Subject: Re: Larry Ellison on FSBs
From: "Ben Tilly" <>
Date: Thu, 27 Apr 2006 20:35:28 -0700

 Thu, 27 Apr 2006 20:35:28 -0700
On 4/27/06, Stephen J. Turnbull <> wrote:
> >>>>> "Ben" == Ben Tilly <> writes:
>     Ben> You were positing that it is a disadvantage for MySQL that
>     Ben> what they put together at cost to themselves can be borrowed
>     Ben> for free by Oracle.  I'm pointing out that maintaining their
>     Ben> software is not a cost center for MySQL, so that isn't a
>     Ben> disadvantage.
> Hm?  *By contract, those programmers cannot work on anything else.*
> That's what "cost" means, resources that you can't use for some other
> prupose any more.  Oracle's programmers can work on anything Oracle
> wants them to, or it can fire them and just distribute (the free
> portions of) MySQL.  That's what "zero cost" means.

I'd be shocked if MySQL has been stupid enough to set up contracts
that are so disadvantageous to itself.  Rather, MySQL gets outside
funding to have programmers do projects on the innards of MySQL. 
Those same programmers are then available to MySQL for other projects,
and bring with them the knowledge and expertise that they developed
working on the innards of MySQL.

That means that MySQL has a better understanding than Oracle does of
the strengths and limitations of MySQL, and what people would like to
be different about MySQL.  For Oracle to develop that expertise would
take time and energy for Oracle that nobody is particularly willing to
pay Oracle to develop.

> Oracle is free in theory to respond to any threat posed by MySQL by
> doing exactly the same.  As you've pointed out, it's harder to walk
> the walk than to talk the talk, but to hurt Oracle, MySQL has to start
> walking Oracle's walk, at least a little bit.  MySQL is *not* free to
> do that, because it has small revenues, small margins on that revenue,
> and much of its revenue is ear-marked for "listening to exactly what
> the customers say, and then doing it to spec".

I didn't say that MySQL was planning to go into direct competition
with Oracle across the spectrum of what Oracle does.  I did say that
they are going to undermine Oracle's core market.  And I said that I
believe that a lot of Oracle's non-licensing revenue is tied in a lot
of ways to their core market.

> *By contract.*  Sure, Oracle is bound by tradition and shackled by its
> own success, but companies do occasionally escape all that, and
> reinvent themselves.  MySQL isn't allowed to.  Oracle sounds like the
> process is well underway.

Leaders of companies frequently make pronouncements that their
companies have made huge changes to face the future.  Typically the
reality falls short of the announcement.

> True, MySQL gets the advantage of being focused on what it must do.
> But it is at the mercy of the market not moving too far, too fast in a
> direction it's not prepared to go.  Oracle has the luxury (and the
> risk) of trying to get to where the market is going before it gets
> there, and winning big.  I wouldn't give odds on it succeeding at
> that, but I'm sure they're not prohibitive if you're a betting kind-a
> guy.

Attempting to pilot without feedback loops in place is generally a
good way to crash.  I'd therefore see Oracle's luxury as more of a
risk if it attempts to seriously exercise it.

>     Ben> If Oracle tries to get involved [in MySQL's market segment],
> I'm sorry.  I admit that my words make it appear that I'm thinking
> about Oracle trying to rub out MySQL, but I'm not.  I don't think they
> would if they could, and I don't think they can without posting a
> spectacular loss (by FSB standards, at any rate).

OK, we are on the same page there.

> All I'm concerned with is the claim that MySQL (PostgreSQL, etc) poses
> a threat to Oracle's profitability in the medium to long term, given
> that LE has openly said that Oracle cannot compete with open source on
> its own terms, and is already successfully moving his business model
> in directions where Oracle won't have to compete with open source (or
> so he says).

Let me reframe what LE said.  LE said that Oracle cannot compete in
the medium to long term with open source, and as a result is preparing
to flee what used to be a profitable market for them.  That's a very
dangerous pattern.

CC's theory says that as disruptive innovation heats up, you expect to
see a pattern where incumbents flee markets as the innovators enter
them.  As long as that motion continues to be one-directional, the
incumbent eventually runs out of places to flee.

>     Ben> What MySQL owns is a competitive advantage in maintaining the
>     Ben> maintainership of an open source project.  That open source
>     Ben> project is a future threat to Oracle's installed base.  It is
>     Ben> both owned and open source.
> Nobody owns a competitive advantage, with the important exception of
> intellectual property.  Lacking a patent, they must earn it, and then
> earn it again tomorrow.  I don't understand what you are talking about.

MySQL is dual-licensed.  As a result, MySQL is the sole maintainer of
that codebase, and will continue to be so unless someone is willing to
fork the project from them.  The difficulty of getting people to
accept such a fork is a competitive advantage that MySQL has in
resisting outside attempts to take over maintainance of the code.

MySQL still has to earn their position as maintainer, and they still
have to earn it again tomorrow.  But they have raised a barrier to
entry for that position, and that barrier to entry makes it  easier 
for them to continue earning that position.

Now if something is easier for me than it is for you, then I have a
competitive advantage over you.  So I'm saying that MySQL has a
competitive advantage over anyone else who wishes to start maintaining

>     Ben> Fine.  I can tell you from direct experience that Oracle's
>     Ben> software is orders of magnitude more difficult to install and
>     Ben> maintain than MySQL.
> Nobody says otherwise.  I'm saying that Oracle-brand MySQL would be at
> most one order of magnitude harder to install than true-blue MySQL,
> and that should be enough to defend Oracle's market from MySQL.

