Thirty years ago, IBM used a paradox mindset to navigate the rise of cloud computing with its existing mainframe and hardware business. What’s the lesson for today’s IT leaders?
A recent piece in Academy of Management Today by Daniel Butcher takes a fresh look at IBM’s pivot to cloud computing, and it’s worth your time. The article walks through IBM’s early exploration of cloud technology in the mid-1990s when competitors like General Magic and Compaq began building business plans around the newly coined term. IBM assigned personnel with experience in enterprise IT services to specialize in cloud-based solutions. Those efforts culminated in 2007 with the official launch of IBM’s cloud computing division. That same year, IBM partnered with Google and six US universities to launch a server farm supporting research projects that needed fast processors to parse massive data sets.
The article centers on an interview with Academy of Management scholar Wendy Smith, who argues that innovation requires senior leaders to have uncomfortable conversations that question the foundation of their companies’ current business models. As she puts it, IBM had to innovate while managing “millions and millions of dollars invested in their existing relationships with their current clients and their current technology.” Smith’s central concept is the “paradox mindset.” The best leaders can hold the past, present, and future in mind at the same time. They can navigate the short term and the long term simultaneously. They can commit to both the existing business and the innovation all at once.
I find this framing refreshing because it’s a smart process. Going back and learning from the past, understanding what we did right and what we did wrong, yields some really good lessons. Too often, the enterprise tech world either canonizes or buries these pivots without actually mining them for insight. IBM’s cloud journey is one of the richest case studies we have, and Smith’s paradox framing captures what most coverage misses: This was never a simple strategic choice. It was a genuine dilemma with no clean answer.
The conundrum IBM faced
Let me be clear about something up front: IBM was early money in the cloud. I remember lots of IBM papers and talks about on-demand computing in its early days in the 1990s. Indeed, I wrote papers on it and covered it in my books and columns when I could, but it was science fiction to many people. The rise of the Internet suddenly gave us access to resources outside our firewalls, and I first saw that value take shape at early cloud computing events. IBM was right there, understanding early that this could have real potential. Dozens and dozens of companies were declaring their interest in the cloud computing space, and IBM was one of them.
But IBM had a trade-off to consider that most of the other players didn’t. Moving to an on-demand model flies in the face of selling hardware, certainly mainframe and legacy hardware, because promoting one path pulls from the other. If you’re pitching legacy hardware, you’re not really promoting on-demand computing; if you’re promoting on-demand computing, you’re not supporting legacy hardware. It was a conundrum. I remember writing 10 or 15 years ago that the worst job in the United States would be the CTO of IBM. I say that from experience as CTO of several technology companies, where I could take them in a single clear direction and never had to straddle two contradictory technological beliefs. IBM did.
This was unlike Google and unlike AWS, which didn’t have to cannibalize a hardware business to build a cloud business. That was certainly going to be the case with IBM, so it had to make some tough decisions.
Over the years, I watched IBM go in different directions with what I thought were scatterbrained marketing efforts that confused most people, including me. Smart Grid? Smarter Planet? A rotating cast of cloud brands and rebrands? Pick a direction, please. But here’s the thing: I don’t think they had a choice. They could either straddle both worlds and hope for the best, ready to pivot if needed or commit to one way or the other and risk making a bad bet that could kill the company. It’s hard to be a Monday morning quarterback when those kinds of stakes are in play.
What I would have done
Still, if I had been in that CTO’s seat, I would have created a unique and innovative strategy unlike anything else in the marketplace. After all, it’s IBM. They had the juice, they had the brainpower no other company had, and they had the ability to pull it off. I would have put all the eggs in that basket and moved forward as quickly as possible, making as many mistakes as I could as fast as I could, readjusting to different market conditions where we could be successful.
I’m not sure that bet would have been building another public cloud. I think I would have doubled down on the existing hardware business and being the absolute best in the world at it—not just mainframes, but small distributed systems, the kind of well-engineered on-premises platforms that companies like Dell provide these days. I would have gone all in and owned that market, even while absorbing endless criticism: “IBM is missing the cloud path!” “How will IBM survive without being in the cloud?” Yada, yada, yada. IBM would have survived just fine. Hardware assets, as we’re seeing with high flyers like Dell, HPE, and Nvidia, are becoming king again, and on-premises solutions are often the best, most cost-effective path. That case could have been made 15 years ago when public clouds started to rise. The criticism would have been loud at first, but it would have faded once the strategy took off.
Of course, these are all hypotheticals. Who knows whether IBM, had it gone all in on public cloud or all in on on-premises resources, would have done any better than it has today. Hindsight is 20/20. I do know they’re a valuable company with lots to offer, and I do know that IBM, like all the other legacy enterprise companies out there, will continue to face challenges as the market shifts and pressure increases.
Here’s my message to them, and to every legacy enterprise vendor reading this: Consider your legacy business and its real value before jumping to the next lily pad that may not turn out to be what you hoped. The cloud isn’t going anywhere, but neither is the mainframe, and sometimes the boldest move is defending the business you already have.
