I recently read a post from MVP Andrew Kelly about his experiences with the SQL2005 Adoption rate. In his article, Andy relates that he asks audiences at his presentations how many of them are using SQL Server 2005. At the PASS 2006 Community Summit last November, Andy said he encountered a 70-80% adoption rate while just a few months earlier, at Microsoft TechEd, he'd only encountered a 30-40% adoption rate. He also expressed surprise that a developer he'd spoken to was hard at work on a new project based on SQL2000.
To add my voice to this discussion, I've been speaking at one or more local user group meeting per month across the country and I am uniformly surprised by the very low adoption rate of SQL Server 2005. It's very common to find IT shops that have SQL2005 installed, but I'm finding that it's much less common to find IT shops where SQL2005 is the primary database platform or even the database platform of intention.
This is clearly bad news. It's hard to tell why SQL2005 is not being implemented as widely or as quickly as expected. The SQL2005 release is potent and a big improvement, but for some reason, it hasn't caught like wildfire.
Now, Microsoft is making noise like it will release a beta of its next version of SQL Server within the calendar year. In a situation like this, I believe we're looking at two possible adoption outcomes. The first is that most users actually leapfrog past SQL2005 to SQL200x whenever that release goes RTM.
The second scenario that I see as more likely is that SQL Server will spin out into a multi-version cosmos very much like what we have in the Oracle world today. When I visit Oracle shops, I commonly find Oracle boxes running everything from Oracle 81 to Oracle 10g R2, and sometimes even Oracle 7. What a support nightmare!
I believe Microsoft doesn't have a lot of options in this space. Better marketing could help a tiny bit. But Microsoft is already better at marketing than almost any body else in this space. Slight changes in programs and incentives might also help, but again, Microsoft is strong in this space. The only major impact I can envision at this point is a major groundswell in ISV support. ISVs still don't seem to fully understand or support SQL Server 2005, and there are literally thousands of them with 10's of thousands of users. Swaying them towards adoption seems to be Microsoft's best bet for ensuring a strong and widespread surge in SQL2005 adoption.