Tip:
Highlight text to annotate it
X
MALE SPEAKER: I'd like maybe to start with a question for
Dave. Why and when did Google get into the enterprise game?
DAVE GIROUARD: Well that was a nice introduction.
My mother couldn't have done better, so I appreciate that.
Well I don't think there was ever a moment in time where
Google decided it was going to be in the enterprise.
I think Larry and Sergey never had a concept
of enterprise consumer.
They pretty much saw people as people.
By the time they were at Google, they were in their
first jobs.
And they discovered they had to spend a lot
of time in the workplace.
And Google always rallies around solving problems that
impact enormous numbers of people, the whole world, if
you will, in very acute ways that they believe computer
science could address.
That's probably the most fundamental thing that Larry
and Sergey wanted to do.
And it so happens that hundreds of millions of us
spend a lot of our lives in the workplace.
And problems of information access are, and have always
been, particularly acute there.
So that's as simple as it gets at Google.
There's no grand plan.
There's no estimating from Gartner how big the market is,
or this and that.
It really is, are there big problems there?
Yes.
Are they ones Google could probably add some value to?
Yes.
Let's get started, OK.
And it just grew from there.
MALE SPEAKER: So whilst I'm asking questions, can you
think of questions as well, because in a minute, I'm going
to turn it over to you to ask those questions that you'd
like to have answered by Dave and Alan.
Another question for you, Dave, before we go on to Alan.
What can Google learn from the enterprise
market, and vice versa?
DAVE GIROUARD: A lot.
I'm in a unique role, because I spent a lot of my time
amongst those in the enterprise market, people like
yourselves, and probably half of my time out, learning,
listening, and the other half of my time inside Google,
understanding Google.
And in some sense, I feel like some sort of crazy transformer
trying to make these worlds understand each other.
But there really is a lot to learn in both ways.
And Google is very special in how it works.
It's very unique, and not without its quirks.
And for sure, Google has had to learn what it means to be
an enterprise player.
And we are continuing to learn that.
But as I always say--
for example, we acquired Postini a couple of years ago.
They were a company that was born as an enterprise company.
And they had real enterprise DNA, so to speak.
And my view is Google's DNA is what Google's DNA is, right?
It's very pure.
Having said that, it can be extended.
It can grow.
And so my task at Google has always been appreciate what is
special about the company, don't ever try to change what
is special about the company, but extend it and get it to
embrace the world of enterprise IT, which has a lot
of very special needs or requirements.
And I think that's been the delicate balance, is I don't
really want to change Google.
I probably need to try to change the enterprise market a
little bit, and hopefully meet somewhere in the middle.
And I think, so far, we've been able to do that.
MALE SPEAKER: So Alan, we've heard today that Google builds
data centers, as well as computers.
How could that make sense?
ALAN EUSTACE: Well I was one of the people that went
through the transformation of consolidation of data centers.
I really thought, with Moore's law and other kinds of things,
that eventually you get to the point where all the
information that you ever needed in your whole world was
on a device like this.
And I was wrong, massively wrong, not slightly wrong,
which is not that unusual.
But the problem is that it is true that these devices are
getting better and better at storing information.
But the amount of information that any one person needs at
any one time has just caught up dramatically.
Sure, I might be able to fit the Library of Congress on
this device, but no longer is that enough.
I need all the satellite imagery.
I need all the maps.
I need all the street view.
I need all news.
I need all the web.
I need--
Everything has to be here.
And so I was wrong on that.
And so when it when it comes to Google, and what caused us
to get where we are, it really is the fact that this whole
mission of organizing the world's information and making
it universally accessible has gotten more complicated.
And as a result, we've had to build more and more
infrastructure to be able to do that.
And no longer will small devices do.
Now it honestly requires huge data centers.
And then it requires networks of data centers around the
world, because you want the latency to be low.
And so then, if you have that model, it requires you to you
know how to deal with fault tolerance.
And then you have to deal with better networking, because the
clusters have to--
it's so much easier programs things with low latency and
high bandwidth networks.
And then, well, now, I have to do my own networking.
