IP: Industry R&D

[Dave Farber <farber () cis upenn edu>]

> From: irving () us ibm com
> Date: Sat, 20 Mar 1999 08:34:53 -0500


> The enclosed excellent article is a very good description of how R&D, and
> research in particular, has evolved in industrial labs like IBM's research
> labs.  For very good market reasons, the research has become more
> "applied", that is, more involved in bringing technologies from the labs to
> markets quickly, and more involved in helping customers solve real tough
> problems.  While that is very, very exciting, and undoubtedely must be the
> top priority of the companies, it raises the serious concern of the
> decreasing amount of long term, high risk research being conducted by
> industry labs, and more important, the decreasing funding of such research
> by the Federal government.
>
> (See attached file: Industry Week.doc)
>
> =================================================================
> Irving Wladawsky-Berger
> General Manager ,  Internet Division, IBM Corporation
> Somers, New York, 10589    914-766-1600 ; Fax: 914-766-8177
> Email: irving () us ibm com
> =================================================================


Industry Week: More firms discover innovation should be a business process as well as a research process.

John Teresko, March 15, 1999

In People and Performance (1977, Harper's College Press), Peter F. Drucker cited marketing and innovation as the two 
primary functions of a business. However, it took the competitive turnarounds and restructuring of high-tech firms in 
the 1990s to show how important it is to integrate them. Unable to increase R&D budgets, they discovered how 
competitive benefits can flow from tightly integrating research into business strategy.  
 
Lou Cole is one president and CEO who found integration to be a critical source of competitive advantage, growth, and 
continuous business regeneration. Actually, he had little choice. When he came to Legato Systems Inc. 10 years ago, the 
R&D budget of the start-up was almost nonexistent. To grow the revenues and drive profitability of the Palo Alto, 
Calif.-based company, he developed a culture that has avoided the insularity that typically develops between R&D and 
other corporate functions.  
 
Cole brought to Legato -- a specialist in enterprise-storage-management software -- a novel way to integrate R&D into 
the business strategy. Instead of compartmentalizing the function internally, Legato built and executed a strategy of 
R&D participation with Original Equipment Manufacturers (OEMs) -- 24 of the world's largest systems and applications 
vendors.  
 
"In the early days of Legato, we set the goal of creating the de facto standard for backup and recovery solutions," 
explains Cole. "But as a small start-up with limited funds, it would have been very difficult to fund and perform the 
R&D, make the product available on every platform, and then find a way to sell it into every one of those vendors' 
customer bases. It looked as if we were doomed to failure if we went about it on our own, working on the code and then 
trying to build a direct sales force. The economics simply could not support that conventional business model. The 
product sold for only $2,000 to $4,000 at that time."  
 
Cole's inspiration was to establish the OEM program as a way of having multiple development departments continually 
working on the technology to evolve product enhancements as the years progressed. "So, instead of having one R&D 
organization, we began with a strategy to develop multiples -- all working on the same fundamental code. We began by 
signing up OEMs, primarily system vendors. Digital Equipment Corp. was first, then Sun Microsystems Inc., Siemens 
Nixdorf, Data General Corp., and eventually the list expanded to 24 partners over the years. In addition, we also 
signed up application vendors such as Oracle Corp, Informix Corp., and Netscape."  
 
"The initial participants were OEMs that sold hardware, and the motivation to collaborate with us was to be able to 
present their customers with a complete solution," adds Ed Cooper, corporate director of strategic marketing. By 
starting with leading OEMs, the strategy snowballed the list of partners to the current roster of 24. The partners 
participate in a licensing agreement that provides them with source code to the product.  
 
The design presumption of the product, called NetWorker, is that it is built on a common code base. So regardless of 
the operating-system environment, functionality is identical. Legato's partners take the code, port it to their 
specific operating environments, and market it as their own product for use by their customers. To coordinate the R&D 
effort, Legato assigns technical people to each licensee, says Cooper. "We share both the vision for future product 
development and the work that's needed to accomplish it."  
 
 From a product point of view, because Legato has so many partners, it is alerted early when operating system changes 
come about.  
 
One of the requirements of the OEM agreement is that interoperability be maintained with Legato's shrink-wrapped 
product. For coordination purposes, Legato maintains a certification lab for interoperability testing, and every two 
years partners are required to submit their code so that all of the value-added R&D can be used to create a new source 
code. Legato then reissues the product back to the OEMs as a new baseline offering.  
 
