What the $2.1 billion bid for Rational tells us, following the $3.5 billion bid for PWC, is that IBM is trying to do nothing more than take over the world of enterprise applications. With a starting position that we all thought was dead and buried: custom-built applications are better than pre-built, packaged applications.
The timing of the IBM announcement is ironic, coming on the heels of a story first reported in ComputerWorld the same day: Computer Sciences Corp. recently botched a $100 million custom implementation for 21st Century Insurance. CSC was acting out the IBM model -- build it, don't buy it -- on a company that has a solid record of buying, not building. The failure was so significant that 21st Century felt obliged to mention it in its most recent 10-Q.
Like most failures of this nature, we'll probably never know what caused the deal to go south. There are a lot of reasons why it could have happened, and all of them and more could have occurred if the implementation had been using packaged software, not a tools and consultants approach.
But don't let mere coincidence ruin the irony of the moment. While any project can turn into a disaster, the IBM/CSC model shouldn't wheedle its way back into the received wisdom of the market without a fight. There were some very good reasons why buy beat build in the early 1990s, and nothing IBM has done so far is really going to change that.
It's hard to know where to begin. Scalability? Upgrade costs? Finding and maintaining the skills needed to run custom software? The army of consultants, the endless overruns, the faint whiff of "make work" in the latest proposal for functional enhancements? And on and on.
Okay, so all these things are still true in a packaged software implementation as well. But implementation cost in packaged software has evolved significantly in the last 10 years. The average ratio of software to consulting dollars used to be anywhere from 1:5 to 1:10, today it's often closer to 1:2 or even 1:1.
Even if we wanted to, we couldn't measure the improvement in pure custom consulting over the last decade: there's nothing to compare it to. And that's the real benefit of the packaged approach: as a user (and implementer) there's lot to compare with and benchmark against. Being part of a community of packaged applications users means you can head for the Web, the chat rooms, the listservs, and the user conferences and swap tips, workarounds, and improvements. You can improve your staff by going to the open market for more talent, and you can use your fellow users as leverage to get things done your way, not the vendors way.
Where do you go for dialogue, tips and workarounds in your custom environment? Guess who? He's the guy with the bus parked outside and the meter running. And if the knowledge transfer process hasn't worked as well as it's supposed, he's your job market too.
There is a genuine need for smart consulting -- both at the technical as well as at the business consulting level. But the IBM vision is that the existing enterprise applications in the corporate ecosystem can only evolve to new functionality and productivity through the application of consulting services.
My real problem with this vision is that it does more to stifle innovation than a software monopoly. The IBM model makes innovation the sole responsibility of the consultant, cutting off the vibrant software culture that is driving packaged software offerings, large and small. If consultants could really write enterprise applications, they'd have done so a long time ago. IBM is trying to remake the entire market in its image, and it's not a pretty sight. The service model can only go so far before its growth becomes limited, and IBM has now hit that wall. They need your packaged software dollars to justify an over $6 billion bet on the wrong strategy. Think twice before you let them have it.