Q&A: Eric Schmidt wants Google in your office
Google services for business customers will become "very profitable" once they reach sufficient scale, CEO Schmidt tells CNET News.
ORLANDO, Fla.--Watch out, business technology managers, because Google has its eyes on your domain.
If Google Chief Executive Eric Schmidt gets his way, the line that separates the computing services used by businesses from those used by consumers will fade fast. And Google, through services such as Google Apps and the new Google Wave, hopes to accelerate the change.
CNET News Poll
The company has done well so far with services that appeal chiefly to consumers, but Schmidt said at the Gartner Symposium here that Google likes services that become part people's lives regardless of whether they are doing work. And because the company covers its costs by charging enterprise accounts $50 per person per year for those services at work, he said it's just a matter of attaining scale before the business becomes "very profitable" for Google.
I spoke to Schmidt after a Gartner Symposium talk in which he said. Here's an edited transcript of the interview.
CNET News: You have a lot of enterprise information technology (IT) background. You were chief technology officer at Sun Microsystems and CEO of Novell. What did you have to learn--and what did you have to unlearn--when you came to Google?
Schmidt: Google was not founded with an enterprise bias, and I came with an enterprise bias. I can remember during the first couple years, I would describe at length the XML architecture and the data architecture corporations used. Larry (Page) and Sergey (Brin, Google's two co-founders) found that humorous: why would you need this? Of course there are reasons why you need it, but they're so specialized. To me the key breakthrough was understanding the browser can be both enterprise- and consumer-capable. The architecture is driven from the browser. That is the story of enterprise IT today.
So what did I learn? Consumer stuff is as hard or harder than enterprise stuff. When I was growing up, I thought enterprise was the hardest. Consumers are both very fickle and time-sensitive. (Services) have to be always there. The architectural assumptions of IT in the 1990s are not the assumptions going forward today.
You are big advocates of cloud computing. I've run into a lot of skepticism here at the show chiefly because of trust issues. What do you need to do to make your cloud-based services trustworthy enough and secure enough that a lot of big businesses will embrace them?
Schmidt: There are some businesses that will never embrace them. For purposes of argument that will be 1 percent. They'll conclude they want absolute control and are willing to pay a premium for that. What is that? Their own data centers, their own security architecture, their own risk management, and so forth. The vast majority, for purposes of argument 99 percent, will conclude that the analogy about the ATM machine is correct. Eventually the convenience of using ATM machines and the bank outweighs carrying the money around with you. Initially you think, "How do I trust the bank?" You work out the problems and (eventually) people have enough experience to know even if there is a problem it will be fixed.
In our case, the uptime of our servers and services appears to be higher than that of corporate services. When you study the reliability, we're trying to get to four nines (99.99 percent availability). Most corporate IT departments are not at that level.
With a lot of people I talk to about this, their analysis is in absolute terms rather than relative terms: something bad might happen, as opposed to the likelihood something bad will happen in the cloud versus their own IT.
Schmidt: This is a race where we just have to be better. Our pricing and flexibility is so much better already.
The message here is it's coming. I think smart people will come depending where they are in the adoption curve. Are they an early adopter, late adopter, in the middle?
The sales model is different (for selling enterprise services). So is the support model. You need more people to get customers to sign on to services, and you need more people to hold hands when things go wrong. I've seen a lot of complaints that there's nobody at Google they can call. Do you need more faces and people out there and an actual phone line to do this?
Schmidt: Be careful to distinguish between paying customer and free customers. Their service levels are quite different.
So if you're a paying customer you don't need to just send an e-mail or fill out a Web form, you can talk to a human being?
Schmidt: That's part of what we sell. If you're the CIO and you're going to take your e-mail system and throw it out, are you going to send an e-mail to somebody you don't know (when something goes wrong)? That's not a credible sale. The first thing a CIO is going to say is, "where is that person and how do I wring their neck?" You're going to have to provide high-quality service or people are going to pay for it.
When you saw the iPhone, the first version was really interesting. But where it got more interesting was the arrival of connectivity with Exchange servers.
Schmidt: And also the App Store.
They were simultaneous and both important, though the App Store was not populated initially.
Speaking as a former board member, Apple executed extremely well.
For me the App Store is really nice, but the Exchange connection is mandatory. Do you think Android (Google's mobile phone operating system) needs that, or are you going to rely on third parties to supply it?
Schmidt: I don't want to talk about product features. The simple answer is we have to solve the problem for the exact same reason the iPhone was able to solve it so well. If you think about it, Android is on its way to being a very, very high-volume smartphone for enterprise use. We need very powerful integration with things like BlackBerry services. You can go through the list.
