Robert McLaws: Windows Edition

Blogging about Windows since before Vista became a bad word

The True Cost of Windows, Part 1 - Introduction

Now that Vista is out in the wild, lots of people are pontificating on whether or not it is worth upgrading to Vista. Ina Fried posted about this subject just yesterday. Sure, over the last few months, I've laid out my own reasons as well. But instead of having a psychological debate over security improvements, usability, et cetera ad infinitum… I wanted to take a different approach, and make the argument in the only way that matters to most people: money.

During all these discussions, the more I thought about it, the more I started to realize that Vista might not cost all that much more that other versions of Windows had in the past. Sure, the MSRP had gone up over the years, but how much had it changed based on inflation? So I decided to take emotion completely out of the picture, and embark on a scientific endeavor to uncover the truth about the cost of Windows.

So I've explained what this report is… here's what it's not:

  • A discussion of Windows v. Mac v. Linux. That discussion has been had to death, and I don't care to interject anything on that subject. Most people don't use either, they've been using Windows. This discussion is targeted at those people.
  • It's not another discussion about TCO. Besides being closely related to reason number one, Microsoft pays companies thousands of dollars to do TCO studies. If you want a TCO discussion, read one of those studies.
  • It's not a discussion about OEM v. Consumer pricing. I understand that most people are introduced to new Windows versions through purchasing new computers. While OEM pricing has an effect on the purchase price, the damage is obfuscated by the overall cost of the machine. Besides, I don't have access to OEM pricing. I'll let Ed Bott cover that discussion.
  • Finally, since we're not talking about OEM stuff, we're not talking about Tablet PC or Media Center SKUs. Prior to Vista, they were only available to OEM channels, and therefore outside the scope of this discussion.

Now that the discussion has been properly framed, let's get right to it. Every scientific study starts with a hypothesis. This allows the researcher to test their conclusions against a predetermined conclusion to see if the researcher's theory is right or wrong. Here's mine:


22 years of inflation has made Windows Vista cheaper than previous versions of Windows.


In order to test this hypothesis effectively, I must dig up the launch price of each major Windows release and find the cost in today's dollars.

The problem with this approach is there isn't any information on the Internet about any version of Windows prior to 1995. That's because Microsoft was unfashionably late to the Internet party, showing up a year later. Anything before that is consigned to some fallout shelter deep under Bill's house or something. I had to find another way to get that information.

I e-mailed a friend over at Waggoner Edstrom, and asked if she knew of any way I could get that information. She was able to get the Microsoft Archive to send me the press releases for every launch of Windows in Microsoft's history. You can check them out here (they'll also be listed under the "Sources" sectio at the end.)

The problem at this point is there have been a lot of different flavors of each release over the years. So how can I make each comparison as similar as possible?

In the end, I decided to group the releases into three categories: Basic Home SKUs, Advanced Home SKUs, and Pro SKUs. That way I could accurately plot the data without skewing the results, and I could make apples-to-apples comparisons. Now, some people may have issue with the way I grouped the SKUs, and that's fine. Keep in mind that SKU grouping doesn't really change the results of the raw data, only the way it is comparatively analyzed.

  • Home SKUs
    • Windows 1.0
    • Windows 2.0
    • Windows 3.0
    • Windows 3.1
    • Windows 95
    • Windows 98
    • Windows ME
    • Windows XP Home
    • Windows Vista Basic
  • Advanced Home SKUs
    • Windows 2.0 for 386
    • Windows Vista Home Premium
  • Pro SKUs
    • Windows 2000 Pro
    • Windows XP Pro
    • Windows Vista Ultimate

NOTE: Microsoft Marketing would have placed Vista Ultimate alone in the high end, and normally I would agree with them. However, this list is organized by the way SKUs were perceived to consumers at the time of release. This means that, to the consumer, XP Home would have been the low end, and XP Pro would have been the high end. Now, that would arguably put Windows 2.0 for 386 in the upper echelon, except that the only differences from the 286 counterpart were related to the processors… which IMO that makes it a midrange model. Also, Windows 2000 Pro is included because it was the beginning of the NT kernel on the desktop,and because it was available for purchase by consumers. Since it was available to consumers at retail, I'm including it. I won't be including Vista Business or Enterprise, however, because they are targeted at business customers, and therefore outside the scope of this discussion (yes, even though Vista Business is available for consumers).

Now that I had the editions grouped, and I had an accurate price list, I now had to calculate the inflation-adjusted prices in today's dollars. There are many schools of thought about how this is done, as well as some cheesy applets that do the calculating. In the end, I settled for the one from the Austin Genealogical Society. Why? Well they basically stole the one from NASA, and added in projections that go all the way back to 1665, from an Oregon State University professor who is referenced often on the subject. Seems legit enough to me, so we'll roll with it.

Ok, enough yakking. On to the data.

Continue to The Evidence -->[NoSharing:True]