If Oracle-brand MySQL is an order of magnitude harder to install than
true-blue MySQL, then Oracle-brand MySQL can be expected to have a
market that is at least an order of magnitude smaller than true-blue
MySQL.  In fact I believe that this is a big part of the reason why
Postgres is so much smaller than MySQL, despite being better than
MySQL by virtually any technical measure that one normally compares
databases by.

>     Ben> The result is that Oracle can't go head to head against MySQL
>     Ben> and outcompete it unless Oracle is willing to throw a very
>     Ben> large amount at the problem.  (And if they did, the result
>     Ben> would be an opening for another company like MySQL.)
> I'm sorry, you're responding to something I should not have written.
> My bad, but can we get back to why MySQL poses a threat to Oracle, if
> MySQL is operating in a market segment that Oracle doesn't even want
> to penetrate, and never will?

MySQL is a threat to Oracle because of an inherent asymmetry in their
relationship.  Oracle can never be interested in MySQL's business, and
MySQL wants Oracle's.  On the margins where they wind up in a head to
head competition, MySQL inevitably wins.

Given that MySQL can improve its ability to deliver on customers'
needs faster than those customers needs change, inevitably MySQL will
cannibalize Oracle's market.

>     Ben> As for Larry's claim, I'm sure that he sees the need to
>     Ben> diversify and I'm sure that he has done a lot of
>     Ben> diversification.  But I strongly suspect that as Oracle the
>     Ben> database falters in the market, Oracle the company will show
>     Ben> the hurt.
> Sure.  *If* the database falters any time soon.  As Mark Twain said,
> "the reports of my death are greatly exaggerated."

Startups 8 years ago virtually inevitably used Oracle.  Startups today
don't.  I've heard about many migrations from Oracle to Postgres or
MySQL.  I haven't heard of migrations the other way.

Given how slowly the database market moves, I wouldn't expect to hear
that Oracle was in great trouble this year or the next.  But over the
next 10 years, the writing is pretty clear on the wall.

>     Ben> I don't believe that open source businesses will get to where
>     Ben> his business did - it is the nature of the game that open
>     Ben> source businesses have thinner profit margins.
> That's not the "there" I'm talking about.  The "there" I'm talking
> about is when open source becomes a threat to Oracle's margins.

Larry Ellison already accepts that in the future his margins don't
include licensing revenue.  How is that not a case where open source
is a threat to Oracle's margins?

Furthermore losing those margins removes a competitive advantage that
Oracle enjoys in the new businesses that it was going into.

I remember a few years ago that Tim O'Reilly was talking about, I
think, Oracle versus Peoplesoft.  Peoplesoft had a big problem - if
someone bought their product they also had to buy a database.  Which
was often Oracle.  By contrast Oracle could go after the same customer
and could sell them the database as well.  So Oracle made more per
customer than Peoplesoft did.  And Oracle made money off of
Peoplesoft's customers.  Which means that Oracle could significantly
underprice Peoplesoft.

Now Peoplesoft was good enough that Oracle couldn't outcompete them,
even with this advantage.  But still this is an advantage that Oracle
currently has in a lot of its new business lines, and Oracle's life
will become harder after it loses this particular edge.

>     Ben> However I don't think that he'll find that his business
>     Ben> forever has the kind of generous margins that it has become
>     Ben> accustomed to.  I could be wrong.
> Depends on how you define "his business".  If you mean "the Oracle
> database licensing business", I'm sure you're right.  If you mean the
> Oracle company, I'm willing to bet that LE will bequeath a very
> healthy margin to his successor.

I meant the Oracle company.  Depending on when LE bequeaths Oracle to
his successor, you may be right.  But I doubt that Oracle will enjoy
its current size and margins in, say, 20 years.

>     Ben> When incumbents try to offer a low-cost alternative they
>     Ben> generally fail.
> Ever heard of Walmart? :-)

Touche, but addressing a different point than the one that I was
making.  Walmart shows, of course, that a mature company can be built
around a discount model.  It doesn't show that a mature company that
isn't built to have razor-thin margins can figure out how to compete
against one that does have such margins.

I was talking about things like Sun's failed attempt to sell Linux
servers.   Sun goldplated the wrong stuff and couldn't deliver what
customers actually wanted at a competitive price.  Hence Sun's
attempts didn't sell.  Christensen cites many similar examples of
incumbents trying to go head to head with disruptive innovators and
losing badly.

>     Ben> Christensen would predict that MySQL will make plenty (for
>     Ben> it) of money in market niches that Oracle can't be interested
>     Ben> in because Oracle can't make enough money there to make a
>     Ben> profit.
> So would Larry Ellison.  What I want to know is how MySQL is going to
> cause Ellison to spend more than 5 minutes thinking about it---until
> he decides to buy it so he can get its programmers.

The fact that Ellison took a public position on open source suggests
to me that he has already spent more than 5 minutes thinking about
MySQL.  Also Christensen's second book documents how hard it is to
purchase a disruptive innovator and not accidentally destroy what
makes it work.

> ObRef FSB:
> As far as I can see the bottom line is still "On the one hand FSB is a
> subclass of lifestyle business, and on the other no real threat to the
> smarter incumbents."

A number of examples show that FSB can be more than just a lifestyle
business.  However I'd heartily agree that the biggest FS business
opportunities are situations where open source enables some other