IBM at 100: A prosperous failure

IBM is celebrating its centenary this week. It has done well to survive for 100 years in an industry where companies die like mayflies.

But those of us who have followed the company for decades can't help but remember when Big Blue was a titan that aspired to dominate not just corporate data processing but the rest of the technology world from typewriters and telephones to communications and cash machines.

People who affect to be horrified by the power of small fry such as Microsoft and Google have no idea what it was like when IBM represented more than 70 percent of the industry — when it was twice as big as every other IT company put together. To have that kind of power today you'd need annual revenues of well over a trillion dollars. IBM has only managed a tenth of that — $100bn (£62bn) — though it is still roughly the size of Microsoft and Intel combined.

Tom Watson Sr, who gave IBM its name in 1924 and turned it into a commercial juggernaut, regularly sat down with American presidents, and negotiated with many other world leaders, including Adolf Hitler.

Foreign governments were so frightened of IBM's commercial and industrial might that they financed "national champions" to process their data, and to provide IBM with some local competition. These included ICL in the UK, Bull in France, Siemens in Germany, Fujitsu in Japan, and some minor players such as Norsk Data.

And Tomorrow the World?

But as we know, IBM has failed to live up to the title of Rex Malik's 1975 book, And Tomorrow the World?

In fact, over the past 20 years, IBM has either been beaten in or exited almost every market it used to dominate. In an unguarded moment, Microsoft's Bill Gates reportedly said he wanted to be "the IBM of software", and barring a few big rivals such as Oracle and SAP, he just about made it.

But Gates wasn't the only winner. Microsoft became the IBM of software, Intel became the IBM of microprocessors, Cisco the IBM of communications, Dell the IBM of personal computers, HP the IBM of printers, and so on, while Google is clearly the IBM of the web.

The IBM PC, launched in 1981, was a turning point for the company. Photo credit: IBM

What had been one giant IBM monopoly simply got subdivided into a lot of smaller monopolies, sometimes with IBM's unintentional help. For example, IBM made Intel and Microsoft dominant by choosing to use their products in the IBM PC, launched in 1981, instead of its own.

While IBM was losing these battles, it was disposing of whole divisions. For example, its Lexington-based printer division became the standalone Lexmark. IBM sold off its satellite business to MCI, its copier business to Eastman Kodak, its Federal Systems Division to Loral, its Global Network to AT&T, its phone business (half of Rolm) to Siemens, its disk-drive business to Hitachi, and its PC business (ThinkPads etc) to Lenovo.

That's not a complete list, but you get the idea. Today's IBM is a shadow of what it was. It's nothing like what it could have been.

Relative decline

We can clearly see IBM's relative decline in financial terms. IBM had annual revenues of $76bn in 1996, which was more than twice as much as Microsoft ($9bn), Oracle ($4bn), Dell ($5bn), Apple ($10bn) and Cisco ($4bn) combined.

Today, Microsoft ($62bn), Oracle ($27bn), Dell ($61bn), Apple ($65bn) and Cisco ($40bn) have increased their combined revenues eight-fold from $32bn to $255bn. IBM's annual revenues would now be $600bn if it had grown at the same rate.

Making the comparison with earlier years just makes IBM's failure look dramatically worse. In 1986, for example, when IBM was a $50bn giant with over 400,000 employees, Apple's annual turnover was $1.8bn, Microsoft's $0.16bn, Dell's $0.073bn, and Oracle's $0.055bn.

Of course, IBM's failure was not due to accident or incompetence: the world changed.

Of course, IBM's failure was not due to accident or incompetence: the world changed.

IBM's enormous power was based on owning and controlling the data-processing systems installed in large companies (and not in small companies, which couldn't afford them). These corporations paid IBM a monthly rent, which enabled IBM to keep equipment in use until it had covered its costs many times over.

That business model started to fail when large companies knew enough to run their own systems, and became brave enough to buy from rivals such as Amdahl (a mainframe company set up by IBM's leading mainframe designer). IBM cashed in by selling off its rental base to its customers, which gave revenues a short-term boost at a long-term cost.

Microprocessor trumps big iron

The second problem was the microprocessor, which eventually became powerful enough to do jobs that had previously required 'big iron', especially IBM's liquid-cooled thermal conduction modules. A small company called Sequent popped up and its founder told me how a couple of Intel 486 chips running Dynix, his multiprocessor version of Unix, could do the job of an IBM system costing 10 times the price.

Read this

IBM offers a glimpse into its history

Read more

IBM did what big companies do: it bought Sequent. Still, it had been making a 70+ percent margin on mainframes that cost up to $20m each, and for some tasks, these could be replaced with multi-processor systems where the manufacturer would make 40 percent (or less) on machines costing $10,000 to $20,000. Sales volumes would explode, but the gross profit on a machine might fall from $7m to $0.007m.

Frankly, I didn't think IBM would survive, and as IBM's share price approached its peak, I compared it to the Titanic sailing towards an iceberg.

In the early 1990s, IBM made the biggest losses in corporate history, and many people thought the old monopoly should be broken up into a series of Baby Blues. That didn't happen. In 1993 outsider Lou Gerstner was brought in to save the company. He did it by focusing on customers and services, while getting out of businesses he didn't consider important. This was a risky strategy because the margin on services was typically about half that on mainframes, but Gerstner made it work.

IBM is still a huge and very profitable company — it has 426,751 staff, and in 2010 it made a profit of $14.8 billion — but it no longer rules the IT industry. As one IBMer wittily told me: "We used to be the Evil Empire, but somebody else has that job now."