Since Daniel references us (Circus Ponies) in such a complimentary manner I could not help but chime in with my two cents about the attractiveness of developing a product for the Mac market.
There is another aspect to the market share discussion that tends to get overlooked, and that is the signal-to-noise ratio in each of these markets. The noise level in both the Windows and Linux markets is deafening. There may be vast numbers of users in these markets, but creating mind share is near-impossible on a bootstrap budget.
In addition, in the Mac market, where the majority of users really "get it" about great software, your cost-per-meaningful lead is much lower, especially if you have a great product. As Paul Hawken says, what consumers really want is not actually that complicated a sell. Customers are people with problems. In our case, that problem amounts to the difficulty of staying organized while drinking from the digital fire hose. Give people a great solution to their problem and listen to their input and they will bend over backwards to help you to build a market for your product -- and contribute great product ideas in the process. Word of mouth is still possible in the Mac market.
As far as I am concerned, Apple's #1 job is to build a healthy and thriving market for us to sell into. The customers who buy into the expensive hardware are committed and passionate and are not averse to improving their experience with new software products that get great reviews and offer value for money.
I have learned that you can succeed only if you are fearless of failure. The Mac market's receptivity to new ideas outweighs its size as a platform for building a software business. All the rest is just commentary and scaling your business model to fit what the market can bear. We have no need to be the size of General Motors.
Circus Ponies Software