There's one more tough decision for enterprise software vendors that I want to cover. Then I can respond to some of the flak that I've taken for Pt 1, when I focused on license revenue vs maintenance/services. Pt 2 examined stack or platform? The third choice is around breadth of functionality.
Integrated suite or best-of-breed? SAP, Oracle and all the rest of the enterprise software vendors are busy service-enabling their application stacks so that customers can easily disassemble and reassemble them into functional components that are more readily adaptable to the changing needs of modern businesses.
Yet they seem oblivious to the self-evident contradiction in preparing for this brave new world of composite applications while simultaneously continuing to insist on a single vendor's fitness to fulfil every conceivable application need that a customer might possibly have. Isn't it the case that customers are pushing them in this very direction precisely so that they can have more choice when it comes to purchasing those individual functional components?
Perhaps it is, but the vendors' first priority has to be to enable composite integrations — enterprise mashups, as they are starting to be called — within their own stacks before they can justify moving on to other mashup frontiers. Their customers demand it just as much as their own self-interests and indeed their stockholders do. Yet while they focus on re-engineering their core stack into composite services, they are neglecting those functional areas on the periphery that are most urgently in need of composite integration.
This leaves an exposed flank open to attack by companies such as Workday, the new venture led by PeopleSoft's founder
Craig Conway oops! Dave Duffield. Workday offers a composite application that deliberately targets a set of processes cutting across multiple application stacks and information sources within (and often outside of) the enterprise. By focusing on those processes that are inherently most composite, Workday maximizes the value it is able to add while minimizing the likelihood that the incumbent vendor will have a competitive offering with which to respond.
This is not something that Conway has invented. Companies like Taleo and Employease are already doing very well out of their ability to bring together functionality and information from many different external services and link it into the enterprise application infrastructure.
Scroll out a couple of years and you can envisage a market that's much more fragmented, with a variety of different vendors attacking these highly composite application opportunities that live between the cracks in the established vendors' suites. But a proliferation of vendors isn't going to make customers very happy. Despite their wish for competition among the vendors they deal with, customers still shudder at the prospect of facing a forest of throats to choke. They want just a handful of trusted vendors to interact with.
So whereas I've previously argued that SAP is following the wrong route when it comes to licensing, I think it's Oracle that is out of step with the market in its failure to open out to integration with third party vendors. SAP's work to build a NetWeaver ecosystem puts it on exactly the right track to combine multi-vendor, best-of-breed functional choice with the single throat-to-choke environment of a single-vendor platform.
Whether Netweaver is the right platform is a moot point, of course, as I mentioned in Pt 2 of this series. Another unknown is the extent to which functionality will be hosted rather than licensed, in which case the likes of Salesforce.com's AppExchange may grow into a potential alternative to on-premises solutions like NetWeaver. A further possibility is that some of the composite vendors will grow a sufficiently large footprint to offer a platform option of their own.