Federico Lucifredi writes:

 >   Postulate: Market analysts tell us that the #1 competitor to F/OSS 
 > companies is internal expertise at the customer site:

This is true of all b2b services.  Consider the business of management
consulting.  If you know how to run your business, why call in Boston
Consulting Group?  And it's often true of proprietary software firms,
too!

But having competition just means you have to be better than the
competition to make a profit.  It doesn't mean you can't get those
customers and make a profit, too.

 > in short, if the local crew is smart and attuned with the state of
 > the software they want to deploy, possibly even maintaining ties to
 > the relevant part of the community, they will deploy and support
 > said software themselves.

Not necessarily.  It depends on what kind of business and what kind of
smart local crew.  If the software is not mission-critical to what the
business is going to do *next* but the smart crew is critical to the
new task, it may make a lot of sense to outsource the basic software.
The trick is recognizing such a business and selling to them---they
may not have recognized that you can do what they need for the launch
platform while the "smart crew" gets to work on the rocket itself.

 >   Given the premise above, it looks like more often than not F/OSS
 > vendors are vying for the business of the customers who do *not*
 > have sufficient on-site expertise - in other words, it looks like
 > one might be selecting customers coming from the most clueless part
 > of the pool!

Small consultancies of whatever kind have this problem.  It's endemic,
and an important factor in the very high bankruptcy rate for such
businesses.  However, "might" != "must", and the successful consultant
has learned to recognize customers lacking both general business and
specialty clue, and charge them accordingly.

 >   Why is this an F/OSS concern? Well, [...] while in the
 > proprietary market, all customers must purchase support from you,
 > in our brave new world, only the less clueful need to.

Well, if you truly have a monopoly on your product, it's true that all
customers must purchase support from you.  However, the smart
customers can, and often do, drive a very hard bargain in return for
that lock-in in terms of up-front concessions---even from monopolies.
And if the proprietary product is one of several, it's theoretically
possible that competition for locked-in customers will drive excess
profit to zero.[1]

So the FLOSS services vendor must emphasize the advantages of the
freedom to change partners, and then make sure the customer wants a
long-term relationship at a fee you can grow rich on.  A tall order,
but it can be done.

I'd like to emphasize a point that several other posters have already
made: there's a difference between business clue and technical clue,
and the sweet spot is a customer with lots of the former and not much
of the latter.

Footnotes: 
[1]  I'm not sure I believe this myself, but judges and juries in
antitrust cases claiming monopolization by such vendors have been
convinced to find for the defendent.