The R&D partners also attend an annual OEM meeting with a focus on how they can take Legato's storage-management 
solution to the next level. Outside this one meeting, the companies compete against each other quite vigorously, notes 
Cole. He describes his R&D approach as a solution for start-ups that find themselves in the position of being able to 
leverage similar OEM relationships.  
 
"In addition to the R&D benefits, the strategy quickly made our software available worldwide with a special credibility 
advantage it derives from its blue-chip OEM partners," he says. "There was no way we could have done that with the 
venture-capital resources, because that would have taken a lot of advertising, a lot of promotion, and a huge sales 
force to make it all happen."  

In the beginning, the OEM R&D partnerships were responsible for a fairly large component of Legato's business -- 35% to 
40%, says Cooper. "Over the years it has become a smaller component because of the strategy's success." That somewhat 
paradoxical statement simply means that the winning market identity established by the OEMs is successfully driving 
Legato's shrink-wrapped version of the product, Cooper explains. "Although our OEM business continues to grow at a very 
nice clip, it is just not growing as fast." He says revenues from OEMs now account for 15% to 20% of the total.  
One measure of Legato's R&D strategy is its market penetration, which Cole estimates at 65%. Another is in financial 
performance. The company went public in 1995 with revenues of $29.8 million. In 1998 revenues reached $143 million with 
a net income of $28 million.  
 
At the other end of the economic spectrum is the research renaissance of IBM Corp. (1998 revenues of $81.7 billion and 
a 1998 R&D budget of almost $6 billion). In December 1992 IBM signaled its intent to rethink its R&D operations with 
the announcement that $1 billion would be cut from its 1993 R&D budget.  
 
That announcement coincided with an analysis of U.S. R&D practices presented by the Massachusetts Institute of 
Technology, Cambridge, and PA Consulting Group Inc., Plainsboro, N.J. Their survey -- "Global Benchmarking of Strategic 
Management of Technology" -- identified a significant gap in the R&D practices of U.S. companies. Their survey 
concluded that the companies were not taking advantage of outside technology influences such as customers, suppliers, 
joint ventures, and universities.  
 
Coincidence or not, since then IBM has evolved a keen understanding of the importance of creating R&D value through 
exploiting those influences. One measure of IBM's success is in IFI/Plenum Data Corp.'s annual tally of patents. For 
the sixth consecutive year the company was awarded the most U.S. patents in 1998, shattering the previous record by 
more than 40%. It received 2,658 U.S. patents in 1998 from the U.S. Patent & Trademark Office, 934 more than it did in 
l997. In second place was Canon with 1,925. The company's intellectual-property portfolio generates more than $1 
billion annually.  
 
Like Legato, IBM is striving to bring researchers closer to cus-tomers and markets. For example, "One of the most 
interesting parts of my job is the interaction with customers," says Paul Horn, senior vice president and director of 
IBM Research. Horn and his researchers interact with customers at IBM Industry Solutions Laboratories (ISLs) in 
Hawthorne, N.Y.; Stuttgart, Germany; and Yamato, Japan.  
 
Designed to help customers harness new technologies, the three labs have had more than 1,000 customer visits in the 
first two years of operation -- nearly 650 at the Hawthorne location alone, says Horn. The day-long visits devote 
mornings to interactive sessions, with afternoons dedicated to workshops on steps that the visiting company might take 
to solve specific problems.  
 
"Customer interactions help us set a significant part of our research agenda," says Horn. "Their visits help give us an 
outward, market-oriented focus, which supplements our existing focus on technology. Many of the component technologies 
that drive the information-technology industry will get faster, cheaper, and more dense by a factor of 50 to 100 over 
the next 10 years. Keeping up with that evolution is a daunting challenge, even for our most technically sophisticated 
customers. They understand that these technologies can be a competitive weapon for them, and they are looking for help 
in dealing with the pace of change that is going on around them. More than ever, they are looking for complete 
solutions to their problems rather than just a piece of hardware or software."  
 
The ISLs have given IBM researchers contact with customers across a broad spectrum of industries, including 
manufacturing, health care, education, telecommunications, and media, as well as banking, finance, and securities. 
About 50% have cross-industry application. Almost all are examples of what Horn cites as the emerging theory of 
competition: "All companies are becoming information companies; all will have to compete as information companies."  
 
For example, Britain's Safeway Stores PLC and the Hawthorne ISL are collaborating on a consumer application that gives 
top customers the ability to conveniently create and maintain personalized grocery shopping lists and pre-order 
groceries using a portable handheld device. Safeway says the technology will be available to its Basingstoke superstore 
customers via palm-sized electronic organizers called Easi-Orders. These portable devices will allow customers to 
preselect their weekly groceries from individually tailored, computerized shopping lists compiled from their own past 
purchases. They also will double as fully functioning personal organizers. Safeway also is researching how to use the 
technology to enable customers to digitally order groceries via devices such as TVs and mobile phones.  
 
