VDI cost models aren't what they're made out to be, so make sure you're getting the full picture before you dive into desktop virtualization.
Most vendors talk about how cheap it can be to deliver a virtual desktop. But some say you can do it for $150 per user, and others say under $500 is the target.
This disparity is usually caused by the vendor's frame of reference. Are they talking about licensing, hardware, endpoints and infrastructure? Or are they leaving out some of those things to make it sound better? Odds are those numbers are born out of an incomplete set of information -- aka "creative marketing."
In fact, trying to come up with a VDI cost model, even if you have good intentions, will likely never catch all the costs involved with a solution. With vendors' "creative" intentions, cost models can be made to say just about anything. Want to show that one type of storage is more expensive than another? No problem. Want to show that VDI is more expensive than traditional desktops? No problem. Cost models can do it all.
I'm sure you've seen the all-inclusive VDI cost models before. They ask questions like "What is the average cost of a help desk call?" or "How many users are in your organization?" and "How would you characterize the majority of your users' workloads? Light, medium or heavy?" Some go beyond that, with detailed numbers based on performance monitoring agents deployed around your organization. Still, at the end of the day, that data is being crunched by a generic algorithm that is using assumed information -- and that's never good.
Brian Madden wrote an article a few years ago that cited five ways you can lie with a cost model, and that list has grown to over 20. That includes things like moving trackable costs to non-tracked areas -- for instance, saying that power and air conditioning costs belong to a different cost center despite the fact that they still exist, or just outright ignoring certain things that don't suit your needs.
More on VDI costs
Controlling VDI costs via storage
How to calculate ROI from VDI
VDI project planning: A beginner's guide
My favorite way, though, is through something called soft costs. Soft costs are the ones that can't actually be tied to money, like user productivity or employee satisfaction.
The example I like to use is that of help desk calls. If the VDI cost model assumes that a help desk call costs $200 (remember, this is hard to quantify, and we need this overall cost to look good), and moving to VDI saves, let's say, 10 help desk calls per week, that means we're going to save $2,000 per week just by going to VDI. That means in the first year we'll save $100,000! That's a mighty big chunk off the cost of a project.
If a VDI cost model were 100% transparent and 100% configurable, you might stand a chance of getting it right. I've tried, the vendors try, and you've probably even tried to come up with something, but no matter what, somewhere along the line a generalization or assumption is made that can skew things wildly.
As we'll get to in an upcoming tip, though, most of these VDI costs are about delivering desktops that are just good enough. It is likely possible to deliver a virtual desktop for less money than a physical desktop, but comparing their costs is like comparing apples and oranges.