IT’s Role in Collaboration and Innovation at Procter & Gamble

Procter & Gamble is famous for being innovative, but the hard truth is that it had better be. The company spends 3.4 percent of its revenue on R&D, more than twice the average of 1.6 percent in the consumer packaged goods industry. But big spending on R&D does not guarantee success: A study by consultancy Booz Allen Hamilton found no correlation between dollars spent on R&D and profitability. What matters is the productivity of that spend—the “hit rate” of ideas that lead to products.

The company behind Swiffer and Crest toothpaste assesses its hit rate at about 80 percent—pretty good, considering that the overall industry success rate is just 30 percent, according to ACNielsen. But in consumer products, the real blockbusters—products that aren’t merely new variations or tweaks of existing products—are few. According to research company ProductScan, of 10,649 new product introductions across the industry in 2005, just 484 were truly innovative—meaning they added significant new benefits in areas such as formulation or technology. And the creativity of the big consumer goods players is waning: A study by McKinsey found the top seven companies, including P&G and competitor Unilever (whose products include Surf detergent), accounted for just 5 percent of all the patents filed for laundry and home-care products between 2000 and 2005.

No wonder then, that the 170-year-old manufacturer has been driving hard to improve its R&D productivity since CEO A.G. Lafley took the reins in 2000. Lafley’s proclamation at the time, that by the end of the decade P&G would get at least half its new product ideas from external sources, was hailed as visionary. But really it was a response to the reality that the days of “not invented here” are over. Today, Lafley claims that outsiders are involved in developing 35 percent of new products, and P&G’s spending on internal R&D has decreased 30 percent in seven years.

But P&G needed to do more than bring more cooks into the kitchen. It needed to change a perfectionist mentality in its research culture. “The front-end ideation process took time,” says Steve David, who retired from a 34-year career at P&G in 2005, the last five years of which he spent as CIO. The company began rewarding researchers for speeding up experiments. “The goal was to encourage people to fail early and often and if the idea isn’t working, either kill it or move it into another organization and let them work on it,” recalls David, who is now a senior adviser to The Boston Consulting Group.

However, increasing the size and speed of the idea pipeline increased the complexity of managing the R&D process. Besides its own R&D organization of 8,000 scientists spread across 28 sites globally, P&G has outsourced portions of its R&D processes, such as routine chemical lab experiments, to lower-cost countries. The company has also begun collaborating more with outside technology companies, who pump new product ideas into P&G, collaborate with P&G scientists or simply develop products themselves with P&G’s guidance and investment. In addition, the company uses its internal network and the Internet to connect with loose networks of research scientists who help solve vexing problems that are getting in the way of new products—a chemical formulation, for example. Coordinating work and screening new ideas are critical to keeping the pipeline moving. “We need to run our global R&D organization as though we’re in a single building,” says Keith Caserta, associate director and head of health care R&D information and decision solutions with P&G.

This is where IT comes in. It takes an army to manage all these constituencies, making coordination and collaboration critical. “Opening up the idea process introduces more fragmentation and friction,” says Navi Radjou, a vice president for research company Forrester. If IT is going to reduce the complexity of managing innovation in a company as widely distributed and diverse as P&G, it has to meet some difficult requirements.

First is flexibility. “We need IT that is flexible enough to meet a broad range of business needs, because we do chemically based products, biologically based products and assembled products,” says Caserta. Next is scalability. “Technology has to stand up to thousands of users in a global installation, and that’s been challenging in some cases,” he says. Finally, the solution should promote cross-functional integration and collaboration.

In an attempt to address all of these requirements, P&G is experimenting with product lifecycle management (PLM) software (which is traditionally applied to the product development phase of R&D for assembled products like diapers and razors) in the experimental research portion of all R&D processes, assembled products, and chemical, biological and mixture-based products. PLM is like ERP for the product development process: a big, feature-rich software platform that companies typically try to standardize across the entire enterprise. P&G has a pilot with vendor UGS using PLM as a backbone for storing and sharing its researchers’ “lab notebooks”—the records of their experiments that are almost always created on paper. If the system becomes a reality, the lab notebook will become an electronic talisman that links internal P&G researchers across P&G’s many global labs, as well as those of its contractors. The project is a reflection of the larger push within P&G to create a more connected, global innovation process.

Converting those notebooks to electronic form is no small matter for most researchers, who see any attempts to standardize or automate the research process as a threat to their creativity—even though they understand the benefits of sharing. “They are resistant to change of any type, and anything that they think affects their freedom they will fight,” jokes Charlie Cruze, systems manager for P&G Pharmaceuticals and a former researcher.

The resistance to IT-based processes has persisted to the point of silliness. Some researchers at P&G are known to write up their experiments using Microsoft Office applications and then glue printouts of their work, page by page, into the notebooks, making them look like witches’ wrinkly cookbooks. Less humorous is the way the notebooks stay hidden on researchers’ shelves while their colleagues in other P&G labs unknowingly duplicate their work, or how the notebooks can be spirited away for months or even years by lawyers who need them as evidence for patent cases.

To accommodate researchers’ discomfort with automation during the PLM pilot, Caserta and his team have been careful not to disrupt how researchers get their data into the system. The researchers can enter the information any way they want, whether in Microsoft Word or through legacy systems that are integrated with the electronic laboratory notebook (ELN) system. The plan is to allow researchers to throw in all sorts of documents—what technologists call unstructured data. They can then apply descriptive tags to the files. The system then converts the files to PDF format, allowing them to be searched and shared by anyone with access to the ELN system. The PLM software is already installed as the global standard for a number of areas in product development across all of P&G. By hooking into this existing system, the lab notebooks become visible to functions farther downstream from the researchers, such as engineers in product development, while functions with tangential involvement in the process, such as the patent lawyers, can access the data without removing it from the system. As a result, P&G increases cross-process integration, what Caserta calls “the horizontal view,” of the innovation process within the company.

No less important are the subtle attempts to automate the research process itself. Templates that automatically display the researcher’s name and the project don’t assault researchers’ delicate sense of independence but do demonstrate a direct benefit by sparing them from rote duties. That paves the way for acceptance of more standardization and automation of the process down the road. “We can template the standard, repetitive experiments so all they do is fill in the numbers,” says Cruze. However, he adds, automation doesn’t merely save individual researchers’ time. It opens up projects to collaboration with researchers in other labs, who can see where one team left off and continue the work.

The pilot is small—only 75 scientists spread among a limited number of U.S. and international locations—and things change dramatically when you attempt to roll out complex enterprise software to thousands (5,500 is the target globally). Meanwhile, trying to create a platform that can meet the needs of biotech researchers as well as engineers designing diapers may be a stretch. “To go from product idea all the way through to [product retirement] is a long time frame and maybe asking too much,” says Robert Cooper, professor of marketing at the DeGroote School of Business at McMaster University. “I’m skeptical that they can hook it all together. It starts to become a 1,000-pound marshmallow.”

The stakes are big. Improvements in speed and productivity in the early idea phase of R&D can have an impact all the way down the line. Concludes Caserta: “If you can improve your productivity in the idea phase, it gives you the opportunity to have a higher success rate at every stage after that.”