And so, it's a ball.
It seems like a simple problem, and it just gets
worse every day.
MALE SPEAKER: OK, so maybe just building on that.
Can Google serve businesses on the same infrastructure that
it services customers, Consumers
ALAN EUSTACE: What's funny about that question-- and Dave
touched on this a little bit-- is sometimes
we bumble into things.
And so, on the day that Gmail launched, many of you
remember, I remember, it was on April 1st. And it was a
gigabyte per person.
And everybody thought that was so outlandish that they
thought it was an April Fool's joke.
And it is always a bad idea, by the way, to launch on April
1st. Because then, half the newspaper said it was an April
Fool's joke, and half tried to give us a real coverage.
And the half that thought it was an April Fool's joke was
really mad at us.
But we launched, and then immediately we got lots of
feedback saying, look, this is fantastic.
It's really good.
I wish I had this in my business.
But I really don't feel like I can use like gmail.com as my
domain name.
Could you please give me any solution.
And so it was actually a relatively hack at that point
to allow you to do Gmail for your domain.
And that's what got us into it.
It wasn't a plan.
It was that there was demand from people that said, look,
this is a better solution than I have in my enterprise.
DAVE GIROUARD: It's funny.
It did grow out of running Gmail, but having your own
custom domain, a simple concept that the Gmail team,
actually, came up with.
And I remember as it came together, we went to Eric and
we said, we're going to launch this.
It's first for universities, and then for businesses.
And what do you think?
And he was all for it.
But at the time, somebody had come up with the genius name
of Google Hosted Services.
And as we were walking out of the room, he goes, by the way,
that name sucks.
Fix it.
Literally, we were launching in two days.
And the best we could come up with was Google Apps, and it
stuck ever since.
ALAN EUSTACE: But what was also funny is that we have the
same problems inside the company that you
guys face and that.
So we started Gmail inside the company,
because we used it ourselves.
And then we decided to launch a consumer offering.
So we did we did.
We cloned it and things.
But we had two different systems, one that we used
internally, and one we used for consumers.
And it was a nightmare, because now we had to keep the
two things consistent.
We had to have 24-7 support for both of them,
and things like that.
And I remember being in a meeting with Larry Page while
the IT people were there.
Larry says, why do you have two systems. Why don't you
have just one.
And the IT were saying, well, what about security.
What's going to happen if our corporate secrets, basically,
get switched over in some things with our consumers.
And Larry looks at him and says, don't you feel like my
secrets are important too?
DAVE GIROUARD: As a user.
ALAN EUSTACE: And you see the jaws drop.
It's like, well, Larry, no.
And we also had teams of people that were doing that.
And so Larry made an edict, said we are switching to the
standard infrastructure.
We're going to eliminate a private
in-house copy of Gmail.
And everybody's going to use a standard system.
And that forced everybody to not develop in two paths, and
to develop security systems and mechanisms that actually
work for both consumer and for enterprise.
And I think it was a good move.
It was an important move for the company.
DAVE GIROUARD: There may well be enough advertising revenue.
But my experience is most businesses, and when I say
businesses, I tend to think of--
once you get beyond a dozen people or so, my opinion is
very small businesses act more or less like consumers.
But when you get into real size businesses, they aren't
really interested in a free product.
It signals lack of commitment.
There's question about whether you're really
investing in this.
So in my opinion, at scale, I don't think a business wants
to use a free service.
They want to have a service, frankly, that's profitable,
because that is indicative that the company is committed
to it., is building a real business, will continue to
invest and make it better in the future, can support you.
And so we talked about that.
It certainly could be an option.
And by the way, we do have a free ad supported version for
very small businesses that's extraordinarily popular.
But Google didn't invent itself as
an advertising company.
And we don't see that as the only way we can earn
revenue over time.
ALAN EUSTACE: And we do differentiate our offerings.
Before it was exactly the same service.
We treated everybody identically,
and things like that.
And I believe that's good in some ways.
But but I think our enterprise customers
actually are more demanding.
And therefore we've changed some
things inside the company.
It's not that we treat normal customers badly, our consumer
customers badly.
