Skip to Content

Old Dogs Can Learn New Tricks

Rethinking Corporate Research: Why and how IBM restored its world-class labs to business relevance.

In my previous column, I argued that the model of the central research and development laboratory had become obsolete. But that raises an obvious question: what will replace it?

We’ll find some clues by studying one classic monolithic R&D organization that transformed itself in exciting ways. When it comes to organizing industrial research, IBM is an old dog that has mastered many new tricks.

The Old School

A decade ago, IBM Research typified the traditional R&D model. Its labs were set up far from its business operations, and they reported to IBM’s CEO separately from IBM’s business units. IBM essentially funded research through a “tax” on its business units.

Its research mission was scientific in nature, so much so that IBM Research was organized very much like a university science and engineering school. IBM had departments of mathematics, materials science, computer science and physics that hired, managed and promoted research staff. IBM competed with universities and national and industrial labs for the brightest scientific graduates. It wasn’t unusual for an IBM research scientist to spend his or her entire career on the properties of one class of polymer.

Rewards reflected this approach. IBM rewarded its scientists on their accomplishments, both to the scientific community as well as to IBM itself. In those days, it was fine to excel in just one of these two areas. Researchers were advised to invest in their scientific reputations in their first years at IBM. They often declined to work on technology transfer issues to move research discoveries into IBM’s businesses. They thought: “I didn’t come to IBM to fight fires on the manufacturing line” or “Working out the operational details would really cut down the number of papers I could publish this year.”

This approach to research generated some notable scientific achievements, including five Nobel prizes, six National Medals of Science and Technology and hundreds of other scientific awards. IBM also benefited from important technological breakthroughs that emanated from its research laboratories.

Getting the Big Blues

But all was not well within the IBM Research Division a decade ago, and within IBM itself.

IBM was a deeply vertically integrated company, and so was its research division: IBM research fed into IBM product development, which was built in IBM manufacturing facilities, which were sold through IBM distribution channels and supported through IBM services and support structures. This deep vertical integration worked well for many years, but by 1992, the formula was obsolete.

IBM Research watched some of its breakthroughs, such as RISC processor architectures and relational databases, fuel the rise of corporations such as Sun and Oracle, while IBM’s own businesses languished. In 1992, the company reported the single largest loss ever for a U.S. corporation. Plans were made to break up the corporation into smaller, more focused businesses. Such a break-up would have devastated the central research group.

Taking Steps

Under this dire threat, IBM’s research division restructured and took the following major steps:

1. Leveraging intellectual property. Since 1993, IBM has been awarded more U.S. patents than any other company. IBM aggressively enforces its intellectual property. In 2000, it received over $1.7 billion in royalties.

2. Restructuring staff. The academic departmental boundaries are gone. No longer can IBM researchers blithely cultivate their scientific reputations and be indifferent to their impact on IBM’s bottom line.

Many IBM research managers now wear two hats. The first hat, their own area of research, remains. But the other hat is to act as a relationship manager between the entire research division and one of IBM’s businesses. If that IBM business unit is seeking a research answer to a pressing problem, the job of the relationship manager is to locate someone in the Research Division who can answer it. So IBM research managers are now more than knowledge generators-they are knowledge brokers. This second role broadens their understanding of their assigned business unit and helps move research discoveries out of the lab and into the market.

3. Changing funding. While much of IBM’s research budget still comes from corporate, a significant and growing percentage of funding comes directly from IBM business units. As a result, IBM researchers are now more sensitized to the needs of IBM’s businesses. Not surprisingly, these businesses are also working more closely with IBM researchers, since these funds now flow directly from their P&<.

4. Connecting researchers to customers. IBM’s First of a Kind program assigns an IBM research scientist to a carefully selected customer, to develop a solution to a customer problem. This solution is really a prototype, and IBM negotiates to receive the rights to the ideas that emerge in order to offer them to other customers later on. This approach boosts IBM’s experimental capacity and, more importantly, links that experimentation to real customer problems. This program recently expanded into the Emerging Business Opportunities program, in which IBM Research works with customers to create advanced solutions to complex problems.

5. Opening up to the outside. Another critical change has resulted from IBM’s rethinking of its deep vertical integration approach. IBM invented some of the fundamental computer languages, yet today it devotes over 2,400 of its staff to Java and related areas, which originated outside IBM. Similarly, the company is making a substantial commitment to the open-source Linux operating system.

IBM continues to produce breakthroughs on its own in technologies such as copper-interconnect technology for semiconductors and Giant Magneto Resistive (GMR) heads for disk drives. Today, though, IBM licenses or sells its technology on the open market, even to companies who compete with other parts of IBM. The technology area within IBM is one of the fastest growing parts of the corporation, along with IBM’s services business, which will service and support equipment and software from any company. As Research Director Paul Horn told me, “this gives IBM more channels for its intellectual capital to get to market.”

6. Increasing the flow of ideas. “We used to locate our labs in somewhat remote areas, where we felt we could control how much got out,” says Horn. “Now we locate them near intellectual centers, in order to stimulate the flow of ideas into our labs.”

Building on Common Principles

Big Blue’s transformation illustrates some principles that will become part of any successful new model for organizing and managing research:

 The importance of internal and external research sources for ideas, and internal and external paths to market for research discoveries. This requires companies to allow upstream technologies and components to be sold to downstream product and systems competitors. It also lets those downstream units buy on the outside.

 The management of intellectual property, both to ensure access to external ideas as well as to profit from one’s own IP.

 Changing the role of research staff from not just generating knowledge but facilitating and brokering it as well. Researchers and their managers must have direct experience with customer problems as they formulate future research agendas.

Keep Reading

Most Popular

Large language models can do jaw-dropping things. But nobody knows exactly why.

And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.

OpenAI teases an amazing new generative video model called Sora

The firm is sharing Sora with a small group of safety testers but the rest of us will have to wait to learn more.

Google’s Gemini is now in everything. Here’s how you can try it out.

Gmail, Docs, and more will now come with Gemini baked in. But Europeans will have to wait before they can download the app.

This baby with a head camera helped teach an AI how kids learn language

A neural network trained on the experiences of a single young child managed to learn one of the core components of language: how to match words to the objects they represent.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at customer-service@technologyreview.com with a list of newsletters you’d like to receive.