Enterprise Search: Dave Girouard on Taking Google to the Corporation

By CIOinsight  |  Posted 05-02-2006

Enterprise Search: Dave Girouard on Taking Google to the Corporation

When it comes to making sense of the overwhelming volume of data that's been unleashed upon the world over the past ten years, no company has had more of an impact than Mountain View, Calif.-based Google Inc. Anyone who had the misfortune of using Alta Vista as their primary search engine in the early 1990s can attest to that.

But as much as Google has insinuated itself into the very fabric of Internet culture, capturing consumer mindshare in leaps and bounds, the company has yet to offer much of value to corporations. In the same way that consumers were set adrift in a sea of endless information in the early days of the Web, enterprises are now drowning in data sets of their own. And the problem is only getting worse.

That's why Google has Dave Girouard, general manager of Google's budding enterprise business. Dave's job is pretty simple: apply to enterprise data the same revolutionary search technology that organized the Web. To do that, Google is working with business application vendors to make the data locked away inside various structured tables and proprietary software available to a simple, intuitive search engine. And in the process, Google is aiming to become the primary interface for all enterprise applications. Executive Editor Dan Briody spoke with Girouard about Google's ambitious plans. An edited version of their conversation follows.

Next page: Getting to Know CIOs

Getting to Know CIOs

CIO INSIGHT: What can Google do for the enterprise?
The enterprise is obviously not where Google began, but the company was founded in 1998 with some very ambitious goals to organize the world's information. That was really the substance of Larry [Page] and Sergey [Brin]'s goal when they started the company, and it explains a lot of the things Google does. You know, organize the world's information and make it universally accessible and useful.

The enterprise is a subset of that. Information access is a big problem in the enterprise. It affects hundreds of millions of people on a daily basis and it's a thorny problem. We think it's one that Google is uniquely qualified to solve. We don't see ourselves as a Web-search company. We don't see ourselves as a consumer company. We see ourselves as an information company.

What Google really needed to do over the past few years was build an organization, and a set of capabilities, so we could go after this enterprise part of the problem. We're able to tap into a lot of the core intellectual property and capabilities of Google, but we also had to develop a lot of new capabilities to channel that expertise into the enterprise.

So have you been meeting with CIOs to gain the knowledge that you needed to address this market?
Well, like a lot of things at Google, we started out very small. We got something into the market very quickly that could provide us with an incredible amount of feedback in a short period of time. That's what we did back in 2002 with the first version of the Google search appliance.

It essentially packaged our Web-search technologies into a box that could be deployed inside a company and provide a Google-like search capability over corporate context. And that was done with a handful of engineers and a few salespeople. A very small amount of resources. That's usually how we start anything.

At a general level, we aren't huge on market research. We believe in getting something out there, talking to customers about what it does and doesn't do for them, and about what would solve their problems at a bigger level.

So we do talk to CIOs every day, and we are very much driven by the needs of the enterprise market. But having said that, what's unique about us, I think, is we're really tapping into the pace of innovation that's happening on the consumer side. That's where I think there's a huge opportunity for us. We want to leverage innovation happening on the consumer side that is end-user focused and channel that in a pragmatic way into enterprise technology.

So what did you learn from the search appliance about the way that data is handled and searched and collated in the enterprise environment?
Well, we learned that people know how to use Google, and they like it. When you put Google inside a company, almost invariably the reaction from the employee base is exceedingly positive. There's almost a sigh of relief that now they'll be able to find their information. For us, that is always the most important thing—how well does it solve the problem for the end-user.

And then, along the way, we've had to learn how to make a product that fits into an enterprise environment and works not just for the end-user but also for the IT organization. So there are different constituencies that you have to think about when you deliver products to the enterprise. We naturally came in with this inclination to focus first on the end-user. That comes really from Larry and Sergey. But we have to bring the sensibility that there are other constituencies involved here.

Did CIOs provide any feedback or concerns?
The questions were mostly about how comprehensive is the product. Questions like: "How much can it solve the big problem for us, versus a small subset of the problem?" It was a limited product, and it could only address a certain subset of the information that needed to be searched. So we needed to add the ability to search across database content or search across secure content. And that's really what we've been driving toward for a few years now, making a product that can become the fundamental search application for very big companies.

Next page: Google's Big Idea


's Big Idea">

So how are you going to make Google relevant to the enterprise?
The Google search appliance is the basic product for searching across all the network information inside your company. And in the new version we're adding what's called OneBox. OneBox refers to the fact that when you go to Google.com and submit a query, you sometimes get a special set of results inserted at the top of the search results.

An example would be if you typed in "weather San Francisco," you would get a five-day weather forecast inserted at the top of the results. Or if you typed in "United 10," you'd get flight information about a United flight. It's just very quick information access through what is, in effect, an über-command-line interface. And we think that could be a very powerful tool inside the enterprise.

