IBM’s strategy in a nutshell: All roads must lead to profit

I had a chance to sit down with IBM Senior Vice President Steve Mills last week (technically, he’s SVP and Group Executive for Software and Systems) and ask him a lot of questions about how the company is riding the waves of big data, cloud computing and open source software. I thought he was pretty frank, even while occasionally laying on the IBM spin.

Here’s what he had to say:

On capitalizing on “big data”

“The tech industry is very fashion-conscious and you have find a way to ride the wave,” Mills said, referring to the fact that analytics have always been a primary use of computers. Computers have pretty much been used to automate business processes and act as calculators, Mills added, and now, thanks to advances in hardware and data storage, “you can do a hell of a lot of arithmetic really fast.”

On how to make money from open source

“The idea of commodification is Darwinian … it’s always occurring,” Mills said, so he expects open source software to eventually affect pricing in every part of the company’s software business. The trick is in finding ways to charge money for stuff on top of the open source layer, and IBM has seen aggregate software margins rise over the last 15 years even after embracing technologies such as Linux and Hadoop and “really put[ting] Apache on the map as a entity,” he added.

Actually, open source can be something of a boon if you know how to do that last part. “I don’t need Cloudera because I’ve been delivering Hadoop for 4 years,” Mills said “… It’s there for the taking.”

Steve Mills

Steve Mills

On buying SoftLayer for $2 billion

“[W]e bought a company. … I’ve bought 120 companies,” Mills said matter of factly, noting that in this case IBM realized it wasn’t capturing certain segments of the cloud market and wasn’t delivering certain capabilities that customers wanted. You can either build those capabilities or buy them, he added, and “at the end of the day you run out of money [to build everything].”

Mills did defend IBM’s cloud legacy, though, going back to then-Google CEO Eric Schmidt calling IBM in 2006 and asking if it would help set up a web-accessible developer cloud “because we had dynamic provisioning and scheduling technology and they did not.”

On getting research right

“Sometimes you try and you can’t figure it out … and then you come back and try again,” Mills explained, noting that IBM never throws away any of the technologies it spends its $6.5 billion R&D budget working on. Some things become easier on the second or third tries because computers get faster, while other times the right use case comes along for something many years down the road.

In fact, he said, IBM’s famous Watson system is an amalgamation of about 40 different technologies, some of which (such as artificial intelligence and natural-language processing) first emerged in the 1960s and became one-off implementations (e.g., advanced gate-screening systems at airports) in the 1980s when computing got cheaper. Ultimately, everything advanced enough to where IBM thought it could create a system capable of winning at Jeopardy!, and the company keeps spotting new commercial use cases in areas “where human beings could not keep up with the information.”

On turning every decision into money

Really, though, my big takeaway from the discussion has to do with IBM’s near-laser focus on the bottom line — something that sometimes comes across as an afterthought when covering web startups or even large web companies such as Google and Facebook. From research to open source software to spending billions buying companies, it’s all about betting on the things that will deliver the most return on the investment for IBM and its shareholders. Walmart has great DNA around volume, Mills explained, while IBM’s DNA is focused on moving value toward the margin.

“The ‘How do you make money?’ challenge is not a small one,” he said. ” … We don’t have permission to not make money. We’re not Amazon.”