But we have gotten to the point where we say that these
customers mean more to us.
And therefore we do distinguish our service.
I'll give you an example.
We have these orders that say we're going to do maintenance
on these particular things.
And because the way we build our systems, we fail over.
Whole data centers can fail.
And it actually causes no problems at all in that.
But it does introduce a little bit of risk, right?
And so normally we just rotate these things around the clock.
But for our business customers, we decided, well,
let's not do them in the normal rotation.
Let's actually shift that over to the weekends for them.
So they never even have the chance, basically, for one of
these things to have some kind of cascading failure.
And so I do think that it's really important for us to
differentiate our enterprise customers and we do that.
MALE SPEAKER: Other questions.
So a hand over here.
We will move across to the other side in a moment.
So please.
AUDIENCE: Dominic Shine, CIO of Reed Exhibitions.
Got a question about Chrome operating system, and how that
relates to enterprise.
What we talked about today, I think, is a very attractive
proposition of moving to Google Apps,
Salesforce, and so on.
But also the reality that they are a number of on-premise
applications that will be around for a while, which lead
us back to the conclusion, oh my God, we've got to stick
with Microsoft Windows, on the desktop and so on.
What are the ambitions for Chrome OS?
And are you planning to make a play on the space that would
give us other options there?
DAVID GIROUARD: Well Chrome OS, the concept there is
everything is in a browser, right?
So clearly it does not conceive of running your thick
client HR applications, etc.
And I wouldn't anticipate it would be a solution for that.
Having said that, a lot of businesses are moving more of
their users to the cloud.
And in some cases, a huge part of your user base may end up
surely using web based applications.
There's companies I hear that are moving toward a stipend
for compute.
Get a laptop.
Lease it.
What ever you do, we don't want to hear about it.
And when you start to think about that model, where, at
least for some, there is nothing installed
other than a browser.
Then I think the client that you would choose would
probably change over time, to something that is more
durable, more secure, very focused on what it does, very
fast at what it does, probably more energy
efficient, things like that.
So I think there will be a place for Chrome OS.
We aren't there saying, oh we're looking to displace
Windows off of all your desktops, etc.
There will be a lot of room for Windows I'm sure for as
long as any of us will be around.
But I think there ought to be some choices.
And some of your employees may well not need anything other
than a browser.
And you may prefer to have a device that lasts longer,
takes less power, doesn't require visits to
the desktop, etc.
So I would say, yes.
It is important to say that we've said Chrome OS is
something we're targeting for the end of next year.
We will start focused on the consumer, because we think
that make sense to do so, and that's where the pace of
innovation can happen quickest. But we definitely
believe over time there's a great opportunity to help your
businesses as well.
ALAN EUSTACE: Let me add to that that Chrome OS is a
really interesting idea.
The idea that most of us spend most of our time in our
browsers was unthinkable a long time ago.
But now literally I spend 98% of my time in a browser.
And if I do spend all that time in the browser, why do I
need all the stuff underneath it.
And why don't I spend--
And by the way, the browser, basically it does
multitasking, it handles fail over, it does all kinds of
things for me anyway.
So what kind of operating system do I
need underneath it?
And it turns out it's really small.
And then why don't I use all the resources that are
available there, like memory, and other things.
Use it to cache applications.
Use it to cache data.
Use it to do all these other things.
The moment I introduce an application that gets
downloaded on to that system, I break
all my security models.
I break all the simplicity.
I break everything else.
And what they channel five, and internet delivery
applications, I can get almost all of that
advantage really quickly.
And I'm even getting the performance I
haven't seen before.
So I really do believe that there are some
big advantages there.
But there are also solutions that are out there.
There's a lot of people who are using virtualization, and
VNCs and other kinds of things to provide access for that
tiny subset of people that really desperately need to run
this Windows application.
But I think it's going to go down over time.
MALE SPEAKER: I promised a question over this
way, so let's go--
AUDIENCE: As an organization that is in the middle of
migrating from a Legacy Mail Solution to Google Apps and
Google Mail.
DAVID GIROUARD: Thank you.
