[A2k] Succeeding at open-source innovation: An interview with Mozilla's Mitchell Baker

[Riaz K Tayob]

Succeeding at open-source innovation: An interview with Mozilla's
Mitchell Baker

The company=92s chairman and former CEO explains the power of the
participatory, open-source model of collaboration.

Lenny T. Mendonca and Robert Sutton

January 2008

As companies reach beyond their boundaries to find and develop ideas,
they are exploring new models to manage innovation. In projects that tap
external talent, questions quickly arise about process management,
intellectual-property rights, and the right to make decisions. Some
executives have been at this game longer than others. Mitchell Baker,
chairman and former chief executive officer of Mozilla Corporation, has
devoted the past ten years to leading an effort that relies extensively
on people outside her company=97not just for creative ideas, but also to
develop products and make decisions. The result: Mozilla=92s Firefox
browser, with 150 million users, has become a rival of Microsoft=92s
market-leading Internet Explorer. MITCHELL BAKER Vital statistics


As Firefox flourished, the process that created it became a model for
participatory, open-source collaboration. Baker=92s role, central from the
beginning, has taken many twists and turns. Ten years ago, she was a
software lawyer at Netscape Communications=97which developed the original
commercial Web browser=97when the company decided to release its product
code to the public. Baker=92s interest in defining and managing the
project quickly earned her a place as one of its leaders. She continued
to guide the project after Netscape was acquired by AOL, led the
subsequent spin-off (to the nonprofit Mozilla Foundation and its
subsidiary, the Mozilla Corporation) to develop the next-generation
Firefox browser, and presided over Firefox=92s impressive growth. In her
role as =93chief lizard wrangler,=941 she balanced and blended Mozilla=92s
commercial needs with the motives and efforts of an army of volunteers
who develop the code and distribute the browser. Over the years, Baker
has helped define the legal and functional model that allows an
open-source community and a corporation to share responsibility for
product development while managing the project and maintaining the
organization=92s momentum=97not to mention its financial viability.

Today, Mozilla and Firefox are successful on several levels. Having
recaptured market share lost to Internet Explorer, Firefox now holds 15
percent of the browser market in the United States and a higher share
elsewhere. In 2006, the company=92s revenue-sharing arrangement with
Google for searches that originate in Firefox delivered revenues three
times greater than Mozilla=92s expenses,2 an impressive rate of return.
Finally, the organization=92s open-source development model is a visible
and well-tested experiment in managing innovation beyond corporate
borders. To learn more about that model, McKinsey director Lenny
Mendonca and Robert Sutton, a professor at Stanford University=92s
Graduate School of Business, met with Baker in her Mountain View office
before her change in roles.

The Quarterly: You=92ve said that Mozilla=92s real contribution isn=92t jus=
t
the browser but the model of participation. How do you manage
participation in this environment?

Mitchell Baker: Our mission is about keeping the Internet safe and open,
but also about building participation. We do that by setting up
frameworks where people can get involved in a very decentralized
fashion. These frameworks embody our values and our goals and get
embedded in other people=92s minds. We attract people who care about those
things, and they go off and participate in the mission in a very
decentralized way.

So for some things at the center, we must have extreme discipline. If
you=92re touching code that goes into Firefox, the process is very
disciplined. But there are lots of areas for participation=97whether it=92s
building an extension or localizing the product or building new
products=97that don=92t need that degree of discipline. And a key point is
for people to =93own=94 what they are doing, not in a financial or legal
sense but in an emotionally committed sense that gives them a chance to
decide, =93I=92m excited about this. I want to do something. I want to writ=
e
an extension. I want to go tell people how to do this.=94 And it also
gives people the success and the relationships to go back out and do more.

The Quarterly: How much of Firefox=92s success depends on people you
employ as opposed to the broader group of volunteers?

Mitchell Baker: I=92d say we need both to be successful. If you took away
our employees, we=92d be a good open-source project but nothing like a
force on the Internet. If you took away the volunteers and everyone
else, we would die. On Firefox, for example, 40 percent of the code is
not from employees=97and that=92s after a recent batch of hires from our
volunteer community over the past year. We had 25 employees two years
ago and now have more than 120. Sometimes we can hire from within our
community, but not always. There are some people with a high degree of
expertise and specialization who you can=92t hire, and we would never find
them if we weren=92t an open project. We would never find these people if
they couldn=92t just step up and contribute. A lot of folks will start at
one level, like fixing bugs, and go on to become star performers.

Actually, people can make a contribution without being either employees
or members of our volunteer community. Firefox has about 150 million
users worldwide, and since it doesn=92t ship on new machines, that=92s 150
million individual decisions to use it. How many people does it take to
do that? I=92m guessing hundreds of thousands of people around the world
who said, =93This is a great product. My family has to have it; my
neighbors need to have it.=94 Hundreds of thousands of different
decisions, and you cannot buy that.

