Subject: Re: Pro Serve Revenue for FSB
From: Rich Bodo <rsb@ostel.com>
Date: Tue, 5 Nov 2002 07:25:58 -0800 (PST)

Doug,

> Services using F/OSS are not equivalent to Support Sellers.[1]

Yes.  Eric's paper is dated by it's categorization of RedHat as a
Support Seller.  He describes components of an F/OSS business model,
rather than business models.  Everyone has space/time constraints,
though.

In your
> examples, the companies which derive value from F/OSS are those which
> provide a service in which the software is infrastructure.

I was actually trying to show three different types of businesses that
each put major resources into developing free software, contribute
back to the F/OSS domain, and offer a complete solution.  It is
possible to run any one of those businesses and either contribute or
not.  I could put together a list of reasons to contribute, I suppose.
Note to self: Make a list of existing FSB types and reasons they
contribute.

> The F/OSS
> software is a replaceable component and not contributing to the revenue
> potential of the company. (F/OSS does contribute to profitability, time
> to market, etc., but not the size of the market nor a higher price for
> the product.) There isn't a revenue incentive to write F/OSS software,
> although the ancillary benefits should be a strong incentive to be
> active in the F/OSS ecosystem.[2]

I don't see a point here.  We can make similar statements about F/OSS
software and proprietary software.  Proprietary software is
replaceable with free software and vice versa.  You can sell free
software at any price, and use it to help you sell other things, and
that's revenue.  How does proprietary software contribute to the size
of the market or a higher price for the product in a way that free
software does not?

> My point is: If you are developing F/OSS with the intention of making
> money based upon your expertise, the market forces work against you. As
> your software becomes successful, your value in the market
> decreases.

I have to agree with you.  That's a good point.  However, I reserve
the right to come up with a slick rebuttal later. ;)

> The consequence is that there is little incentive[3] to be a
> developer of F/OSS as a core business.

We have discussed some business incentives on the list in the past.
FSB's may be easier to bootstrap or provide a strategic advantage.
There are, of course, the personal reasons you mention as well.
Making money based on your expertise and hard work seems honorable, if
not the wisest business decision.

> Your "product" - that which makes you money - must meet one of the
> following criteria: (1) complexity (as defined in my previous post)
> (2) dual license to meet market needs Or (3) Not be the F/OSS at
> all, but some proprietary software, hardware or service you sell
> alongside[4].

These are all pretty useful ideas for FOSS bus models.

> This is important when building a business model to
> fund the development of F/OSS inside a company or as a new
> venture.

> As has been discussed here, the margins on proprietary SW
> distribution, support and professional services are high and
> sustainable. I've read business plans for F/OSS development which
> assume margins for similar functions provided by the original
> developers will also be high & sustainable. I believe that to be an
> exception, not the rule. Not having a sustainable model makes it
> very hard to bring new F/OSS projects to market.


> [4] See Frank Heckler's article: Setting Up Shop
> http://hecker.org/writings/setting-up-shop.html for other models.

I'll check that out.  Thanks.

> TIVO & Web ASP don't match the definition of "Service Enabler" as
> defined by Frank Heckler. They do not create the software on which they
> base their business. They chose F/OSS as 'best of breed' or lowest cost.

O.K., they might not match the Service Enabler definition, but TIVO
does create a lot of the software they base their business on.  Yahoo
does too.  They hire prominent developers, and they generally
contribute their improvements back to the FOSS domain.

> The call center and the telephone manufacturers in your previous post
> appear to be examples of "Widget Frosting" where the advantage of
> interoperability outweighs any added value proprietary sw would bring to
> the product as a whole. BTW: That advantage is not exclusive to F/OSS,
> but is true of any standard in the public domain.

I install call centers.  It's what I would call a hybrid business
model.  I wouldn't say we primarily sell hardware.  We develop the
core server software as free software, we write "custom software" for
each customer (that is not typically generally applicable and
copyright the customer), we sell hardware and support.  In many cases,
we get calls for support or service or custom development from people
who have not bought hardware from us.  We will also be releasing a
line of turnkey servers soon.  It's even more complicated than that,
way too complicated, in fact, but it seems to be working.  In our
business, most people want a complete solution for a variety of
reasons.  That's true for both free and proprietary solutions.

Wish I had more time to discuss this one today.

-Rich