Betwixt and Between: Can IBM afford to stay out of the applications business?

For a decade or so, IBM has pursued a determinedly ISV-friendly course when it comes to the enterprise applications business.  During this time IBM has gone to great lengths to convince ISV’s to build their applications on top of IBM’s infrastructure software (primarily Websphere and DB2) all the while assuring those ISV’s that IBM had no plans to compete with them in the applications space.

This strategy has arguably served IBM very well.  It has helped build broad acceptance and support for their infrastructure products within many ISVs and has been well received by customers who want a vibrant and competitive applications market.  However, unfortunately for IBM, the software industry is evolving to the point where it may be impossible for IBM to stay out of the applications business if it wants to remain competitive over the long term.

A Tale of Two Forces

Two separate, yet simultaneous, forces are pushing IBM towards the applications business.  The first force is vertical consolidation set in motion by IBM’s competitors and the second is increasing price competition within the core infrastructure software business.

In terms of the software industry’s vertical consolidation, the most recent and high profile example of this consolidation has been Oracle’s acquisition of Peoplesoft (which I wrote about last week).  With this acquisition Oracle is trying to solidify and legitimize its vertical software suite, which starts at Oracle 10g, moves up to Oracle Application Server, and then ends with Oracle’s Application suite (and whatever remnants of Peoplesoft that are left over). For IBM, the Oracle deal is somewhat bad news as it had announced with great fanfare just four months ago that Peoplesoft was going to base all its products on IBM’s infrastructure software.  Instead, Oracle will now try to migrate all of those customers to an Oracle-only software stack which could hurt both IBM’s Websphere and DB2 businesses.

At the same time that Oracle is trying to move its center of gravity up the stack into applications, another major software company, SAP, is trying to move down from applications and into the infrastructure.  While it hasn’t got a lot of attention, SAP’s NetWeaver application server and suite of infrastructure software products is clearly designed to try and create a Websphere-like layer of branded infrastructure software that SAP can sell to its customers … in place of IBM’s software.   Can an SAP branded relational database be far behind?

Both SAP and Oracle appear to be taking a page out of the playbook of Microsoft.  By tightly tying its office applications to operating systems Microsoft has been able to create a tremendous 1-2 punch that compels customers to install Windows and .NET if they want to get the full impact of Exchange, Excel, and Word.  One can imagine similar efforts underway at SAP and Oracle where engineers are tightly integrating applications into their supposedly “open” application servers to the point where those applications only seem to work best on each company’s infrastructure stack.

This trend of consolidation reached its logical absurdity when Microsoft and SAP seriously considered merging in late 2003.  While most people laughed at the idea of Microsoft and SAP combining from a cultural and management perspective (including me), IBM software executives must have found the discussions to be quite sobering.  On paper at least, an SAP/Microsoft combination would be a vertically integrated behemoth that could really give IBM a run for its money across almost all platforms and markets (which is no doubt why Microsoft broached the topic).

The Price Is Wrong

Outside of consolidation, IBM’s ISV friendly strategy is also under attack from pricing pressures, particularly within the infrastructure software space.  This pricing pressure is coming from two fronts: the integrated vertical players and the open source players. As the integrated vertical players build out their software stacks they can start to selectively target competitors, such as IBM, with specific pricing strategies. For example, SAP can lower prices on NetWeaver because it knows it will make up much of the lost revenue by selling its ERP applications.  IBM can’t respond to such “bundling” tactics because it lacks an applications business.

On the open source front, while IBM masterfully embraced open source and Linux as a way to defeat Sun’s and Microsoft’s push into the glass house, they have created somewhat of an uncontrollable monster that threatens to erode margins in their application server and messaging products. With corporations embracing Linux as an operating system, many are now looking at open source applications servers (such as JBoss) and databases (such as MySQL) as potential ways to cut additional licensing fees. This pricing pressure is starting to be felt in the infrastructure space with per CPU pricing for application server licenses declining quickly even as IBM and BEA bundle more and more functionality into their products in an effort to justify their premiums relative to open source alternatives.

The Empire Strikes Back

It appears as though IBM’s response to these trends has been to redouble its efforts to make its infrastructure the center of the software universe by fragmenting applications into lots of little pieces.  By heavily pushing J2EE, grid computing and SOA’s, IBM is trying to fragment the application layer into lots of little pieces.  Fragmenting the application layer lets customers mix and match best-of-breed application components and therefore undermines much of the utility and rationale for monolithic application suites such as R3.  It also benefits IBM’s services business by making it easy for customers to custom develop application components and integrate them with 3^rd party software.

While in the long term such architectures may predominate, in the short term customers are still looking for solutions and it’s still very difficult to piece together a robust enterprise application using next generation technologies and architectures.

All Over But the Buying?

Thus, with competitors acquiring IBM’s best ISV channels and with pricing eroding thanks to bundled pricing and open source alternatives, IBM may be forced to enter back into the enterprise applications business, if anything, as a purely defensive move against its chief rivals.    If IBM does enter the applications business it will likely have to buy its way in, although with many of the most attractive ISV’s already acquired, IBM would have to cobble together a solution by buying several companies such as Siebel and Lawson.

On the other hand, IBM may simply fight a rear guard action and be content to gradually let its software growth and margins decline as software isn’t really its core business now anyway. From this perspective, IBM will accept gradually lower margins and growth in its software business as simply the price to pay for keeping its services and hardware businesses in the black.  After all, IBM had its chance to get into the applications market by buying Peoplesoft, and instead of playing the white knight (which Peoplesoft clearly desperately wanted IBM to do), it just stood by and let Oracle acquire Peoplesoft without batting an eye.

It’s hard to say which strategy IBM will settle on, but it is clear that thanks to deals like Oracle/Peoplesoft and trends like open source, the software world is changing and the status quo won’t survive over the long term.