The Quarterly: How do you motivate people to contribute to Mozilla,
especially after ten years?

Mitchell Baker: I think that for the people who have kept Mozilla alive,
the desire to maintain an open and participatory Internet has been very
important. The Internet is hidden to human beings except for this piece
of software we call the browser. Years ago, we could see that there was
some risk of people not being able to reach the Web except through a
browser that was part of a business plan. And by the year 2000, we were
seeing pop-up ads, spyware, and other things that slowed down the whole
computer. I think of this as abuse of the consumer, but it is a
perfectly rational business decision for some companies to do that
without considering it evil or nasty. But many people feel there should
be an alternative, and that dedication to an open Internet has helped us.

The Quarterly: What else?

Second, our product makes a giant difference in the lives of our
volunteers, and they take ownership of it. I don=92t know if you could
build this degree of motivation for something that really didn=92t change
people=92s lives, something that they weren=92t emotionally committed to.
But the number of people who feel that Firefox is partly theirs is very
high.

That=92s a tricky management challenge, but we work at it really hard. We
see ourselves as part of a community, some of which is inside the
organization and some that is outside it. Issues constantly come up
within our walls, and we have to say, =93This needs to be a public
discussion; it needs to go up on the mailing list because other people
are involved.=94 The community is reinforcing once you get started. We
can=92t ship Firefox or get it onto people=92s machines without that
community. So that means it=92s very much a two-way street, and if we
start to think of ourselves as the center, we will fail.

It=92s a very exceptional emotional state to feel like you=92re part of a
healthy community and that you=92re in trouble unless you=92re reaching out
and lots of people are reaching back. We also are extremely sensitive to
community criticisms and desires=97probably oversensitive sometimes. So
when some significant part of the community gets upset, we pay a lot of
attention. Sometimes our responses are defensive at first, but I think
we=92re pretty good at opening up. It=92s pretty interesting to look at wha=
t
somebody is complaining about and find the truth behind that. We also
try to be very low spin. In fact, sometimes we joke that we=92re negative
spin. We don=92t need the press or anybody else to do that; we=92ll do it
ourselves.

The Quarterly: The line between back stage and front stage appears to be
pretty thin.

Mitchell Baker: Yes, and quite permeable. And that is a management
challenge we haven=92t quite solved yet. What=92s the correct group of
people for information to reach? The easy default is employees because
we see each other regularly and they have signed confidentiality
agreements. But that=92s an unhealthy default for us because we=92re not
successful based just on employees. The community is as real as we are.

The Quarterly: How do you think about your role in enabling innovation
in the communities?

Mitchell Baker: Sometimes, just giving people permission does wonders.
Consider our quality control process. We have a public process for
finding, tracking, and correcting bugs in the code we=92re developing, and
thousands of people are involved. When several people within the
community began to take leadership in that effort, someone who worked
with me said, =93All we need to do is tell these people it=92s OK.=94 So
that=92s what we did. We said to the leader, =93You=92re awesome; keep doin=
g
what you=92re doing.=94 And after that, he became our release driver. There
are more people like that than you would expect.

Second, we create scaffolding for people to work from, so that even if
we=92re not innovating ourselves, other people can. You can see, with the
extensions and the customization, that there are thousands of people
doing interesting things we haven=92t thought of, and they don=92t have to
tell us or ask us.

Third, we=92ve assembled a set of people here who are really motivated by
seeing other people do interesting things. So if somebody appears, out
in another community, doing something interesting, we don=92t have a
not-invented-here culture; we just say, =93Wow!=94

Another thing: not just celebrating when people do great things but
knowing how to react when people do things that are troublesome. There
are days when somebody=92s done something and you wonder, =93What were they
thinking?=94 At that point, you have to look really carefully and evaluate
what=92s just uncomfortable and what really must be fixed. And then you
try to keep that latter category to a minimum. A healthy community will
do a lot of self-correction.

The Quarterly: What kind of people do well here as employees?

Mitchell Baker: Typically, people whose motivations line up very
strongly with either our mission or our technical vision. Also, people
who can handle large amounts of their work being public. People here are
following the bugs; they=92re watching each other, watching their
progress. They know how quickly you=92re working, and they know if you=92re
stuck on something. So you have to be able to live not just your social
life in public but your work life as well. We called it =93life in the
fishbowl=94 long before Facebook.

The Quarterly: What would be an example of a red flag that comes up
during the hiring process?

