Microsoft is reorganizing part of its research-and-development operation to create new products faster, and to compete with the seemingly vast array of innovative consumer software and services that companies like Google and Yahoo bring to market on a weekly basis.

Its new organization, called Live Labs, consists of some 85 researchers drawn from two existing divisions, Microsoft Research and the Microsoft Network (MSN). Both organizations are heavily involved in creating new Microsoft offerings, such as MSN Search, introduced last year. But Live Labs is designed to act as a “perpetual startup” within Microsoft, in the words of the organization’s new director Gary Flake – an incubator where software engineers can rapidly test ideas for Web-based services and other software, then shepherd the best concepts to market.

The “Live” label comes from Windows Live, a still-embryonic set of Web-based services announced by Microsoft chairman Bill Gates in October. Windows Live is separate from the company’s other Web offerings, such as MSN.com, and is intended as a central platform for personalized Internet services – current offerings include news and weather information as well as e-mail and instant messaging – and capabilities such as Web publishing that enhance Microsoft Office and other existing products. Live Labs will use Windows Live as a testing ground for more Live services.

It’s all part of a broader attempt within Microsoft to adapt to the reality that much of the software that people now use every day – from Google’s Gmail service to Salesforce.com’s online customer relationship management tools – is delivered via the Web and runs inside a browser.

“There is a sea change going on within Microsoft,” says Kevin Schofield, general manager of Microsoft Research. “We are looking at how software and services are offered to people and companies, and it’s changing. A lot of the work we’re going to be doing in Live Labs is applied research about how we execute on that. In the world of creating ‘Live’ services and software…you want to be fast and nimble and try new things.”

“Nimble” is probably not a word that comes to most people’s minds when they think about Microsoft. The company is not known for rushing innovative new products to market to counter its competitors. Instead, Microsoft has often taken months or years to develop its own versions of popular products – for instance, Internet Explorer 1.0 appeared eight months after Netscape Navigator 1.0, and MSN Search was launched seven years after Google rewrote the rules of Web searching (and has relied on the ubiquity of the Windows operating system to help spread them).

The formation of Live Labs appears to constitute an admission by Microsoft that its traditional, gradualist approach to research, code development, testing, and marketing is not well suited for an era when younger competitors post beta versions of latest software on the Web almost as soon as their programmers have dreamed them up, then let them evolve in response to user feedback.

Google, for example, asks its technical staff to spend 20 percent of their time developing software concepts that might evolve into new businesses for the company, and tries out many of these ideas on the public through its Google Labs site.

Microsoft, of course, is not portraying Live Labs as an imitation of Google’s or any other company’s research models. “We obviously keep a close eye on Yahoo and Google and the little players and startups, and we’re always looking very broadly at how other people are organizing themselves,” says Schofield. “But we love the way we do basic research and the way we do product development. The question was, ‘How can we do this more and do it better?’.”

The idea of building a startup-like applied research organization within a large company is neither new – nor has it been uniformly successful. Business analysts such as Clay Christensen, author of The Innovator’s Dilemma, have long stressed the difficulties encountered by inventors who try to promote “disruptive technologies” that might undermine their companies’ sales of existing products. But Schofield says Live Labs will embody “a Microsoft twist on the startup mentality,” meaning that it will have close ties with the scientists at Microsoft Research and developers behind MSN.com, where some of Live Labs’ products may appear, and that its researchers will be able to lean on Microsoft’s existing infrastructure for necessities such as capital and human resources.

How would the Live Labs model work in practice? As an example, Schofield mentions one area Live Labs is likely to explore: searching multimedia content on the Internet. “Today, there’s lots of video recordings out there on the Internet with audio tracks, and you’d love to be able to index all of that great video content just by their audio tracks. We have speech recognition technology to do that, and it’s getting more advanced. But suppose you really wanted to build an Internet-scale multimedia search engine. A key research question is: What’s good enough? Is there some way we can make the search a little fuzzier, so we don’t have to get the transcription exactly right? We could do a lot of research and write a lot of papers on that, but, you know, if we build something and it works that gets you a lot closer toward shipping it as a product. That’s the kind of value we see in Live Labs, and that’s why it’s a deep partnership between MSN and the research groups.”

Schofield says executives at every level of the company, including Gates and CEO Steve Ballmer, have bought into the Live Labs concept. “They’ve been interested and signed off on it,” he says. “Bill spends a lot of time with MSR as it is, and I’m sure he’ll be spending a lot of time with the Live Labs folks, understanding the projects they’re working on and new products they could do.”

Yet some industry observers consider Live Labs too late to help Microsoft compete with the likes of Google. “I don’t think anything like [Live Labs] is going to have any effect in a time horizon that will be relevant to the challenge they are facing from Google,” says Charles Ferguson, an investor and journalist who developed the Web-authoring program Front Page in the mid-1990s and sold his company to Microsoft. “Someone coming up with some great idea [at Live Labs] that seven years from now will be a significant market – that’s way too late.”