Another example from the Hawthorne ISL is a recently completed project that automated the crew-pairing process for 
Southwest Airlines Co.'s fleet of 271 Boeing 737 aircraft. Specialists in mathematics from IBM Research worked with IBM 
Global Services and IBM Travel & Transportation. The challenge: pairing Southwest's 2,700 pilots and 4,500 flight 
attendants with more than 2,400 daily departures, in a point-to-point route structure, and in accordance with FAA 
regulations and union rules governing crews.  
 
"It previously would take a team of six or eight people several weeks to complete the entire crew-pairing process with 
every schedule change, which was happening with increasing frequency, given market dynamics and new aircraft 
deliveries," says Al Davis, Southwest's vice president of special projects, Dallas. "Now, the daily, weekend, and 
transition pairing solutions are generated in a fraction of that time."  
 
But from a national perspective, a presidential advisory committee questions whether that is enough to seed the growth 
of technology-based businesses over the next 10 to 15 years. The 26-member commission, about evenly divided between 
university researchers and information technology companies, has recommended that federal financing of IT research be 
doubled over the next five years.  
 
In making its case, the commission points to the economic and social benefits that have come from federal funding of 
research into computers and telecommunications. It says that those sectors are responsible for a substantial portion of 
the economic growth that has occurred since 1992. Those sectors were beneficiaries of funding provided prior to 1980, a 
time when more than 50% of all R&D was funded by the federal government, says Jules J. Duga, senior researcher, 
Battelle Memorial Institute, Columbus. In the last 18 years, however, the government's share 
of R&D has dropped to about 30%.  
 
Unfortunately, he sees little evidence that the trend will be reversed in the near future. He says the federal 
government will spend $68.1 billion for R&D in 1999, a slight increase over 1998 spending, barely keeping pace with 
inflation. The commission recommends that the funding increase be directed at four areas: software, updating the 
Internet infrastructure, high-end computing, and researching IT's effects on the workforce and society.  
 
Other scientists at IBM may have more conventional R&D missions, but special attention is given to breaking down the 
insularity of R&D and increasing speed-to-market. One example is the work of Robert L. Wisnieff, manager of the 
advanced-display-technology laboratory at the Thomas J. Watson Research Center. Working to commercialize a flat-panel 
display with four times the resolution of the best commercial computer screens, his project was able to go from 
prototype to production in 12 weeks. The manufacturing challenge lies in the way the increased resolution is achieved. 
Conventional liquid crystal displays (LCDs) have a resolution of up to 120 pixels per inch (ppi), but the active-matrix 
LCD (code-named Roentgen) that Wisnieff was readying has a resolution of 200 ppi, which greatly challenges the yield in 
the manufacturing step.  
 
Because marketing and manufacturability are such critical issues, researcher Wisnieff is part of two IBM organizations. 
He reports to Eric Harris, director of the Subsystems Technologies & Applications Laboratory at IBM Research, and to 
Takahisa Hashimoto, general manager of the Display Business Unit of IBM Japan Ltd. at Yamato.  
 
To produce the flat-panel displays, IBM depends on another teaming effort. Because manufacturing plants for flat-panel 
displays can cost $500 million or more, IBM partnered with Toshiba Corp. in a joint venture, Display Technologies Inc. 
(DTI), with facilities in Yasu and Himeji, Japan. To be able to go from prototype to production samples in just 12 
weeks meant that Wisnieff, in effect, integrated DTI's manufacturing facilities into his research efforts.  
 
"We had to make sure that what we were doing in research would fit into the manufacturing process." Wisnieff expects 
hospital radiologists and CAD/CAM design teams to be among the first users of the high-resolution display. According to 
one estimate, the display may appear in high-end ThinkPad notebook computers in late 1999 or early 2000. He believes 
that users of the display will be excited by the sense of virtual presence the technology provides.  
 
The examples of Legato and IBM are not isolated exceptions to yesterday's practice of R&D. Benchmarking studies show 
that most technology-based companies are about twice as effective at developing products as they were 10 years ago, 
says Pittiglio Rabin Todd & McGrath, Weston, Mass.-based management consultants to technology-based business. Its data 
show a 75% improvement in R&D effectiveness during the 1992-96 period alone.