Mitchell Baker: If we ask, =93What do you do if someone disagrees with
you? What do you do if you think something needs to happen and it seems
to be slow or stopped?=94 And the answer is, =93Well, I tell them I=92m in
charge.=94 Bing. Even our employees rarely get told that, because I
believe that many of the things that work in open-source management are
also very valuable for your employees. You can try to tell an employee
what to do, but if the two of you disagree the employee may be right.
There=92s much more negotiation here, like a professional partnership.

The Quarterly: Has the culture of the open-source group changed the
culture of the core organization over time?

Mitchell Baker: I wouldn=92t say =93over time,=94 because I think we were b=
orn
out of that organization. I would say it infused us from the beginning
because even back at Netscape, leadership had nothing to do with
employment status. In fact, sometimes the managers of our project
members were demanding that they do things very contrary to what we at
Mozilla thought should happen.

The Quarterly: More traditional organizations that are now looking
outside themselves may not be used to this management style.

Mitchell Baker: Well, there=92s a real dividing line between simply
getting input from outside, though that can be very valuable, and what
we do. Our decision-making process is highly distributed and unrelated
to employment status, and some of the people who make decisions about
code are not employees. But what ships as Firefox, with the Mozilla name
and brand on it=97that=92s going to be a Mozilla decision, even though othe=
r
things are not.

The Quarterly: What has been the biggest surprise in the time you=92ve
been working at Mozilla?

Mitchell Baker: That we had exactly what was needed at exactly the right
moment. You often see this in start-ups that burst onto the scene and
grow dramatically. There=92s a lot of hard work and smarts, but also some
piece of timing is right. Those things, you can=92t control; you need to
be ready.

The Quarterly: Do you think your success in timing was related to the
fact that you had so many more =93sensors=94 in the community than you woul=
d
have if you were a group of 40 developers sitting in Silicon Valley?

Mitchell Baker: Oh, absolutely. We could not have succeeded if we=92d been
a closed little area. Yes, we had not only the right product but also
the right community of tens of thousands of people all those years, and
some sense of hope that we were the alternative to a closed Internet.
All of those things mattered. We knew we had a community because we had
been living in it for quite a while. All that came together in the
product=92s success. There=92s just no way we could have been or continue t=
o
be as successful without being this very diffuse organization.

The Quarterly: Looking ahead, what do you worry about for Firefox?

Mitchell Baker: That Firefox is only a part of what=92s necessary for the
Internet to remain open and participatory. We=92re the part that touches
human beings, and that=92s a powerful part. But we=92re just one element.
There=92s so much value and revenue in the Internet that it makes economic
sense for companies to try to create proprietary places there. And of
course, there=92s room for companies to do that and generate revenue for
their shareholders.

But there also needs to be a section of the Internet that=92s open, where
people can participate. Open source has been a phenomenal force in
pushing us in that direction. Firefox needs to remain strong enough and
innovative enough that we=92re able to continue to show the industry that
you can give people control or choice in an elegant manner and still be
a professional vendor and that there are revenue opportunities in this.
That=92s my greatest concern.

The Quarterly: What can other leaders learn from the Mozilla project
about running an innovative company?

Mitchell Baker: Turning people loose is really valuable. You have to
figure out what space and what range, but you get a lot more than you
would expect out of them, because they=92re not you.

Second, figure out where you want input. There are different varieties
of input and user-generated content. Figuring out what you really want
is very important because you can get benefits out of any of those
things. But if you=92re doing one thing and sending out a message that
you=92re doing another, I think you=92re dead.

Third, look hard at whether there are areas where you can give up some
control, because the returns are great. And if you can=92t, then stay away
from this type of model. If you have a good group of people around
you=97people you trust=97sometimes just stepping back when you don=92t like
something is really valuable. Let the problem play out a little bit. The
idea that a single individual is the best decision maker for everything
and should have ultimate control works only some of the time. I think
for Steve Jobs it works because he=92s so good at what he does. But if
you=92re not Steve Jobs, I have found that, sometimes, even when I don=92t
like something, there=92s often real value in stepping back and asking
questions. When you just ask people to stop what they are doing, you
lose their creative thought. And this approach can get even harder when
that person shows that you=92re making a mistake. In a lot of
organizations, people don=92t really admit when they make a mistake, which
I think is delusional because we all know that no one=92s perfect. About
the Authors

Lenny Mendonca is a director in McKinsey=92s San Francisco office; Robert
Sutton is professor of management science and engineering at Stanford
University. Notes

1The title refers to Mozilla=92s cartoon mascot=97a Godzilla-like lizard.

2See Baker=92s blog post about Mozilla=92s 2006 finances.

This email was cleaned by emailStripper, available for free from
http://www.papercut.biz/emailStripper.htm