Case study: IBM’s reputation for innovation

The story

In the 1990s, IBM was struggling. The rise of Silicon Valley as a hotbed of innovation had created a perception that everything innovative in the technology sector was created in small, high-growth companies. IBM, by contrast, was seen as oversized and decrepit.

The challenge: The company was finding it hard to attract talented computer scientists and salespeople, who were instead going to Silicon Valley. Reversing the perception that IBM was not an innovative company became a priority for senior management.

The solution

IBM decided to build BlueGene, a supercomputer. High-performance computers were used by meteorologists, security and intelligence analysts, particle physicists and geneticists. But in their efforts to push the boundaries of their respective fields, they were constrained by the speed of the world’s fastest computers. IBM believed that if it could meet this demand, it would be in a position to make a strong, symbolic statement about its prowess at innovation.

Impetus from the top: Lou Gerstner, chief executive, understood the power of symbols. In 1997, for example, the company generated a lot of attention by creating Deep Blue, a machine that defeated Garry Kasparov, the reigning world chess champion.

In 1999, when the company was more stable but still struggling to attract talent, the most exciting field of scientific inquiry was biotechnology. Scientists were close to completing a map of the whole human genome.

Rethink the model

Ambuj Goyal, head of computer science at IBM Research, led the project to create the first “petaflop” computer. Computer speeds are measured in “flops”, floating point operations per second. At the time, the world’s fastest computers operated at speeds of about 1,000bn flops. A petaflop was 1,000 times faster than a teraflop.

To create a new machine, however, IBM had to abandon the existing paradigm for how to design supercomputers. Traditionally, advances in computational power came through increases in processor speed and sophistication. But faster processors consumed more power. So, the company decided it needed to build a supercomputer while at the same time reducing the speed and complexity of the microprocessors. This had never been done before.

Tell the world

Whether the company’s theory could be turned into reality was uncertain but Mr Gerstner and his PR team seized on the opportunity to make their ambition public. IBM held a press conference in November 1999 to announce it would invest $100m over five years in a new machine, to be dubbed BlueGene.

There was a shift in press coverage of the industry: a New York Times article, for example, declared that only an organisation of IBM’s scale could assemble the expertise in computers, mathematics, biology, chemistry and physics to build such a machine.

By September 2004, the company had surpassed its original ambition of creating the world’s fastest computer: BlueGene was running with more than 8,000 chips and just over 36 teraflops.

BlueGene did not break the petaflop barrier until 2007, but IBM was still able to get attention in 2004 by publicising its ambition.

IBM expected the investment to signal a redoubled commitment to research that would in turn help recruit the world’s greatest minds – and it did. When asked about the project’s return on investment, Mr Goyal responded: “BlueGene has attracted a number of the world’s most talented graduate students to IBM. What’s the return on investment of that?”

Key lessons

First, IBM is defined by innovation and cutting edge technology. Attracting top talent requires understanding the purpose of your company and fulfilling it better than anyone else.

Second, IBM used disciplined experimentation, meaning it innovated in incremental steps and tested the progress at each stage. In so doing, it was able to learn quickly, make good decisions and get better results at the least expense possible.

The writers are professors at the Tuck School of Business at Dartmouth College, and co-authors of ‘The Other Side of Innovation’

- Financial Times, 29 Sep 2010