Gerstner refocused IBM on software and services. The product line was unified under the Linux operating system. Thousands of people were let go, and thousands more arrived through acquisitions such as Lotus Development, Sequent, Informix, Tivoli Systems and PricewaterhouseCoopers' consulting business.
Gerstner was able to hand off power to an operations man, Sam Palmisano, in 2002. Palmisano in turn was able to hand off last year to Virginia Rometty, then the head of marketing.
When Rometty was named, I tried to write a cautionary piece for another Web site, but it was rejected by the pay side of that site and eventually came online as a mere blog post. No one wanted to listen.
But the world has changed. As was the case 20 years ago, value is moving "up the stack," toward the companies that imagine and run applications, away from the companies that write them and manage them.
So IBM needs a new vision. The company understands where the computing world is heading, but it's losing more on its older lines than it can make on its newer ones. The loss of the CIA's cloud contract to Amazon.com (AMZN) should have sent alarm bells ringing throughout the executive offices.
Instead of hearing those alarm bells as an attack on its vision, IBM fought the award and lost.
The problem for IBM is that everywhere it looks in cloud it finds commodities. Cloud services are a commodity. Cloud management, cloud software and cloud operations are all becoming commodity, with dirt cheap prices that are only going to get cheaper. You can't maintain IBM's human infrastructure on commodities. You need enormous value-add.
My own view is that IBM needs to start running the services created by cloud. But I'm just an old reporter.
I would be buying transaction processing companies and business-focused cloud services. I'd be building a system for writing, deploying and managing apps quickly, and making app developers my new OEMs. I'd be selling anything that didn't fit this vision, even spinning out the mainframe division.
I'd be finding ways to productize cloud, creating a "cloud in a box" that any organization could buy, holding the crown jewels while maintaining compatibility with public cloud infrastructure. I'd be advertising a bridge from client-server to the new, cheaper infrastructure, one in which computing is managed, like electricity, rather than sold as a product.
IBM needs to rush "up the stack," and that will take a visionary, not a marketer.
At the time of publication, the author owned no shares in companies mentioned in this story.
This article is commentary by an independent contributor, separate from TheStreet's regular news coverage.