AUDIENCE: It's not without its difficulties.
It's not without its challenges, in terms of
encryption, and security, and single sign-on,
and things like that.
Are Google going to develop, if you like, a suite of
migration tools that de-risk the migration, and perhaps
help make it more predictable?
DAVID GIROUARD: That's a fair statement.
Let me say when we first launched Google Apps, we
frankly had no migration tools.
So full knowledge, a lot of the earlier customers that
moved to Google Apps had little choice but to not take
their data with them, or to go through some fairly painful
processes to move data client by client, etc.
So we've come a long way since then, but we certainly aren't
all the way there.
We happen to be, what I would consider, very solid, quite
good moving Lotus to Google, for whatever reason.
A heck of a lot of people seem to have done
that, and pushed that.
And our tools are very good there.
And I should say, around messaging, and
calendering, etc.
Moving Lotus applications is a whole different question, as
most of you are probably aware.
On the Microsoft side, we are improving, but were not as
quite there yet.
We will this quarter finally release the full server side
version of migration tools, to get all the data you need to
out of Exchange.
And we're beginning down the path of having tools to move
things out of SharePoint into Google Sites, etc.
So we are working very *** migration tools.
They're not perfect, though they are much, much better
than they were even six months ago.
We will continue to have work to do, particularly on the
applications side.
And you think about your Lotus apps, or your SharePoint apps,
or whatever.
You want to get rid of those, shut those servers down.
It will never be point and click for those things.
But we do believe on the messaging side, frankly, it
ought to be point and click.
And you do have to still think about things, like how much
data do you want to move through your firewall, how
much email do you want to keep.
The data model of calendars is actually horrendously complex,
right, trying to move calendars from
one system to another.
How Lotus handles holidays, and how
it reschedules meetings.
There's all sorts of weird corner cases.
So it will never be perfect.
But our goal should be you can make a few quick decisions
about what you want to move and what trade offs you're
willing to make.
And it can be entirely predictable
from that point on.
And I think we're actually getting better at that, though
certainly not perfect.
MALE SPEAKER: Any questions over this side, where we
haven't had any from over here.
Carson.
And then I saw a few hands over here.
We'll come to you afterwards.
AUDIENCE: I just, very brief, want to address the innovation
model that Google is using, this idea that you, that the
difference between a product and a service is generally
that you can own a product, but you don't own a service.
So of course, all the nice features that suddenly are
there next time I log on.
Its excellent as a consumer.
And you're sort of more prepared for that kind of
innovation that happens from the center out to the edges.
So I'm just wondering whether you have been considering the
challenge that might emerge for that innovation models,
once you engage with large corporations, who might expect
a different kind of process?
Or might it be so that they'll have to get used to this is
the new world of innovation processes.
DAVID GIROUARD: There's probably no place more ripe
for where does Google meet enterprise, and have to sort
of navigate than this area, which is how is technology
delivered and consumed.
And without question, if you are in a traditional IT role,
where you are used to having upgrade is coming next year.
We're going to prep.
We're going to sandbox.
We're going to see who--
we're going to set up training courses.
Gone, all that's gone.
Now we do a bit better than just saying, poof, shows up
Monday morning and everybody has a new calendar system.
The truth is these applications do grow and
change slowly and surely every week.
Literally most teams are on two week launch cycles,
sometimes things you would never notice, a little
performance thing, sometimes small tweaks.
Now, when there are big changes--
I want to talk about big changes and small changes--
when there are big changes, things that anybody would
notice, we are fairly good at now giving you the choice.
Opt in or opt out, or automatically turn on, or wait
and let you evaluate it.
So the IT, the administrator of the system, has a lot of
choices when big changes come.
So I don't think we have issues there.
Small changes-- and of the course where the line is
probably a good question--
there's a lot of trust involved, because we don't
have an opt in button for every small change.
What we do have is a fairly significant focus group of 100
million people or so, who we will, for example, if we're
going to move a button to the right a little bit in Gmail,
we'll release it to a tenth of a percent of people for ten
minutes, and have a few million data points about how
people react to that.
