Don't forget that there are other industries having such remarkable properties as well: another example is pharmaceuticals and generally biotechnology. Or the motion picture and music recording industries.
Industries where the resources of the participating companies produce IP, not products (or products where the marginal cost of production is insignificant compared to the cost to design).
Biotech has startup costs that are orders of magnitude more than software - especially web stuff. People with PhD's + lab + equipment == lots of dollars. Multiply that by years for them to make anything close to being testable on people and it's a big chunk of money.
A big website or widely distributed software package is still extremely expensive to create and maintain. Less than a pharmaceutical company to start, perhaps, but pharmaceuticals don't have to be maintained once invented, and they're relatively protected from competition for the lifetime of a patent. Software companies must invest continuously in development, or their window of profitable operation will be narrow.
My point is that the massively profitable, "free" website run by three people in a garage is probably a myth. The massively profitable online retailer run by three people in a garage is definitely a myth. To reach the scale where non-trivial profits are possible from the internet, you've got non-trivial expenses.
You're dreaming if you think Craigslist is three guys in a garage. They have ~30 employees, and I can guarantee that they have large server, colo and bandwidth costs. They also make their money off of direct payments (as opposed to advertising), and their annual revenue is around $100 million, by the estimates I've seen. So they're profitable, but not hugely so. (By way of comparison, Netflix makes a bit over $20 million net per quarter, on over $300 million gross per quarter.)
I'll grant you that they're exceptionally small for a website of their size -- but being "exceptional" means only that they're the exception to the rule. The rule is what's important.
Actually, they are quite exceptional. Craigslist gets more traffic than either Amazon or Ebay and those companies respectively have 16,000 and 20,000 employees.
You've missed the point: Craigslist is so far and away the exception to the rule, it's practically non-reproducible. And for what it's worth, they're also not nearly as profitable as either Amazon or EBay, regardless of their traffic.
There's simply nothing about Craigslist that you can count on reproducing. If you're creating an internet company today, and you're aiming for hundreds of millions in revenue, it's 99% probable that you will need to spend more money than they do.
Well, of course you can reproduce their design. But reproducing their design won't reproduce their success. If it were that simple, every website would look like Craigslist.
Again: you're missing the point. Craigslist is the exception to the rule. It's a product of its time. For any value of X, if you tried to say "this is the way Craigslist did X, therefore I should too," you'd very likely be wrong. And in any case, Craiglist is pretty far from the romantic notion of three guys in a garage. It's an expensive site to run, even if it's more cheaply run than other big sites.
