Microsoft's Information Worker unit has warmed up an internal incubator to foster new product and business development.

"Project Greenhouse," with about 30 to 50 people, is one of several incubators within the company, said Bobby Kishore, general manager of Greenhouse. The group got underway in 2003, but now has a more formal way to vet ideas and decide on funding, Kishore told CRN. (Project Greenhouse is not to be confused with 'Project Green' the code name for Microsoft's next-generation ERP product code base.)

Kishore, a Microsoft veteran, took over the reins when Peter Rinearson, a corporate vice president in the Information Worker Business Group left last year.

A Microsoft source said the rationale is to encourage employees who have good ideas, but not the Ph.D. required for a Microsoft Research gig, to come forward. The last thing Microsoft wants is to lose some great new notion to Google or a startup, said this long-time Microsoft watcher.

Some say Microsoft has grown so large and bureaucratic that it discourages innovation--hardly a problem unique to that company. Greenhouse is at least an attempt to correct that.

Another Microsoft insider said Greenhouse is modeled on a similar program at Intel but said it lacked "real teeth."

There are smart, capable people there, but "they are ultimately hampered by the fact that most of them are [Microsoft] newbies and therefore don't have the relationships and tech knowledge that is needed to work behind the scenes, which is really necessary to get stuff done at the company at this stage in the game," this veteran said. Additionally, interesting work that Greenhouse might accomplish can run into direct conflict with existing products and "resultant turf wars tank the new stuff," this veteran said. Ringcam reportedly came out of the Greenhouse endeavor.

Meanwhile, Microsoft's latest arch-rival Google has won plaudits for encouraging employees to explore ideas that are often not initially viewed as part of its core business. Sometimes these "orthogonal" ideas pay off, observers say.

Kishore said one key goal is to separate what could be a net new product from what would be better considered a new feature for an existing offering.

"We have to make sure the elements are right. Ideas can come from anywhere [in the company]. People can come in with an idea and we then see how it fits. Is it just a feature? Or is there something meaty there that could be a product?" Kishore said.

What the idea has to have in its favor is that needs to be "disruptive" he said.

If it does have the makings of a new product, the team moves to a rapid prototype stage. The prototype is then evaluated and if it passes muster, the team has to come up with a complete plan as to who will build it, ship it and support it, all the resources needed to follow through, Kishore said. As per Microsoft's mandate, the team must early on figure out the impact on partners and what opportunities it may spark. There the group has to tread a fine line. "We want feedback, but we also don't want to go out there to partners with something that is very speculative, where we're not sure of the go-to-markets. I don't want to start promising the world to someone and get them to invest when we're not sure on the plan yet," he said.

Kishore would not disclose specifics about current projects, but said the group did work closely with the SharePoint team, Microsoft Research and MSN desktop search team to deliver desktop search. Other current work will surface in the Office 12 wave, he said. Office 12 and its related client, server and services offerings are slated to roll out in 2006. The company is quietly working on a range of new Office-labeled servers, including what was once called InfoPath Server, and is now being called the Office Forms Server, sources said.

If an idea is accepted, the person who brought it forward gets the freedom to test it out. "No one's breathing down your neck on a daily basis. The dependencies [between product groups] are not there," Kishore said. The real benefit is being able to watch the concept from birth to execution and see it have an impact, he said. Material rewards are another matter. "Will someone in Greenhouse become a billionaire? I don't think so. But if I'm driven by money, then maybe I should go somewhere else. If I want to do something for millions of users that have an impact, I can do it here."

Initially Greenhouse relied on an internal "board" from across the company to OK projects and allocate funding. "It served a very good purpose but at that point there was no solid muscle memory or empirical evidence to determine what kinds of ideas we funded or didn't fund. We've now matured . And are focused more internally," Kishore said. Now decisions are made by the GM, group vice president Jeff Raikes and the Information Worker senior leadership team, Kishore said.

As is usual for Microsoft, the company researched and set out guidelines for new projects. A green-lighted project should address one of four target areas: information overload and mobile access needs; collaboration or information sharing; easing tracking and management of multiple projects; or connecting business people and their processes.

Some observers say such strictures negate the whole point of an incubator, which ideally should come up with what is totally new and different.

Microsoft is caught in a dilemma, says Paul Degroot, analyst with Directions on Microsoft, a Kirkland, Wash., research firm. On the one hand, the company is trying to do so much now it is seen as having no focus and increasingly has difficulty bringing much-hyped products--Longhorn, Yukon, Whidbey--to market in a timely manner.

On the other, some product groups, who've been told to focus on their core businesses, often cite customer research as showing users don't want X new product or feature. "So, management declines to develop something, and it turns out yeah, maybe their existing customers don't want it , but maybe new customers would want it," Degroot said.

Degroot also sounds the now-common refrain that Microsoft has difficulty recruiting and retaining key people and keeping them excited about their mission. "The company is becoming like IBM. It's really large and one look at the Longhorn timeline shows you that even getting innovative, exciting technology out of the company can quickly turn into a marathon."

No one should forget, Degroot says, that Longhorn was due in 2004. Now the client is slated for 2006 and the server for 2007.