Subject: Free Software Business Models
From: "D. V. Henkel-Wallace" <gumby@zembu.com>
Date: Tue, 01 Sep 1998 13:51:51 -0700

Wow, some excellent responses on my question and comment.

I may have been a little unclear in that I mixed two issues in the
same message.  The first issue is: there's more to a Free Software
Business than just a consulting business.  The second is: for such an
alternative, is it possible to build solid brand equity?

The first issue is very important.  I think everybody pretty much
understands the consultant model or labour theory of value.  It has
been the basis of so-called "fair" commerce around the world up until
new models began to appear in late 18th century Europe, and still has
many many adherents today, even in supposed capitalist powerhouses
like the USA (not to mention some denizens of this mailing list).

That model has a lot to recommend it, but I'm not convinced that it
can undertake significant investments (medium-sized examples like
Linux merely prove (i.e. test) the rule, they don't contradict it).  I
think we all understand that a libre approach can raise a barn but
probably can't build a suspension bridge or skyscraper.  I think the
same is true of software: a model designed solely around short-term
cost recovery cannot afford the risk necessary to undertake a major
project, (both direct prodction and promotion) and, because of such a
model, will have trouble finding enough "investors" in the pay-for-it
phase since the closer the project gets to payoff the less value in
paying.

The FSBs have a chance of evolving a 21st century business model to
join the 19th century and prehistoric ones.  Will we actually do so?


Some specific comments on others' comments, then onto the second
point:

   Date: 1 Sep 1998 03:07:36 -0000 
   From: Russell Nelson <nelson@crynwr.com> 

   In a free market, the only way to get rich is to sell the same
   thing repeatedly.  One way to do this is to sell the same
   organizing services to many people (usually called employees) at
   the same time.  I think this is the "non-linearly with the labour
   expended" you refer to above.
   
There is a small non-linearity in this that's true.  Though the real
value in this approach is the aggregation and coordination of
domain-specific skills.  But the change in slope is small: look at the
ROI of companies like EDS for example.  In fact the labor-selling
companies that do this the best (e.g. Macdonalds or Wal-Mart) do in
fact strain as much labour _out_ of the process as possible -- and
extract value at other points.

Another way to get rich is to sit on a critical resource, which you
can sell once.  Real estate is one example (we can't repeal gravity!)
but information can be another.  I take it as given that all FSBs
(libre or gratis) are by definition rejections of this approach.  Am I
wrong?

   Date: Mon, 31 Aug 1998 21:55:48 -0700 
   From: Brian Behlendorf <brian@hyperreal.org> 

   At 09:10 PM 8/31/98 -0700, Chris Maeda wrote:
   >Only the second business
   >model ("non-linear return on investment") can satisfy these people;
   >to them, the words "free software" are an oxymoron.

   Are these two concepts completely in conflict, never to be reconciled?

Precisely my point (and more pithily presented, John's): much of the
discussion on this list of late presumes that they cannot be
reconciled.  I offer Cygnus and Red Hat as (unfortunately the only
examples) demonstrating otherwise.

Another example of how hard it is to come to grips with this
distinction is Frank Hecker's thoughtful note.  At its root, his
argument turns, as yours does Brian, on the issue of "justifiable"
(his word) profit.

The investment issue is more complex then that.  Large profits have an
additional benefit quite apart from attracting outside investment.
They are also the _source_ of _internal_ investment (either through
reinvested profits or new external infusions of cash) in new
projects/R&D.  They are not only the inducement but are exactly what
permit one to take large development risks.  In this regard software
businesses are no different from any other.


Now on the brand identity issue:

   Date: 1 Sep 1998 03:07:36 -0000 
   From: Russell Nelson <nelson@crynwr.com> 

   [I said -- g]L
   > ...How does one build sufficient brand equity with libre (or
   > gratis) software?  That's one way to spike the exploitation
   > problem -- but it's tough to swing on the consultant model.

   [...]
   It comes down the the services that RedHat sells.  If they can
   perform a service once (e.g. identifying a reliable Linux kernel
   version, or creating a RPM of some free software package), and sell
   it many times under the RedHat label, there's your brand equity.

Does that have a natural limit?  For example Signum sells the Red Hat
distribution on a Signum disk.  Notice that I didn't say "resell".
Also Signum appears to tell people that it's the Red Hat distribution
because that (presumably) has demonstrable value to the actual
consumer.  I am guessing that Signum could get away with removing or
burying the RH name while keeping to the letter, if not the sprit, of
the GPL (no intent to besmirch Signum here, it's just an example!).
So yes, RH has achieved some success in branding.

But let me now tie my two issues together (business model/scale and
branding):

What happens if/as RH becomes _very_ successful?  All of a sudden
there's incentive for people to undercut the "official" distributions
perhaps by adding small, even proprietary pieces at the margins.  Yes,
they will continue to be small time players themselves, but the
question is, will syphoning off that revenue also limit RH to a
significant degree?  Could that act as a deterrent to investment?

Interestingly much of the popular opposition (such as there is) to 
Microsoft appears to be due to jealousy of their financial
success[*].  Will this same phenomenon start to cut back into the
revenues of a successful FSB?

Note that I'm presupposing that branding is the only way to
address this issue.  That could be wrong.

If I didn't think that these problems were soluble I wouldn't be on
this list.  And I don't know that either RH or Cygnus want to
publically discuss how they handle these issues.  But for the rest of
us (I include myself in there now since I'm no longer a Cygnus
employee), I think this is the crucial issue to resolve if FSBs (for
gratis or libre) are to be significant alternatives to traditional
software businesses.

g

[*] I also feel that Microsoft is a menace.  But that's due to certain
    of their business practices plus the inherent effects of their
    size vs the size of their industry and that industry's imaturity.
    Unfortunately I think most people dislike Microsoft merely because
    they're successful.