Why IBM's Software Position Makes Sense

I have opined here more than once that Microsoft (MSFT) is facing a major business-model decision between being either an information technology [IT] supplier or an IT services provider. IBM (IBM) already crossed that bridge. The acquisition of PwCC and divestiture of the PCs to Lenovo (and other such strategic shifts) are prominently illustrated as the in and out of the shift on page 1 in IBM’s recently released 10K.

Knowing which business-model camp an IT company has its feet in is important in IT investment research because the market dynamics are entirely different in the two approaches. Is a company selling commodities or the aggregate of its brainpower? Do investors realize the benefit of low-cost mass production overcoming low unit margins or the benefits of high “unit margins” on thousands of costly time-sink one-off engagements? Can the supplier combine the two by pricing the commodity by the drink (Software as a Service—SaaS)? Can it give the commodity away and make money on backup business and technical acumen services (open source software—OSS)? Can it do all of the above? Very few companies, maybe not even IBM, can do it all in my opinion.

That’s why you might find it strange that IBM continues to emphasize its software technology market position and revenue stream to its investors. I think there is some logic in the IBM message and position however, even a double dose of logic:

Software is the key to whatever leverage companies can exercise in a traditional IT services business model or a PwCC assignment: having to change a few lines of code for each of those thousands of costly time-sink high-margin engagements lowers each engagement’s cost and means potentially better profits than the net on millions of low- margin commodity technology sales of equal revenue, no matter how cost effectively the commodity is produced, because SG&A is lower.

Underlying infrastructure software such as that marketed by IBM is also an important factor in cost effectively providing applications SaaS to enterprises and consumers. Currently IBM only supports partners that go to market via that model, as opposed to doing SaaS itself. But it would not be a big step for IBM to build or buy its way into the SaaS space.

Use of OSS is a major tactic in executing both strategies, and IBM is a major OSS player. In fact, in my opinion, IBM and the rest of the IT Top 12 have just about co-opted the OSS movement.

So software is THE fallback position for IBM should it ever decide that the PwCC acquisition and like moves over the last few years was too much of a departure from its historical roots.
The position is not without risk. IBM has put all of its (very large sock full of) marbles on the services oriented architecture [SOA] concept. SOA, according to IBM, in a 3/13 announcement, is “a business strategy that enables a company to more closely align and reuse existing technology to achieve business goals.” To IBM, which no longer markets a full enterprise application portfolio, infrastructure software (which IBM calls middleware) is the key SOA deliverable.

The IBM tactic of marketing SOA to business executives is an interesting approach however because for many of us (and the arbiters of such things at Carnegie Mellon, which only defines ‘service’ and ‘architecture’ separately), SOA is a commodity. SOA is to C-level execs what rack and pinion steering was to car buyers in the 90s: gotta have it; no idea what it means.

Technically SOA refers to

layering a 10-year-old service-level-agreement mechanism (“the contract”)

on top of the decades-old concepts of set-theory-based object-oriented polymorphism, encapsulation and inheritance

in a 20-year-old client/server relationship

To some that study SOA, the slightly more modern Internet and service bus (“loosely coupled” objects) technologies are required to make it truly SOA. But if IBM wants to say SOA is a business strategy, it has the upcoming golf-season and March-Madness advertising power to make it so.

Another risk of course is that very little of IBM’s current software revenue stream is SOA-based yet. IBM says all of its software revenue grew 8% in 2006 over 2005, from $16 billion to $17.3billion. But unfortunately, revenue for its old-architecture products are deflating almost as fast as revenue for the new SOA stuff is rising.

A little more than half of the $17.3 billion was for so-called branded products such as WebSphere, Tivoli, Lotus, etc. This category grew 17% by IBM’s calculations but I think as much as half of the 2006 “branded” growth came from the Filenet, Micromuse, MRO, and other software acquisitions. (And billions of the branded revenue flow are for renting or licensing earlier versions of WebSphere Application Server, MQ/Series now called WebSphere MQ, etc., not SOA.)

At the same time, “unbranded” products (e.g., IBM operating systems; CICS, ironically one of the highest grossing IT brands ever marketed; etc.) deflated 3%, partially cancelling out the organic revenue growth for branded software

A billion dollars worth of product lifecycle management (PLM) applications from Dassault flow through IBM, about half of which goes back to Dassault.

Finally when last reported separately, most of this software ran on the IBM “proprietary” platforms, the z and “i” series. So as much as 60% to 70% of the IBM software business is “captive.” No one else can compete for it on a level playing field (although BEA can and does try to convince 30-year CICS mainframe users to switch to Tuxedo on UNIX, and Microsoft gets longtime AS/400 users to move to Windows).

It is also important to note that another $2 billion in IBM software revenue in 2006—both SOA and traditional—passed through other IBM divisions, primarily IBM Global Services. That flow is not as good a test of market acceptance as software that flows through outside third-party channels or direct to enterprises. It’s real money of course and there’s nothing wrong with blocking out a few competitors any way you can. However, if I am reading this March 12, 2007 IBM IR Note correctly, IBM will no longer report the “internal” revenue.

IBM has emphasized an important SOA issue in its latest announcement: quality management. It says its new portfolio includes Rational, Tivoli and WebSphere software as well as new technology services that will help improve the overall quality of technology in an SOA bundle before it is deployed. The techniques are in fact probably best practices that grew out of its IGS or Lab Software Services activities. With IBM's new SOA quality management approach, clients will be working with SOA the way IBM does, always a good thing for an IT supplier whether specializing in IT as a commodity or as a service.

Disclosure: Author has no position in IBM.