Changing circumstances make it difficult to measure adoption rate for Exchange 2013

A journalist asked me whether companies are deploying Exchange 2013 faster or slower than previous versions and what reasons are driving its adoption. I thought about the question a little and then replied that it is impossible to compare the adoption rate of one Exchange version against another because of the different circumstances that surrounded each release. Here’s what I meant.

When Exchange 2000 was launched in 1999, we had Windows 2000 and Active Directory to deal with. Windows was growing up from its PC LAN heritage on the start of the road to becoming an enterprise platform and we were preoccupied with the need to validate NT applications (for Y2K and Windows 2000) and to make sure that Active Directory worked smoothly before attempting to introduce the Active Directory Connector (ADC) and then start the process of moving from Exchange 5.5. Exchange 2000 wasn’t a particularly good release and its roll-out was slow. On the other hand, Exchange 5.5 was a great mail server that is still in use today, even if Microsoft would like to pretend that it went away some time ago.

Exchange 2003 appeared to repair the damage that Exchange 2000 had caused to the product’s reputation. This release was rather good. We had become used to Active Directory by this point and concerns around replication had largely proven unfounded. Some still believed that the domain was a security boundary, but that’s another story. Exchange 2003 came along with many new and exciting developments like the much better client synchronization with Outlook 2003 and the appearance of Exchange ActiveSync. Adoption from Exchange 2000 was fast, mostly because we couldn’t wait to get off that version.

Exchange 2007 came along in 2006. This was the first 64-bit release and, for the first time, customers were required to buy and deploy new server hardware as in-place upgrades weren’t supported. The need to procure new hardware slowed adoption (who needed all those extra bits anyway?) but once you got to Exchange 2007, you found that this was another good release. PowerShell made its first appearance and we all made jokes about the Unix-style syntax. We also experience log shipping for the first time, and even if Local Copy Replication (LCR) proved more of an interesting side street than useful in production, its CCR and SCR cousins proved that multiple copies of databases were the future and killed off the single-copy cluster (SCC) as effectively as a stake through a vampire’s heart.

The initial release of Exchange 2010 was a bit wobbly. Too many features were half-baked when Microsoft shipped Exchange 2010 and we had to wait until SP1 appeared before the true glory of the Database Availability Group and features like block mode replication and single page patching blossomed. Much for the better, Outlook Web App was rewritten (again) and renamed (access was so passé). Exchange 2010 was the first release to appear in the full glare of tweets, blogs, and other social commentary, so it took a while to lose its initial poor reputation. Now we look at Exchange 2010 and regard it as a very good mail server indeed.

And then we come to Exchange 2013. The big difference with this release is not the software and the technical innovation that exists in the software. Rather, it’s the prime-time billing that cloud email now has in the hearts of Microsoft and large sections of the customer base. The big decision that companies have to make around Exchange 2013 is whether to move to Office 365 or stay on-premises, and I think that the strategic and long-term impact of this choice is what makes it difficult and therefore creates a delay in adoption. Sure, there’s been some quality issues too, but these have mostly been small and irritating instead of large and fundamental and Exchange 2013 is now a solid email server, even if it has a tremendous appetite for CPU and memory.

So the journey over the last fifteen years has bumped into many challenges and changing circumstances that influence and affect the capability of customers to decide when they upgrade to new software. From Active Directory and Windows 2000 to Office 365 and Exchange 2013, there’s never been an identical yardstick to measure the adoption of one version against another. Even if you were to compare a single customer that has moved from Exchange 5.5 through each upgrade, I bet you’d find that internal events were different at each point, from capital scarcity to management change to regulatory oversight, all of which influenced how that company made their decision about when to upgrade.

Journalists love simple answers to questions. Migrations to a new version of Exchange are never easy. Is it any wonder why the adoption rate has flexed over time to a point when an easy answer is impossible?

Follow Tony @12Knocksinna