So while yes, you have to have a level of trust, that these
applications are going to grow and change.
That trust, I think, over time, people will get very
comfortable with.
And it's funny, somebody was asking about this.
I said for every ten complaints I hear about when
are you going to give us x, y, and z features, when you are
going to finish your mail migrate, whatever it would be,
I hear zero about we really didn't want
that, literally zero.
It's one of those things, I think, that is intuitive, that
it should be a problem, but have yet to hear the problem,
for the reasons I describe.
It is a different model, for sure.
And it takes a while for people to get
comfortable with it.
ALAN EUSTACE: This is a philosophical thing too.
I remember Toolbar, when we first released Toolbar.
Larry was insistent, I want it to auto update.
And at the time, nobody auto updated because the
risks are too high.
We have millions and millions of toolbars out there.
What happens if they all stop working?
What happens if we can't auto update again,
if we make a mistake?
Larry says, I want it auto updated all the time.
He says, if we have a bug, I want to fix it.
don't want to wait.
If there's a security issue, I want it fixed.
I don't want us to deal with a thousand different versions of
toolbar out there.
I want there to be one.
You do not have a choice.
Use our software.
You get it auto updated.
And then we force our teams to just test really, really
heavily, to make sure that we're not screwing them up.
But the result is, if you compare that model to other
companies, where you're constantly being bombarded--
well we got a new update for you, do you want to accept it.
They don't tell you what the heck it is.
They just say, do you want your software updated.
What choice are they really giving me?
So I think there's a philosophy of updating things.
And I think it's kind of thrilling to come in every
once and a while, and say, wow, video chat, who would
have though?
Now maybe somebody doesn't want video chat, and
that kind of stuff.
But just that breaks the whole model.
I didn't have to pay $50 for it.
I didn't have to schedule it far in advance.
All of sudden I got something.
Translation, it just appeared.
You get this thing in your mailbox that's in a different
language, and it says do you want to translate it.
And you hit yes and it's there.
I think consumers will expect that in the future.
And I don't think, when it comes right down to it, if you
had, like we do, a great administrator panel that says,
yes or no to video chat, I think most people will say,
yeah, yeah, we'll take that.
MALE SPEAKER: OK, so over to you, please.
AUDIENCE: How is your position to service the corporate
market in the future, coming back to Jeffrey Moore's model
of, do you want to do this on a standalone basis, or on a
network with outsourcing service providers and
integrators.
DAVID GIROUARD: Well I would say, philosophically, we kind
of answered this question when we started into enterprise.
And we decided that Google was a product company.
We didn't aim to be a global services, solution, blah,
blah, blah.
There's a lot of those and they're very good at.
So we really said-- and frankly, it probably slowed us
down, because it's always easier to go into a company
and say we will do everything for you, just sign
here, and sit back--
but in reality, we did not aim to become ExEnsure, or IBM
global services, or anything.
So we purposefully pulled partners in, because no matter
how good we make the services, the integration to the
enterprise environment will always require some help,
migration of data, single sign-on integration, active
directory, and whatever it would be.
And so we are growing partnerships with people like
ExEnsure, and Capgemini, and CSC and others.
And those are the big guys.
There's probably a hundred times the number of guys
you've never heard of who are becoming
cloud service providers.
And they're just experts at the change management, the
technical bits.
But we're very clear.
Our goal is to make this as much turn on
as possible, right?
And we don't want to be in that business.
We want to make this a service you can turn on when you need
it, and turn off when you don't, and stick with what
we're good at.
I don't know if it's the ultimate right answer.
I should say, by the way, for the big companies like Jaguar,
Land Rover, and others who are with us early, we're taking a
little different approach.
We are very hands on.
And we make a 100% sure that their project is successful,
because they're amongst the earlier adopters.
But we don't bill for services and do that kind of stuff at
all, because our goal really is to enable an
ecosystem to grow.
MALE SPEAKER: There was another question over here.
AUDIENCE: I'm Vikram from Wipro technologies.
Two questions, so one, we've heard a lot about Gdrive.
Is it really getting launched?
