Subject: Re: Cygnus and proprietary software
From: John Gilmore <gnu@toad.com>
Date: Wed, 24 Dec 1997 13:23:47 -0800

> In the early stage, the proprietary software company (company P)
> is worth more than the free software company (company F).  Why?
> Because P is going to own their software and charge a lot of money
> for it while F is going to give it away and charge for support contracts.
> This is probably an order of magnitude difference in revenue expectations.

There won't necessarily be an order of magnitude difference in actual
revenue -- just in "revenue expectations".  I.e. this is a marketing
problem, not a reality problem.

Chris's analysis points out the problem pretty exactly.  He has focused
on all the downsides of a free software company, without noticing any of
the upsides.  So of course he concludes that the free software company
"can make a lot less money".

Cygnus dealt with this marketing problem for years by not caring what
the company was "worth".  If you don't borrow money or sell equity,
you never have to ask or answer that question.  Employees and founders
made personal judgments about whether it was "worth" it to accept
company stock as compensation.  As Cygnus got bigger and more
interested in its valuation, it started addressing the problem
directly, by researching the similarities and differences between its
own business and other well-understood businesses.  And then by
communicating the results to the community concerned with the "worth"
of businesses, the investment markets.

Cygnus' venture capitalists (we let two of them buy in, in January,
giving the company a $30M+ valuation) were not concerned that the
company's software was free.  They understood the free software
business model and believed that putting their money into it would
result in comparable returns versus other investments they could make.

One could say, "A proprietary program that had the market share that
gcc has, would be making tens of millions of dollars a year for its
owners."  But that ignores the question of how such a program would
get such a market share.  It also ignores the relative costs of
developing, supporting, marketing, and selling such a program.

Here's a short, incomplete list of free software business model differences:

	*  Companies become dependent on your product without ever
	   having to decide to buy it.
	*  Employees familier with your product are readily available
	   out in the world; they tend to seek you out!
	*  Users evolve your product to fill niches that you never knew
	   about, and some of these niches are much larger or more
	   profitable than your original niche.
	*  Your users generally regard you as less of an adversary and 
	   more of a partner or collaborator.
	*  Users are less willing to pay for software distribution, and
	   more willing to pay for custom services.

I know how little I know about marketing -- it's a vast subject of
similar complexity to computer programming, and just as essential to
the functioning of modern society.  Business management is another
equally vast and essential subject.  I have no formal training in
either; I'm just scratching the surface here.  But a one-sided 
treatment of "free" versus "proprietary" isn't useful to us in 
understanding how the world actually treats those models.

	John