Today Hewlett-Packard (HPQ), the valley’s largest tech company, wrestles with a similar problem. Though HP’s advanced research group once invented wonders such as the scientific calculator, the thermal inkjet printer and commercial LED lighting, these days executives feel HP Labs and its $150 million annual budget could do more to boost the company’s bottom line.

This issue has fresh urgency now that HP has topped $100 billion in annual sales. The company will need big, marketable ideas to fuel future growth. Can HP Labs deliver?

The challenge falls to Prith Banerjee, a noted scientist and startup veteran who joined HP last May as director of HP Labs. Banerjee, 47, was previously dean of the College of Engineering at University of Illinois at Chicago; he has also founded two electronic design software startups. AccelChip was sold to Xilinx early last year, and he still serves as chairman and chief scientist of Binachip.

At HP, Banerjee wants to transform the labs from an idea farm into a hit factory. Rather than send the 600 HP Labs scientists out to pursue dozens of personal-interest projects, Banerjee will encourage them to pool their talents and tackle big problems that are likely to yield money-making discoveries. Shane Robison, HP’s chief strategy and technology officer, says the retooled organization ideally should help add hundreds of millions of dollars in revenue to HP’s core businesses of personal computing, imaging and printing, enterprise systems and software.

A Different Focus

“HP Labs today does a lot of cool things — about 150 research projects – but it does so many things that none of the projects get enough resources to have a high impact,” says Banerjee. “My message was, let’s try to narrow the focus” to about 30.

Banerjee’s vision borrows heavily from the culture of Silicon Valley venture capital firms, which fund fresh new ideas and try to nurture them into the next Intel (INTC) or Google (GOOG). Traditionally, HP Labs scientists have had little incentive to transform concepts into products. But in the future, research efforts that fail to meet benchmarks won’t get more funding, according to HP executives; instead, more resources will go to larger projects that are meeting goals and are likely to mature into profitable new businesses.

Much like in a VC firm, once a project is ready to begin the process of becoming a product, researchers who worked on it will move to the product team. “When a startup company gets formed, you don’t just bring in the technology,” Banerjee says. “You convince the people who created the technology go along, and you bring in people who are experts in building products. That’s the same thing we plan to do at HP Labs.”

It’s a colossal shift in the way innovation happens inside a Silicon Valley giant, and competitors will be watching. If Banerjee succeeds at leading the reinvention of HP’s research culture, it could both drive earnings and provide a blueprint for other tech behemoths that struggle with growth. But this is also not the first time a mature enterprise has tried to change its culture and adopt the creativity of a nimble startup. Perhaps because big companies are run by managers and not by entrepreneurs, these kinds of efforts typically fail.

Still, HP can find inspiration in stories like Chandrakant Patel’s. As a Fellow in HP Labs, Patel began experimenting a decade ago with large-scale cooling technologies, based on a gut sense that the Internet would give rise to energy-guzzling data centers full of computers that would run too hot. Though HP managers didn’t grasp the importance of his research at first, Patel soldiered on. Using social skills he honed during college when he took a job selling encyclopedias door to door, Patel found allies within the company who helped him find a market for a new kind of software-driven thermostat system designed for server farms.

Today, Patel’s project has bloomed into a product called Dynamic Smart Cooling. And since electricity costs have become the most expensive part of running a data center, demand for the invention is obvious in the global warming era; HP expects it to generate hundreds of millions of dollars in sales.

Another project called BRAIN, a system for predicting the future through betting patterns, also offers an example of the type of work HP Labs wants to encourage. HP economists and engineers worked together to design a system that analyzes a group of people’s bets in a given market, factors in each person’s appetite for risk, and uses that information to predict what will actually happen. Though the project began as an academic exercise, Bernardo Huberman, an HP Labs manager, was determined to put it to profitable use. The financial division suggested using it at the beginning of a quarter to predict what revenues would be at the end; when that experiment went well, HP began using it to predict component prices, helping the company to buy the critical building blocks of its products at the lowest cost and thus increase profitability.

Uncertainty Ahead

HP is not alone in its ambition to improve the profitability of its advanced research efforts. At Cisco (CSCO), the company has also been working on system to get more bang for its R&D buck. “Innovation typically is when you take a number of inventions and you put them together to create something that is disruptive,” says Cisco senior vice president Marthin De Beer, who leads the company’s Emerging Technologies Group. “We really focus now on the latter part — on innovation.” That focus means Cisco more widely solicits ideas, chooses the best to develop, and creates new startup-like business units to incubate them. The jury is still out on how well the system will work for Cisco, though one of the first products to emerge, TelePresence video conferencing, has generated buzz.

Likewise, it will take years before investors can be sure whether HP’s revamp of its labs is bearing fruit. Many of the research areas that the company identifies in 2008 won’t yield products until 2013, though the new model could produce some benefits sooner. For instance, HP’s personal computing group has set up a team called the Innovation Program Office that seeks to grab good ideas from inside and outside the company and turn them into products more quickly.

Along the way, there are many things that could go wrong as companies try to improve the batting average of advanced research. Chuck Geschke, who left Xerox PARC with his friend John Warnock and 25 years ago co-founded Adobe Systems, notes that no one has yet cracked the unique code for making advanced research consistently pay off. “There must be something special, because not everyone’s able to do it as well as they anticipate,” Geschke says. And perhaps there is no perfect formula at all. “You have to accept the fact that if you’re really pushing the envelope, some of what you do won’t work out.”

HP executives are conscious of the balance between focusing the research and leaving room for the unexpected. Phil McKinney, chief technology officer for the PC group, said Banerjee and the company’s chief technology officers have been working to make sure that under the new structure there’s room to pursue ideas that at first blush don’t seem like home runs.

“You want to be careful that you don’t become so restrictive in your strategic direction that you end up with blind spots,” he says. “How do we give enough flexibility in the model to let the researchers work on something that they’re personally interested in, and they think it might be something, but it’s very early-stage?”

It will be important for executives to get this right. The researchers who work in the top advanced technology labs are highly sought after, and will have no trouble finding work if they become unhappy at HP.

As Patel talks about his work with data center cooling, a similar theme emerges. He was free to pursue his research even before the company saw its value, and that freedom is one of the primary reasons he has remained with the company for more than 20 years.

“My boss could have said, ‘Why are you wasting time on this?’ But to his credit, I can’t remember anybody in HP Labs questioning me about all of this modeling that I was getting into,” Patel says. “What remained constant was autonomy. Nobody really questioned me.”

As HP Labs undergoes a transformation, there certainly will be more questions. But executives hope they’re the type that spark innovation rather than stifle it.