HP Labs braces for impact

Palo Alto, Calif. -- Hewlett-Packard Labs director Prith Banerjee has been a busy man.

Since joining the company in August, he has reorganized HP's storied R&D organization into 23 labs, each with a new director focused on high-impact projects. He has put a review structure in place to link HP's business units with the 600-person research division, and he has set up new programs to connect HP Labs to universities and the venture capital community.

Banerjee is getting kudos from around the industry for the reorg, announced with much fanfare March 6. The moves should energize and open up the group, which some had seen as too insular and lacking in impact.

Now for the hard part.

The bigger challenge for both HP and archrival IBM Corp. is defining the role of the modern computer company. Intel and AMD will increasingly dictate system-level hardware in their multicore processors; Microsoft and Google will drive the software stack and Internet platform. Squeezed between them, HP, one of the world's biggest electronics companies, needs to carve out its own space to innovate. By all accounts, however, it's off to a good start.

"In general, I think [the HP Labs reorg] is a very positive change," said Bill Dally, chairman of the computer science department at Stanford University, who met with Banerjee recently. "Previously I didn't think HP Labs was appropriately connected outside--to academia--or inside, to its own business units. They were perceived as somewhat insular. They didn't publish much and were seen as something of a black hole."

"I'm impressed; it's a bold change from where they were," said Mark Dean, an IBM fellow who runs IBM's 400-person Almaden Research Center in San Jose, Calif., part of IBM's 3,000-strong research division. Banerjee, Dean said, is "well respected and obviously willing to make changes and take risks, which is important in research."

"HP needed to do something to be competitive. You have to find a way to contribute to the corporate success, and you can get caught if you are not ready to adapt," Dean added.

Click here for larger image

IBM spent $6.2 billion in 2007 on what it calls R&D and engineering--an expenditure that has risen steadily at the company over the past five years. By contrast, HP spends just $3.6 billion, a figure that has edged up marginally since 2002, despite the company's revenue rise from $72 billion to $104 billion over that time.

In addition, IBM has maintained its position as the company with the most patents granted each year, while HP has bounced from No. 3 to the fifth and ninth slots during the period.

The plan
To turn the tide, Banerjee has directed each of his 20- to 30-person labs to pitch big projects that combine fundamental research and advanced development. The first round of proposals, filed last week, will be reviewed by a board comprising equal numbers of lab researchers, HP business unit managers and technologists from around the company.

The board will meet quarterly and review approved projects annually. A new technology transfer office in HP will help commercialize the labs' work, and a new Web site will share some of the group work publicly.

"This represents a huge step for sharpened focus at HP Labs," said Banerjee, an entrepreneur and former dean of the college of engineering at the University of Illinois at Chicago. "Our resources remain the same, but we are reallocating them to fewer, bigger bets to improve our return on investment."

"There are a lot of mini transformations going on all over HP, and this is one of them," said Mark Hurd, HP's chairman and chief executive, speaking at a March 6 coming-out party for HP Labs at the company's Palo Alto headquarters.

Banerjee also aims to increase the amount of basic research done at the lab. "We used to do less than 10 percent of our work in blue-sky exploratory research, but that will a third of our work going forward," he said.

"This is one of the places where there is still some real R left in R&D," said Hurd.

The division is also adopting the approach of nearby Sand Hill Road venture capitalists by starting an entrepreneur-in-residence program. Foundation Capital (Menlo Park, Calif.) will be the first to base one of its executives at the lab starting later this year.