Can you really herd researchers so that they focus on the most pressing needs of entrepreneurship? That seems to be the belief driving Hewlett-Packard as it announced today that it has revamped its vaunted HP Labs central research division.
Central research labs have not had a great history. Xerox famously failed to act on innovations like the mouse and the personal computer, fumbling the future to Apple. HP has had better luck, creating new markets with research inventions such as the laser printer and inkjet printer. But Mark Hurd, HP’s new CEO, is a famously results-oriented manager and the remake of HP Labs clearly reflects his get-with-it attitude. HP Labs, smaller than it used to be, is clearly under the gun to deliver.
Today, the tech executives are out to show that they get it and can move fast on the most pervasive trends that start-ups are embracing. They are talking about crowdsourcing, working with universities, and working closely with venture capital firms such as Foundation Capital. (Clearly, with Tom Perkins no longer on HP’s board, the VC firm Kleiner Perkins Caulfield & Byers no longer has a premiere spot within HP).
HP has revamped its central research centers to have a sharper focus on larger projects in five different areas: sustainability, information explosion, dynamic cloud services, content transformation, and intelligent infrastructure. The company says it will organize itself so that it can accelerate the conversion of research into commercial ideas.
“Balance exploratory research with an entrepreneurial approach,” said Michael Mendenhall, chief marketing officer at HP.
Prith Banerjee, director of HP Labs, said HP will focus on 20 to 30 big bets rather than 150 or so smaller projects that it used to work on.
“We hope to solve the most complex, challenging problems facing our customers and drive HP’s growth,” he said.
About a third of research will be basic, a third applied, and a third in product development. Now there are 600 researchers in 23 newly formed labs in seven locations. (I can remember when it had many times more that — at least 900 not long ago and many more than that a decade ago). IBM, by contrast, has 3,200 researchers.
Each of the 23 labs has 20 to 30 researchers, Banerjee said. Shane Robison, chief strategy and technology officer, said that this is “not a cost-cutting exercise.”
Some context is useful here. About half of HP Labs’ original team went to Agilent when the big company split into two parts in 1999. But HP’s R&D spending peaked at $3.7 billion in 2003. It has been relatively flat at around $3.6 billion for the last four years, even though revenues have dramatically expanded. HP’s rank on the top ten patent recipients in the U.S. has fallen from No. 3 in 2005 to No. 9 in 2007.
The interesting thing is that there is no obvious research area focused on hardware. The first area of focus, information explosion, is aimed at helping people sort things out. “Developing a better way to acquire, analyze, and deliver information so it can be acted on,” said Banerjee.
The new projects include Face Bubble, a face recognition technology is based on digitized content. It allows you to search through a bunch of photos and videos to find what you’re looking for. Another focus area is dynamic cloud services. These include CloudView, which provides instantaneous data to mobile devices.
Content transformation is another focus area. This is about transforming analog content into digital assets across a bunch of devices. It’s also about moving digital content to other digital devices or to the physical world. Another area of focus is intelligent infrastructure. That means “new smart, computing devices, networks and architecture, that are secure and operate at Internet speeds and scale,” Banerjee said.
Lastly, sustainability means creating products for the “lower-carbon economy” where you can address waste issues from cradle to grave.
HP is also announcing an “open innovation” program with university alliances, an entrepreneur-in-residence program, and HP IdeaLab, which allows a peak at early stage research at HP. Anyone can help HP out with projects. This move resembles Eli Lilly’s own InnoCentive program, which posts problems the company needs to solve and rewards outsiders who submit the solutions to those problems.
The university partners will be challenged to come up with projects in HP’s five focus areas. HP will fund 100 PhD students. The EIR program resembles the venture capital industry’s practice of using entrepreneurs to help them see new trends. Foundation Capital in Menlo Park will be the first to participate in the program.
“We plan to know what the market trends and business opportunities,” Bannerjee said. “I want to inject start-up DNA into HP Labs.” HP IdeaLab will make technologies available on the web so that independent researchers can add features to the technologies. It’s an idea that resembles the “crowdsourcing” ideas executed at places such as Eli Lilly.
For tech transfer, HP will set up an office to speed innovations to market and license it to whoever wants it. One of those ideas to be licensed is Memory Spot, a tiny chip that allows you to transfer information on documents and other small objects.
Mark Hurd, CEO of HP, was the headline attraction. But he appeared on stage for just a couple of minutes. He said HP Labs is one of the places “where there is still R left in R&D. Most of it out there is D. We are placing major bets on the R.” The entire IT ecosystem will support the creation and distribution of digital content. He said the entire company is excited about an ongoing transformation at HP and that this “mini transformation” within HP Labs will generate results.
Robison talked about “everything as a service” as I pointed out in the earlier post. The Internet is a real platform and cloud computing is real. That means companies can launch all sorts of applications that don’t depend on any particular device and instead can run on anything. He also mentioned BRAIN, a kind of wisdom of the crowds prediction service. HP is using this to make more accurate financial forecasts and it involves tapping the collective smarts of the company’s employees. Robison said that it was able to more accurately predict the pricing of dynamic random access memory chips, which HP buys in the the tens of millions.
I tried out a little game where a group of reporters placed bets anonymously using the BRAIN system. You make a prediction, put your money down, see the results, and then you get a ranking for how good you are. (And we won!) . Prediction markets, of course, all all the rage among big and little companies.
His five trends to watch include:
1. The digital world will converge with the physical world.
2. The era of device-centric computing is over.
3. Publishing will be democratized. (Amen to that, baby).
4. Crowdsourcing is going mainstream and will change the rules of the game forever.
5. Enterprises will use radically different tools to make key business decisions, including systems to accurately predict the future. This idea comes from the ideas popularized by James Surowiecki’s 2004 book “The Wisdom of Crowds.” Xpree is a startup that recognizes this, and HP is a giant that has learned the same thing.