Google Voice is an interesting service. It has some nice features for consumers, but I think as a business case it's got a lot more merit.
Schmidt: We've been having this debate. How do we turbocharge Google Voice? It's obviously incredibly useful to a CIO. We could take a couple paths. One would be to get more connected into PBXs (private branch exchanges, or corporate phone networks). Another would be to adopt more voice over Internet services (VoIP). All those are possible within the enterprise.
When you talk about launching more VoIP services for Google Voice...
Schmidt: You can think of Google Voice as a launching pad for telephony. The more you cross-integrate the stack, the more efficiency there can be.
With your higher-touch sales and support model for enterprise customers, is that totally offset by $50 per user per year? Or is this a lower-profit business than your present, more consumer-facing business?
Schmidt: It's certainly lower for gross margin. The text-ad business has very high gross margins. We don't do the math the way you ask. We try to say, does it cover our costs? The answer is sure. The price was set, to be very honest, arbitrarily. Because it's such a scale business, whatever number we set will be the perfect number because we'll grow into it. It'll be a very profitable business at scale.
I see a lot of interest in Gmail but not nearly as much interest in Google Docs--spreadsheet, word processor--and I'd put calendar in between. In the future, are those going to reach parity, or is Docs just going to be this bonus feature?
Schmidt: The way it plays out tactically is almost every sale is e-mail, calendar, and instant messaging. It doesn't start with Google Docs. They're playing with it, but fundamentally it's about e-mail and calendar. That's a pretty good project for a year for a significant company. They have to do a trial, convert the existing system, train users and support people. When you talk to those customers, they will tell you they will use Google Docs in conjunction with the e-mail accounts that they're already putting in place. That's how it plays out. It's first an e-mail sale, but once you have that, then you get the benefit of Google Docs.
The nontechnical press describe it as Microsoft Office versus Google Docs. They're not comparable.
Schmidt: This year. The reason they're not comparable is Microsoft Office is expensive and ours is free or cheap. The other thing is there are an awful lot of workflow features in Microsoft Office we don't have today. What we're doing is adding appropriate functions to Google Docs from the bottom. We're adding the common cases. We're not trying to build a full copy of Microsoft Office. I don't think that's good use of our time. What will happen is a corporation will end up having both around for awhile.
I personally think the spreadsheets are much more compelling. As a data capture mechanism built into the document (through online forms that fill spreadsheet entries) it's quite powerful.
It mediates against sharing. It's the perfect enterprise solution. You just throw the thing up there, it takes three seconds, and this interesting experience occurs.
The object is that if you add insight, you will have a new way of thinking about workflow and collaboration. That ultimately may relate to Google Wave and other things. You can think of the Microsoft Office model as the incumbent. Ours is a different way of solving the same problems.
You guys like to build your own servers, your own software, your own networks. Why not use off-the-shelf technology? Are the companies that do use off-the-shelf technology misguided?
Schmidt: Give me an example of off-the-shelf technology.
Servers from various companies.
Schmidt: Everybody thinks we use PCs. What we really do is build supercomputers. The supercomputers we build are made out of PC parts. They're highly specialized to the data architecture Google uses. It's a very pragmatic cost-benefit analysis.
So are other companies misguided?
Schmidt: I would never criticize another company. The Google model is sufficiently specialized that I don't think you can even compare. There have been a series companies created to build the Google architecture as a generalized rack server. I don't know how well they're doing.
Google is unique in many ways, but the general trajectory of very large data centers with a lot of x86 servers and reliability at a higher level rather than individual server level--that's a general trend. Maybe the rest of the market will intercept you at some point.
Schmidt: Maybe we are the specialized solution the generalized solution learns from. In many markets there's an iconic solution so that's so overengineered, but there's this trickle-down of ideas. I have lots friends in the venture industry working with start-ups with products. I say, "Ship us one and we'll beat the crap out of it. We'll tell you very precisely whether your claims are any good or not.
OK, to wrap it up, when people talk about Microsoft or Oracle, they understand what their enterprise strategy is. How would you sum up Google's enterprise strategy?
Schmidt: We're trying to build to bring the benefits of the consumer architecture of the Internet to all the people using enterprise services today with the same or better level of security and control.
And in terms of the actual services you offer?
Schmidt: I describe it as horizontal--those that will effect everybody in the organization. The ones everybody uses. We try to avoid specialized business logic like enterprise resource planning (ERP) systems. It's very interesting, but it's highly unlikely Google will be interested.