Savio has a point, much as I don't want to admit it. However, it might not be the point he's thinking that he's making. Or, rather, the data points to an entirely different point.
Wayne Waddoups of SAIC sent this slide deck along to me from a presentation delivered by Michael Cusumanoa (MIT) at Carnegie Mellon University, and I found it fascinating. The data clearly shows a (strong) decline in enterprise software sales over the last few years, with the only exceptions being "hits" and "platform leaders." In other words, those who get lucky and those who have built a massive lock-in ecosystem.
As shown, software is clearly on the decline, while services revenue is on a strong upswing. This, as Cusumanoa posits, may well lead the industry to invest in the next big area of innovation: Services innovation.
It's not just a matter of revenues, either, but also of exits. If enterprise (proprietary) software is healthy, you'd expect the public markets to welcome such companies with open arms. But they're not. The number of public software companies is in steep decline:
Why? Well, it could be that this stems from Oracle buying every available public software company. Actually, consolidation probably does account for some of this.
But I suspect the real connection has much to do with the model being fundamentally broken. Over time, we will see public open source companies (i.e., those companies that make the business of selling support/services around open source, without appreciable proprietary software revenues) grow. Right now we have just one company (RHT) that fits this bill. By next year, we should have two. In 2009, I suspect we'll have four to six, if they're not acquired first.
This model has legs. It's not a pure services play in the way that an Accenture is. It's not a software play in the model of IBM. It's both. You get service when on the face of it you think you're getting software. That's a customer-friendly model, and I believe it's the future.