Subject: Re: Successful FSBs
From: "Benjamin J. Tilly " <>
Date: Thu, 31 Oct 2002 20:53:20 +0500

eric hamilton <> wrote:
> > This somehow triggered a line of thought, and I now know why I
> > believe
> > Tim is wrong.  Ie, why I believe that defining FSB, and defining it
> > to
> > involve _developing_ open source software, is important.
> I agree on this point.  A company that relies on, but does not
> contribute to open-source development should not be considered an FSB. 

What does it mean to "contribute to open-source
development"?  Do good bug reports count?

> If it is not a large part of the company product or brand, how can we
> look to them for guidance?  Sure, they may have money in the bank, but
> how does that help me evolve MY business model if we're doing
> completely different things?  How does it help the solitary open-source
> software developer turn his geek-hobby into a business oportunity?

Bob Young in describing how he started Red Hat reasoned
that if customers wanted to buy something, then he could
find a way to sell it to them.  First you have to figure
out what they want, and why.  Then figure out how to
sell that.  Then figure out what it means about your

For some customers, that thing is for it to come in a
nice box.  For a sample success story of this see

Several companies (Sleepycat and Aladdin are two that
get mentioned a lot) arrange that only some people are
willing to be customers of the free product, and then
sell the proprietary one to those who need it.

In other cases you have to sell them something that is
tangentially related.  Consulting and training are both
mentioned quite often.

One important point to note.  Many of the plausible
business models around OSS are very low margin.  With
low margin businesses you want to cut expenses and
then try to create a fast turnaround from money spent
to money made.  Remember that selling twice as fast is
as good as doubling your profit margin.

The unfortunate fallout of this is that you can make
money, but you cannot afford to turn around and directly
fund a lot of development in the process like you would
like to. :-(

> > What we are saying at this point is that getting private financing
> > (which is where the money is; ask any country, including the U.S. or
> > Japan, faced with the prospect of fighting off George Soros &cie)
> > seems to be nearly impossible for FS.
> The money in private financing is dependent on the idea that eventually
> (3-5 years down the road), the company will be profitable enough to buy
> out the investment.  The question at hand then, is how can we make that
> happen -- let alone sell the idea to investors?  Trying to figure out
> how to achieve financing before you have a provable business model is
> putting the cart before the horse.

No cart/horse problem here.  A common characteristic of
OSS business models seems to be that they are low-margin
commodity businesses.  Such businesses can be stable and
profitable - just look at your local supermarket - but
do not offer the kind of rapid growth that venture
capital salivates over.

This means that there are viable business models.  Just
not ones that venture capital is likely to invest in.

> > Nor are the companies Tim is lobbying going to suddenly come to their
> > right minds and start funding gobs of FS development.  They're
> > already
> > in their right minds, and that's why they _don't_ fund FS
> > development:
> > it's a net loser for them, even in the long run.
> If that were completely true, companies like IBM would ignore the
> open-source community.  That obviously isn't happening.

Received wisdom on fsb seems to be that IBM has noted
that it is far better to be a customer of free software
than a vendor.  Yes, it funds many free software
projects.  And then pays for it with money saved by not
having to buy that software from elsewhere.

I am not sure what facts and figures actually justify
that received wisdom.  But it is a very reasonable
reason to develop free software.  Improve free software
that you need to use, and save money.

> Who says that the open-source software itself needs to be "readily
> commercialized"?  I can think of a number of ways to sell OSS projects
> to Uncle Sam.  The right strategy depends largely on your intended
> business model, though.  Obviously, simply selling the software (or
> some proprietary extension of it) is not enough.

Convince Uncle Sam that it wants to be a customer, and
like IBM it can sell itself what it needs.

You might sell them services to help them accomplish
that goal.  You might sell them something built around
it.  But there are no cents in developing things when
you think you can find someone who will either develop
it for themselves or (even better) will pay you to add

> > So maybe we have to call 1-800-Mephistopheles and try weakening
> > _some_
> > of the OSD strictures against proprietary licensing, while
> > maintaining
> > the ones that have the most social benefit.
> I doubt this will help much, asside from allowing more software to work
> together.  In terms of market climate, though, it won't change a thing.

See how Sleepycat works please.

Also how IBM actually makes money off of Linux and

Get your free email from

Powered by Outblaze