Subject: Re: a business model
From: "Stephen J. Turnbull" <>
Date: Fri, 24 Mar 2006 22:50:43 +0900

>>>>> "Thomas" == Thomas Lord <> writes:

    Thomas> Stephen, please don't get me back in trouble on this list
    Thomas> by giving me too much to respond to.

I wouldn't worry about it, as long as you are responsive.  In any
case, I've been chided off-list for an excess of acerbity, so you have
my apologies.

    Thomas>         Since A and B are closely matched in the larger
    Thomas> competition, they both arrive at the same conclusion about
    Thomas> the value of the program.  Either they can both want those
    Thomas> ($L - $K) points against C or neither of them do.
    Thomas> Optimally, they each spend either $0 or $K/2 towards the
    Thomas> creation of the program.

There is symmetry, it's true, but when L > K, here it leads to a pair
of *antisymmetric* equilibria, in which one or the other pays $K.

I can think of reasons why that might not be the expected outcome as
easily as you, but you need to put those reasons into the model, or
the model is just a false face for an informal verbal argument.
(Nothing wrong with informal, but couching it in mathematics tends to
mislead the naive.)

    Thomas>   The free software innovator's problem is that she is a
    Thomas> commodity.  The customers need not conspire against a
    Thomas> greedy innovator: they can get equivalent product from
    Thomas> anyone willing to work for just the marginal cost of
    Thomas> production.

This is precisely the kind of facile assumption I object to.  You have
made the point that you think you're a "high-percentile" software
engineer, and that your product "excels".  I think that is important;
this industry is not about "anyone willing to work for just the
marginal cost of production."  OTOH, if the product really *is*
equivalent, I see no excuse for you getting anything but ... a job, at
minimum wage.  Me, I'm going to deny the equivalence.

    Thomas>   This is part of the point of my plan: an innovator
    Thomas> formed as a corporation (in it for the long haul, starting
    Thomas> from liquidity, dedicated to an unbounded number of
    Thomas> different projects)

IMHO this is a fatal contradiction.  The larger the number of
projects, the less *attention* the corpovator can give any one.  The
customers are not going to want to give the corpovator free rein;
rather, they are going to want (something resembling) an employment
contract.  "We pay the money, we call the shots."  If it's mission
critical software, they're going to want a controlling interest (and
that _would_ be an employment contract).

    Thomas>   Debian comes kind of close to smashing RH's price but,
    Thomas> because of structural elements in its constitution, will
    Thomas> never quite get there.  UserLinux took a stab at closing
    Thomas> the gap.  Ubuntu is taking a stab.  It's just a matter of
    Thomas> time before the community gets its organizational act
    Thomas> together well enough that RH's profit under the current
    Thomas> model evaporates.  From another direction, CentOS inspires
    Thomas> RH to clean up their own act.

I don't agree; those are all distributions.  RH is not in the business
of selling distros; those are not their competitors.  Or do you see
Debian, UserLinux, Ubuntu as organizations providing support?  Dunno
about Ubuntu/Canonical, but I disagree about the other two competing
with Red Hat in that area.

    Stephen> You mention "value-based pricing."  Well, that is exactly
    Stephen> what this is,[2] to the extent that it can be measured.

    Thomas>   Right.  RH's model is precisely value-based pricing.  I
    Thomas> think that's what I wrote.  And it's a bug because in the
    Thomas> non-rival, labor-rich world of free software, only pricing
    Thomas> equal to marginal cost of production is sustainable.

Once again, RH got out of the business of selling free software.
That's what Fedora Core (and in another direction, the FTP site that
is just as available to CentOS as to any of RH's paying customers) is

AFAICS, RH's current model is straight out of ESR (or Shapiro and
Varian, for that matter).  Build the software because it's useful to
your real business, then give it away because it *isn't* your real
business.  What does RH get from FS devo activity?  (1) Work-for-hire
fees, except that because of the GPL, unlike the usual work for hire
they are in a good position to get that software back for their other
businesses.  (2) More-reliable software making its support business
less risky.  (3) More-qualified support staff.

These factors allow it to make a non-zero profit sustainably as far as
I can see, because (as you point out) the attention of support
engineers is a rival good.  Value-based pricing is simply a way to
extract yet more profit.  The next Microsoft?  I think not.  But that
doesn't mean they can't sustain a profit.

    Stephen> The problem here is arranging that the "free software
    Stephen> innovator" *listen* to the paying customers.  You point
    Stephen> out how Canonical's entry dried up your GNU Arch market;
    Stephen> well, if your public utterances are any sample, you made
    Stephen> it plain to them that their perceived needs were the
    Stephen> *wrong goals* for you to devote effort to.  

    Thomas>   Close enough, yes.  And not just the wrong goal for me:
    Thomas> while Canonical Corp. is achieving market penetration I am
    Thomas> not aware of them yet having a sustainable business model.
    Thomas> I would point to my differences of opinion with
    Thomas> Mr. Shuttleworth as the areas where they have gone wrong.

So?  As I understand it, you weren't expected to ante anything.  Just
don't buy any stock (more than is required to keep up appearances with
your colleagues), and polish your resume while accumulating experience
holding back the flood.  When the dike fails, you should be able to
land a better job, maybe even with a company with a business plan you
can respect.

Point is, it's his money to bet.  He doesn't have to prove anything to
you or anyone else.  And Mark Shuttleworth has enough that there will
probably be a round two if it doesn't fly this time.  He might listen
to you then, if you listened to him the first time.  And he might even
turn out to be right enough to evolve the present business into a
success.  I agree, it would be nice if somebody who knows what he's
doing were running things.  But that's not the way the world works;
you have to "earn"[1] the right to run things.

What bothers me about your schemes is that as much as you talk about
listening to the customer, in the end the incentives the developer
would face (by construction) say "do what you think is right, because
you'll get paid anyway."

You could argue that that is merely aligning financial incentive with
professional ethics, but I don't think so.  It's aligning financial
incentive with the professional's personal interest.  If the
professional is expert and genuinely ethical, that's cool.  But if the
professional is incompetent, venal, or arrogant, it's not---and it's
often difficult to distinguish professional arrogance from
professional ethics.  (At least that's true of me, by introspection
and reflection on personal history.)  Who watches the watchers,
especially when they're watching themselves?

[1]  As we all know, "earned" altogether too often means "got lucky",
but nonetheless, that's the way things work.

Graduate School of Systems and Information Engineering   University of Tsukuba        Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
        Economics of Information Communication and Computation Systems
          Experimental Economics, Microeconomic Theory, Game Theory