak, we would have made about as much money as we did
with Marimba, personally. But would I do it differently? No, I had a great time
at Marimba.
Livingston: Did you have regrets?
van Hoff: When it's your first startup, there are a lot of people involved. You
take advice from a lot of people, and that advice is not always the best advice.
Very often, your intuition tells you to do something different, but then you go
with the advice from the experienced guys anyway. And there were a few occasions
where I look back and think, "If only I had gone with my intuition, things
Arthur van Hoff 159
might have been different." So I might rely more on my intuition if I were to do
it again.
Livingston: Were you the ringleader to start the company?
van Hoff: Well, in a way. Jonathan and I came up with the initial idea to do a
startup, but you're talking about a difference in weeks. Very quickly it became
the four of us. Then you need to make some decisions about when do you want
to leave and how much money are you going to put in. Then once you've left, it
gets quite interesting. Because then you've got to go for it whether you like it
or not.
Livingston: Were the founding shares divided equally?
van Hoff: Yes, we split it four ways. We were very lucky because at the time of
the IPO, the founders still had a fair amount of stock. Financially, the company
was structured really well. That's mostly because early on, we had some really
good VC deals. Especially with four founders, if you're not careful, you end up
with such a small portion of the company.
Livingston: What advice would you give to a group of people who worked
together and wanted to go out on their own?
van Hoff: Don't take anything with you. Especially if you go and do something
that is somewhat competitive with your previous employer. Although you might
not have actually taken anything--ideas, physical things, or time--if you're successful,
they'll come and sue you just for fun. They'll have a really good starting
point because you are a previous employee. You must have taken something
because you're successful now, right?
So unless you go into a completely different area, you have to be very careful
about the intellectual property. So really what you've got to do is: don't plan
anything, don't write anything down. Talk about it over a beer and then leave.
And then you start. Don't use any office equipment or email.
It's irrelevant if the company fails, but if the company succeeds, that can be
a big problem. The funny thing is that they won't sue you until you're successful,
because why sue someone who is a failure? And this is particularly important
if you start out at a big company like Google or Amazon, because they have
a lot of time and money to spend on these kinds of things.