Subject: Re: dual-branding (was: Re: Paper on dual licensing)
From: Adam Turoff <>
Date: Wed, 4 Dec 2002 16:21:04 -0500

[apologies for the length.  Z.]

On Wed, Dec 04, 2002 at 02:59:52PM -0500, David Kaufman wrote:
> while coincidentally open-source, so that the laziest customers, with deep
> pockets and a desire to fulfill their instant gratification needs, can buy,
> plug and play, while those customers digging a little deeper can still find
> the value added by the commercial version... "okay, if i want to compile it
> myself, i can get the free version, but if i want the GUI installer and
> online updates, i can pay to get that" and this lets those customers (if
> they buy) at least feel that they made an informed decision.  unhappy
> customers who bought it and realized, only after the sale, that the
> software was free (of cost) elsewhere, could loudly complain and provide
> an awful lot of ugly anti-product viral marketing...

Why is it we're still perpetuating the same five-year-old memes as
if they're the gospel?  esr had a handful of speculative business
models for open source development that remain untried, unproven
or unsupported.  I don't buy that an FSB predicated on the customer
gullibility and customer laziness is a valid (or viable) business model.

It's been five years.  What have we seen?  Mikko has found a novel
trend: dual-licensing of open source software.  Each of the businesses
profiled in this report are profitable, and have been for a number
of years.  The thread that unites them is that they each started
an open source project, and maintain the copyright over that code.
That gives them the very valuable option of making commercial
licenses avaiable.  

None of the companies profiled in this paper use the "gullible
customer" model.  While RedHat and others may be using a model that
*looks* like this, there's a deeper drive behind open source business
than "give me convenience or else".

One common theme that I *do* find between the three companies that
Mikko has profiled as well as SourceFire/Snort, Easy Software
(CUPS), RedHat, et. al. is that they're taking a two-fold approach
on software adoption.  As Tim O'Reilly is fond of pointing out,
it's the alpha geeks that will make or break a new product.  How
do open source software appeal to alpha geeks?  By being totally
unencumbered.  Open source software doesn't require NDAs, purchase
orders, or come with restrictions (30-day evaluations, limited use
and the like).  A lot of it is also highly portable, or somewhat
easily ported.  So open source software at least has a *chance* of
being seen and used by the alpha geeks.

The whole issue of "you can fix bugs!" is a red herring.

What happens after Berkeley DB, MySQL and Qt get in the hands of
alpha geeks?  They get incorporated into other software.  The
internet becomes one great showroom for these products, and everyone
has a chance to kick the tires.  Why?  Because they're unencumbered.

That's not the only benefit to the dual-licensing vendor.  Alpha
geeks provide a very effective form of viral marketing.  They also
provide excellent feedback.  They will be your best source of bug
reports and feature requests.  On a good day, they may even help
your documentation project.  On a very good day, they may even
submit patches to fix bugs or add features.  None of these companies
rely on the world to be beating a path to their servers with
docpatches or bugfixes, though; they all know it's just icing.

The next phase for any product is crossing the chasm, or making it
into the mainstream.  This is where customers who want to incorporate
open source software into a proprietary product have issues.  Paying
for a commercial license to a dual-licensed product is an advantage
because a mature open source product (MySQL, Snort, Berkeley DB)
will have both an active, easy-to-find user community (for support),
as well as a viable vendor.  Therefore, incorporating this software
into your product is a safer bet than using some component you find
in the Programmer's Paradise catalog by some fly-by-night software
vendor you've never heard of before.  Plus, dual-licensed products
are a very safe bet; if the vendor does go out of business or stop
supporting its product, there are a lot of people with a vested
interest in seeing bugs fixed and features added -- and they have
the means to make those changes happen.

This to me is a much more plausible explanation of how and why
MySQL, Sleepycat, SourceFire and TrollTech are in business today.