How Microsoft quantifies corporate IT culture to forecast Vista adoption

InformationWeek has an interesting story on some recent internal research at Microsoft that attempts to project Vista adoption on corporate desktops.

Despite the delays, missteps and uncertainty that have cast a shadow on Microsoft's long march towards Vista, it would be wrong to assume the company doesn't have a detailed plan for taking the product to the all-important corporate desktop. In typical Microsoft fashion, they have an extremely precise quantitative forecast of who will adopt Vista, when, and why. Their numbers might be wrong, but it's instructive to follow their reasoning.

In a nutshell, based on its surveys of Vista beta testers and its experience with previous big Windows releases (e.g. XP), Microsoft expects that somewhere in the low double digits of companies will convert to Vista desktops in the first year following release, probably 12% to 16%. The takeaway from this is not the number itself (whether you think it is low or high), but the implication that Windows client revenue in FY07 is not as vulnerable to unexpected delays in Vista uptake as one might imagine.

By an interesting analogy, because Vista adoption will be driven partly by the degree to which corporate cultures are ready for high levels of IT automation, the analysis also has implications for the adoption of SOA.

Following Gartner, Microsoft classifies companies into three bins depending on their approach to managing PCs. About 70% are still managing their PCs manually, i.e. they have tech support people running all over the company popping CDs into users' machines when it's time to upgrade. 28% of PC shops are semi-automated, while only 2% are fully automated or what Microsoft calls "rationalized" (i.e. install and update all PC software automatically over the network, and lockdown PCs to prevent users from tampering with their configurations). Guess which group Redmond is aiming Vista at in the first wave? If you guessed the 28 percenters, you guessed right.

Microsoft tells InformationWeek that back in 2001 only a "single digit" percentage of the corporate desktop installed base migrated to XP in the first year, but it expects to double that for Vista. Assuming "single digit" means 6% to 8% at best (otherwise they would have said "nearly 10%"), that means they're counting on 12% to 16% Vista migration at best for year one, or roughly half of the semi-automated segment. Presumably they're only referring to the corporate market in North America and western Europe, because the cost of the high-end PC hardware needed to run Vista will delay adoption elsewhere. But that still works out to 12 to 16 million Vista licenses in North America alone.

What's interesting is the precision with which Microsoft has quantified the benefits it expects the semi-automated PC shops to reap from Vista. (This is the most market-research driven software vendor in the industry, bar none – some of the other big enterprise software players could afford to take a few lessons from Redmond in this regard, it might make their revenue forecasts less off the wall.) Microsoft estimates that the typical semi-automated shop running XP spends $542 per year on IT labor to manage each PC. Switching to Vista will, according to Microsoft, save them $342 per year per PC on average. So the payback period is going to be pretty short, especially if you're getting Vista under Software Assurance (it will cost you roughly $150 per license if you don't).

InformationWeek goes on to point out that not everyone agrees with this rosy estimate of Vista ROI. Gartner thinks the savings per PC will be closer to $200, and points out that the decision to migrate is driven not just by a straight ROI calculation but by corporate culture. Not everyone understands where their best interests lie, and not all those who do are able to act on that understanding. Some companies are just plain messed up. But it seems a pretty good bet that the shops who have semi-automated PC management already are likely to see the benefit of moving to more automation with Vista.

There is an interesting corollary to this analysis which applies to enterprise software at the server level, that is to say to the domain of Services Oriented Architecture (SOA). Lots of CIOs and consultants view SOA as overhyped, and it's true the recent promising license revenue growth numbers from SOA players are from a small base. For example, BEA reported about $27 million in license for its AquaLogic line in the July quarter, and probably only about half of that is pure SOA. But the analogy between Vista and SOA adoption is that both require a "rationalizing" approach to IT. Despite the widespread sentiment that enterprise IT is mature or even "non-strategic", there are in reality a huge number of big and midsize companies who still use their IT assets in shockingly inefficient ways.

Squeezing out these inefficiencies is going to put dollars on the table for the management teams that drive their companies to do it. That will hurt some IT vendors (think server hardware vendors squeezed by virtualization). But it is bound to benefit many software infrastructure and application vendors (think BEA, Oracle and SAP), providing of course that these vendors can run their own operations efficiently enough to capture some of those user cost savings.