Siebel 7 is pretty new. Is something wrong with Siebel 7 that requires a rewrite so soon?
Architecturally Siebel 7 is sound. In fact, the announcement is really Siebel’s product release schedule giving customers an understanding of what infrastructure changes to expect over the next 24 months. CIOs want vendors to be more candid about major upcoming product line changes, seeking sufficient lead time to prepare. The Siebel, Microsoft, and IBM developments won’t happen overnight, but will be delivered in stages over the next 30 months, and most users will appreciate the advance notice. By the time these changes are generally available, the current release of Siebel 7 will be at least four years old and ready for an upgrade in functionality and supported environments.
How can any supplier afford to implement a product twice (once on .net and once on J2EE)? How will Siebel keep the two versions from diverging? Which one is destined to become the least favoured?
The questions harken back to the days of multiple databases: IBM DB2, Microsoft SQL Server, Sybase, and Oracle. In many cases, software vendors looked to the database vendors to give them guarantees that their code would operate with the latest release of a given database, while ensuring that the application vendor would be the first line of support for any problems a customer might experience with that database. Today, it is quite common for software vendors to support multiple databases. It is also quite possible for software vendors, if their code is designed properly, to support native application servers. It just takes the right code structure, suppliers, and relationships to pull it off.
The same goes for this deal, especially with the code structure of the Siebel product line. The meta-driven architecture allows Siebel to design functionality in a way that is abstracted from the underlying application server. Siebel is working closely with Microsoft and IBM to rewrite its object manager, the executable manager code, to operate on the native Microsoft .NET or IBM WebSphere J2EE platform.
According to Siebel, the object manager accounts for roughly 5% of engineering effort of its total product line. Think of it this way: Siebel is outsourcing development of non-strategic infrastructure functionality to leading suppliers of infrastructure. By doing so, Siebel can concentrate on more valuable functionality that expands its capabilities in customer relationship management (CRM). As AMR Research’s Jim Shepherd pointed out at Strategy 21, many other industries have created a supply chain of components that are embedded into a final product; it’s about time for the software industry to do the same. Siebel is clearly taking the lead on this subject, and should be commended for it.
What assurance do we have that they will get it right: meeting timeframes and functional expectations for deliverables?
Here is the big question. Plans are great, but execution is everything. Software vendors have significant room for improvement here. Customers of enterprise software providers often discount what vendors say when making statements about deliverable timeframes and product specifications. This is no different. Siebel 7 was delivered six to eight months late from expectations that were set early in their plans--so don’t bet too much on it hitting promised dates.
Aren't customers looking at a large upgrade effort moving from one code line to another? Does anyone remember the Siebel 7 “One Button Upgrade” propaganda?
Yes, Siebel 7 is a one-button upgrade, “…but you have to push it 1,000 times", customers have commented. Furthermore, Siebel has suggested that Siebel 7 would be the easiest enterprise application to upgrade, boasting that in the first year alone, 90% of customers would upgrade to it from Siebel 6. It hasn’t happened and no enterprise software upgrade can be done as a one-button upgrade, so software vendors should set expectations properly. The first year of availability only converted 15% to 20% of Siebel customers to Siebel 7, a reasonable percentage and one that’s considered a relatively successful migration in the enterprise software market. The good news is that Siebel hasn’t made the same sort of overblown claims, having learned from its past missteps.