There's a simple reason why, after 20 years, we still have analysts with big firms who are ignorant about open source.

There just isn't a lot of money in it.

(As always when you see this guy on this blog, I'm going to point you to the wonderful game from the good folks at Hasbro that has taught generations of kids money is just a game of picking cards and rolling dice.)

This is a feature and not a bug.

By its nature, open source reduces distribution costs to zero and marketing costs to near-zero. Even when an open source company has a sales department they're not really selling. They are converting. They are trying to turn people who have already downloaded and started using the software into paying customers. Their success is determined, in large part, by the prospect's success in getting value from the software.

In other words, the stuff sells itself.

Market analysis is not journalism. It's marketing. Analysts look for opportunities on behalf of buyers and sellers in a marketplace. They sell this knowledge in order to build markets. But what is there to sell if buyers can just download the stuff, and sellers are just converting users into buyers?

Long story short, open source is a direct threat to market research. Typical market research processes don't work in the open source market. Sure, you can get gross comparisons of lots of open source projects, to help you decide which one to put your time and money into. For a company of size this is valuable stuff. But the typical “magic quadrant” analysis just doesn't have a market. Sorry.

Rather than acknowledge this reality, firms like Forrester, Gartner and IDC dance past it, or pretend it's not real. They don't train their people to understand open source because there's not enough money in it – for them. So they don't know it. And they probably never will.