Virtualization Projections Deserve Scrutiny

This is a cautionary tale about the need to carefully examine analyst firm projections and not just highlights of that data, summarized by someone else. Most complaints about analyst projections focus on the wildly optimistic numbers they assign to putative market sizes—you know the “[We] (insert name of analyst firm) estimate the left-handed mouse market will be $17 billion by 2012.” If you took all the market projections by all the analyst firms and added them up, they’d probably sum to more than global GDP, leaving no room for all those unimportant sectors of the economy like manufacturing, agriculture, construction, and, of course, pet psychotherapy.

But this post isn’t going to complain about over-optimistic analyst projections. Instead, it’s about the need to examine them—because in this instance, the analyst firm did a good job of estimation about the financial benefits of virtualization, but their numbers were misquoted, leading someone who would take them at face value to conclude that client virtualization doesn’t really offer much financial payback—a conclusion at odds with the firm’s actual numbers, a disservice to its work, and misleading to anyone who was looking for help to decide whether to pursue a client virtualization strategy.

The story: I received a whitepaper on virtualization. I won’t disclose the source, as it was clear during my quest for information that this whitepaper was little more than a compendium, a pastiche of information lifted from different sources, put together by someone without any familiarity with the subject. So, discretion will be the watchword here. In any case, the originator of the paper is less relevant than the point of this post, which is the need to carefully evaluate the information you review when making decisions.

The whitepaper covered a number of different aspects of virtualization, all with a positive perspective on the benefits of the technology. One fact cited in the paper really stuck out at me: On the first page, in the “Key Findings” box, the paper noted “Firms could save GBP 78,000 (USD 159, 000) for every 1,000 PCs a year by replacing the full desktop PC kit with a server-hosted desktop virtualization set up.”

I looked at that and thought, “Hang on a second: $159K savings for 1,000 PCs? That’s only $159 per year per PC! That’s not much of an argument for client virtualization.” I sat down with a calculator and figured out that, at loaded personnel costs, that meant that client virtualization would only save around four hours of labor per client. In my opinion, that level of payback wouldn’t justify pursuing client virtualization. And of course, there are a number of other areas that should offer payback from client virtualization (I addressed them in earlier posts here and here ). So $159 per machine wouldn’t really justify pursuing client virtualization as an option.

On the other hand, many anecdotes about client virtualization paint a very positive picture about the benefits of the technology—certainly well beyond $159 per machine per year. So how to explain this inconsistent assessment?

Intrigued, I decided to track this down. The whitepaper referenced a company called Datamonitor. Its webpage cited a study by a research firm called Butler Group that was described as stating the findings about client virtualization. So, even further research was called for. I went to the Butler Group’s website (btw, Datamonitor is a information company and is the parent company of Butler Group). I poked around its website and found a summary of its study. While I was unfamiliar with the Butler Group, I believed it to be similar to Gartner or Forrester, an analyst firm that tracks trends, does studies, etc. Therefore, I wasn’t surprised that the entire study wasn’t available on its website.

However, I did find enough information to see that the Butler Group’s study did not conclude that the total payoff for virtualizing 1,000 client machines is $159,000. Its study indicated that power savings alone from virtualizing 1,000 client machines would total $159,000. That’s a completely different kettle of fish, as the saying goes. Given that power savings is only one component of client virtualization (indeed, it’s seldom cited as a client virtualization benefit, so would typically not even be addressed in the anecdotes I mentioned earlier), the financial benefits of client virtualization would undoubtedly be significantly higher. Probably high enough to make examining the option worthwhile.

The point of this post is not to criticize the Butler Group or Datamonitor or the whitepaper writer. A mistake crept in somewhere along the line—but that kind of thing happens. The point is, when making virtualization decisions, it’s critical to really examine one’s assumptions and the information used to make decisions. Accepting information from trusted sources without understanding its accuracy can skew a decision in the wrong direction—to your misfortune. Rather than blindly swallowing assertions made by an analyst firm—or anyone else, for that matter—you should evaluate the potential payback for client virtualization in your own environment.

In my book, Virtualization for Dummies, I strongly recommend performing a pilot assessment as a mechanism for exercising a potential virtualization solution. One of the aspects of a pilot program is to evaluate the operational characteristics of the solution, assessing the labor required to manage the solution, the ease of migrating existing systems, and so on. That information can be used to perform a financial assessment of the potential payoff of a virtualization solution (I also include spreadsheet templates in the book to ease the evaluation). From that assessment, you can draw a picture of the potential financial benefits of virtualization in your environment, which is surely the only important one, no?

So let this be a cautionary tale about the dangers of believing everything you read. This is particularly true in emerging technologies, where expectations and best practices are unsettled. Always examine your assumptions and look carefully at information put before you. And, above all, don’t let data that seems to diminish the potential for a solution sway you from looking at it carefully. There may be more, much more, to the story—and to your benefit.

Bernard Golden is CEO of consulting firm HyperStratus, which specializes in virtualization, cloud computing and related issues. He is also the author of “Virtualization for Dummies,” the best-selling book on virtualization to date.