GE Taps a World of Local Skills

In the flat agricultural landscape just north of Munich, a gleaming structure of steel and glass is nearing completion. This summer it will become the main European research centre for General Electric, the huge US industrial company that is the world's most valuable business by market capitalisation. The $52m (£28m) building represents the final component of an effort by GE to improve its ability to tap the skills of scientists and technologists around the world.

Until four years ago, almost all GE's spending on research - as opposed to product and process development - took place in the US. But as part of the attempt by Jeffrey Immelt, chairman and chief executive, to make GE less focused on the US, as well as increase its presence in new technology, GE has opened up two new research operations, in Bangalore and Shanghai. It has also pushed up its research and development expenditure to about $2.6bn last year, from $2.2bn in 2000. The Munich centre will complete the company's aim for the first time to have a global research presence.

In taking this path, GE is following a decade-long trend among big companies, such as Siemens of Germany, Philips of the Netherlands and International Business Machines of the US - though arguably it lags behind them. IBM, widely considered the leader in the move to globalise research, has eight laboratories, just three of which are in the US.

By spreading out their research base, these companies hope to access a broader range of technological disciplines than would be available in their home country.

Spearheading the effort at GE is Scott Donnelly, a long-time GE industrial engineer who became the company's vice-president in charge of R&D in 2000 after periods in aerospace, semiconductors and medical equipment.

GE has spent about $280m building the new laboratories and modernising its original research unit in New York. Mr Donnelly's task is to ensure that the extra outlays on R&D pay off in the form of increasing GE's presence in its main industrial sectors, which cover products from power stations to X-ray scanners. Last year GE had sales of $134bn, with just over half coming from industrial and consumer products and the rest from financial services including insurance, plus the NBC television station.

Central to Mr Donnelly's philosophy is his belief that - thanks to the breadth of GE activities - a number of critical research technologies can be "leveraged" across GE's 11 main business divisions. Also crucial are the links between the 2,300 researchers directly under Mr Donnelly's control and the much larger number of engineers and other business people developing new ideas in GE's sales divisions.

To get the most value out of GE's spending on its research centres of some $400m a year - equivalent to roughly a fifth of the sum the company spends on development - the four units are, says Mr Donnelly, organised in a "globally transparent" fashion. Different people work on the same technologies in different parts of the world, connected by e-mail and in frequent contact. The centres focus on 10 to 12 core or "enabling" technologies, covering areas such as metallurgy, organic chemistry, solid state physics, bio-sciences, electronics and imaging technology.

GE, says Mr Donnelly, has a huge advantage in that it can use the same idea in one of these topics in many ways. "In ceramics, for instance . . . we can direct our knowledge to divisions of the company that might seem unrelated.

"Ceramics technology has applications in the detectors that you need in medical scanners. It is also important in the materials used to coat turbine blades to make them withstand high temperatures. The blades are used both in our power systems division and also in [the GE division making] aircraft engines. New materials derived from ceramic technology are also important in lighting." This also saves money, says Mr Donnelly.

"If our engines division had to research an area such as ceramics for itself, it would have to do similar work to what we're already doing with applications for another part of the company, such as power systems."

A crucial part of the research management challenge is to connect the people in research with those in development. Mr Donnelly explains that to make the connection work smoothly "a mixture of push and pull [in technology]" is required. "Our approach is that we need the technology to end up in products. If this happens then we've been successful."

Roughly half the research centres' budget comes from specific technology programmes requested by the business divisions. Several dozen of these projects are taking place at any one time, lasting between six months and five years.

The power systems division, for instance, might tell the corporate research unit that it needs a new kind of gas turbine with a high thermal efficiency, and turbine blades that can withstand particularly high temperatures. In such an instance, Mr Donnelly says, his staff would assemble a worldwide team of GE researchers drawn from different technology disciplines.

"[We would look at] such areas as combustion, firing systems, the design of new super-alloys for the blades, ceramics, structures, and modelling the whole design so it can cope with high temperatures and is capable of being controlled. The team would discuss their approach with the people from the business division. We would set up a series of milestones in the project to ensure that we can monitor how well progress is being made."

The other half of the annual $400m spending of the research centres is directed, says Mr Donnelly, towards longer-term projects that are not tied to specific goals but relate to what the business divisions are interested in. Here the programmes are worked out through discussion between Mr Donnelly and his research staff and people representing the divisions.

"We have to decide on areas of activity that we think are relevant to the longer-term future for GE," says Mr Donnelly. "For instance, we are working on areas such as emissions from power systems and jet engines, and new materials which have a linkage with our plastics division. Or we could be looking at new sensors or control equipment that apply to a number of business divisions. We don't do real blue-sky research, which is looking too far ahead."

Personal contact is all-important. "For each business division, I have a person on my [research] staff who knows the business and manages the interface between the research effort and that division . . . He or she is basically a technology transfer specialist ...a networker, who knows not just one person but a lot of people in each of the divisions," says Mr Donnelly.

It will no doubt be some time before GE works out whether it is moving in the right direction. The sheer complexity of putting into place the proper links between the different groups of researchers - and people in the business divisions - is taking up a lot of Mr Donnelly's time.

How will he know if he is doing his job properly? Mr Donnelly says there is no easy answer. "Our basic method [in tackling this issue] is to look at the market share of the business divisions and their profitability. Our job [in research] is to play a part in providing technology. The job of the business divisions is to execute product development programmes and sell the products in the market place. But it's difficult to be too prescriptive about the part that each of these groups plays. It's a team effort. Technology is important but it's just one part of the total picture."

A full transcript of the interview with Scott Donnelly is available online at www.ft.com/donnelly