That's one.
And second is you've heard about so many customer case
studies, so what I want to really ask is are the
customers shifting all their users to Google Apps, or is a
hybrid market more prevalent?
We are shifting most of the users, but they still have a
core set of users in the enterprise with the Legacy
environment.
So what's the trend?
DAVID GIROUARD: Gdrive, let me answer that one.
We obviously don't announce things that haven't been
launched, so we don't have any product to announce today.
We generally want the cloud to be a place where
you store your data.
It doesn't always have to be in our formats.
We think you should be able to store photos, and be able to
store cad diagrams, and anything
you feel like storing.
So you're going to see a lot more from us in that regard.
So Gdrive is little bit of imagining the world as you see
it today, which is a mounted hard drive on your desktop.
To me, it may or may not be interesting.
The interesting part is can you actually move the mass of
your data into the cloud, either as a consumer or as a
business user, access it from any device, and have really
full fidelity, reliable, fast access from anywhere.
That we will do.
Exactly how we will do it, I would hold off on.
ALAN EUSTACE: And the Chrome OS is a good example, because
I said, we don't have applications.
We don't have storage.
So Chrome OS, how are you going to store your data?
Well, you have to store it in the cloud.
So for that model to work, you have to have a
cloud storage model.
And you have to have the software to be
able to support that.
So it is something, obviously, that's important to us as a
company to develop, because it's a prerequisite for some
of the other things that we're doing.
DAVID GIROUARD: What was the other half of your question?
AUDIENCE: [INAUDIBLE]
DAVID GIROUARD: Oh, well, let me tell you what I think most
of the model looks like.
What I see more often than not-- we always thought in the
early days, we'd get the deskless workers, the easy
people, and the hard core people.
Generally what we're seeing is people move 100% of employees
to Gmail, and Calendar, and Google Talk, because you don't
generally want more systems than that.
And you can use Outlook, if you want.
And they think about it.
Usually they start down the path saying, oh, we'll just go
with the Asian offices, where we could use low cost. In the
end, nine times out of ten, they come to the conclusion,
we're moving everybody there.
Now Docs, or you might call the applications, what have
you, the kind that are sort of analogous to Office.
That is very different.
There's virtually nobody at 20,000 people who are just,
Office, you're out, Google Docs, you're in.
What you see-- and Todd mentioned this from Genentech.
And I'll tell you the way it worked at Google.
Google has about 20,000 employees.
When Gmail was live, we had 20,000 people
on Gmail, day one.
Docs was launched--
ALAN EUSTACE: We didn't have that many number of people
when Gmail was launched.
DAVID GIROUARD: That's true.
We didn't have that many.
But it was 100% uptake, right?
What you see, because we look at active users by every
service, and we can see by domain-- and at Google, Docs
and Sites, since then, has now become about the same as
Gmail, seven day active users, the entire company.
And that's what you see from others, which is
they go live 100%.
And then the Docs and Sites, the collaborative products,
over usually a year, or something like that, will grow
to be just as much used.
But it's organic, as opposed to a big, big switch.
MALE SPEAKER: So we probably have time for
just one more question.
Gentleman over here.
AUDIENCE: OK.
[UNINTELLIGIBLE]
from [UNINTELLIGIBLE].
And the question is related to the fact that some countries
do have the firewalls.
Do the enterprises need to be worried about the Chinese
government firewall and its impact on the Google
cloud-based services.
DAVID GIROUARD: Do you want that one?
ALAN EUSTACE: No, I don't want that one.
This is where I pull rank.
You take it.
DAVID GIROUARD: We don't store user data in China.
And we very purposefully have done that, for either
consumers or businesses, because we absolutely need to
make sure we are able to protect our users data against
any improper encroachments on it.
So there's one thing.
You can have users in China accessing our systems. And we
do occasionally have spats--
I don't know, how would you describe that?--
issues about accessibility of our content from China.
But generally, through our applications, it works quite
well that data is not hosted in China.
That's kind of the short version.
MALE SPEAKER: So we're *** out of time now.
So can you all just thank Alan and Dave.