So, for Google OneBox for Enterprise we went out and talked to a lot of business application vendors about making their information much more accessible through a simple search interface. For example, we talked to Oracle, Cognos, SAS Institute, Salesforce.com, Cisco and a few others. But essentially the idea is that just the same way you could look up a weather forecast on Google, you could easily tap into your Salesforce.com system and find out about a sales opportunity.

Will Google and Microsoft Live Happily Ever After? Click here to read more.

Wasn't the Google search appliance already able to search those applications?
Yeah, it was, but if you think about general search, it's about getting a bunch of information back and relevancy ranking it into documents. So if you want to find a phone number of an employee, do you really want a document-level search where you have to look through three and four documents, or do you want it—Boom! —at the top of the screen.

Basically, it means real-time access to another system. There's no lag. It's real-time access to a piece of information that resides in another system. Yes, you could crawl your CRM system, or your Exchange system with our product and with our competitor's products and have your search term show up in a list of documents. But that's very different from the five-day weather forecast showing up at the top of the search results.

Can companies set it up to feed them whatever results they want?
Yes, because it's a development environment. Any given company may have all sorts of information that they would like to make available, and they can make it all keyword triggered. You could type the word "contact" and then a name and it would go to Exchange. It's really up to the administrators to decide how they want to trigger it.

But the user experience—and this is really important to us—entirely mimics how Google.com works. So, you don't have to get training; you can discover it over time; a friend can show you a OneBox that they think is particularly useful. For example, one of our partners is Oracle, and you'll be able to look up a purchase-order in your Oracle financial system because Google will recognize what a purchase order number looks like. Just like Google.com recognizes a UPS tracking number. The Enterprise system will know what an Oracle purchase order looks like, and it will insert that information right at the top.

Next page: Google's Lofty Goal


's Lofty Goal">

What do you think the impact of this technology could be?
We expect that search will become the preferred way to access these systems. If I want to access my CRM system, if I want to quickly grab a piece of information out of my business intelligence system, I don't have to be an expert in those systems anymore. I can type a few keywords into Google and get that information.

So, it has the potential to make all these systems that are already installed, already paid for, already being used by experts—all of a sudden that information can be much more quickly and broadly available to the rest of the company.

Let's just talk about Cognos for a second. In most companies that have Cognos—which obviously has a huge market share, it's a great, successful BI company—probably 5 or 10 percent of the employees in that company even know it's there and have access to it. And they've been trained on it and they use it every day.

Well, there's probably another huge percentage—40 or 50 or 60 percent of the employee base—to which that Cognos system could have a lot of value if it was accessible, and if they could get the information in a short, meaningful way. But they're not experts at building cubes or extracting information from a BI system.

But a company will get a lot more usage out of systems it's already paid for, and I think that's probably the most important thing for people to consider. It is a lightweight, fast, easy way to provide access to systems which, until now, have only been under the domain of the experts in those systems.

What about information outside the enterprise?
Let's say your company has signed up for premium content from somebody, say a some market research firm. Instead of having to log in to the system from IDC, or wherever, why shouldn't that just show up at the top of my search results?

Even that's a huge improvement, because it used to be that you had to call your research person, who would send the report to you. But now I have a searchbox in the middle of my page, and I can access the IDC reports that my company's paying for with a few clicks of the mouse. Or keyboard.

So this helps to find information. Will Google do anything to help companies work with the information they find?
I've been asked questions like that before. Questions like, "Will this be the end of business intelligence?" And my reaction is that asking a search engine, "Go tell me which of my customers are the most profitable ones," is an incredibly difficult artificial-intelligence problem. If I assigned that to a staffer, they would need to ask me six or seven qualifying questions before they could find that out. So search isn't going to have that level of impact in the next few years. In ten years, who knows? But, to have an engine that has that level of omniscience is pretty far out there. It's certainly something we dig into.

Having said that, a little more here-and-now and pragmatic is, if you wanted to ask, for example, "When might George, Mike and I all be available today?" Obviously, a search engine can do that if you can imagine a Microsoft Exchange or some form of calendar plug-in that could do that from a simple query. That's simplistic, but I think search will do more and more in terms of not just looking for information but actually taking action.

So what's next for Google?
We're still in the first or second inning of search, so we'll never be done with that. But we really want to reach the point where, in your average company, a very high proportion of the company's information of all flavors, structured and unstructured, is available through a single search box. And like I said, we're very early in that game, so there's a lot of work to be done there.

Does Google want to become the operating system?
Well, I can say that, for years, most of the business-application vendors have been moving their primary interface to a browser-based interface from the clients. So maybe we're just the next generation of that trend. Now that everything is, in effect, a Web app, all of a sudden a simple Web-style search engine that you know and love may become the preferred jumping off point to get into those applications. That's how I see things evolving here. It's the next generation of Web technology pushing its way into the enterprise.