Subject: Re: Successful FSBs
From: "Tim O'Reilly" <>
Date: Mon, 28 Oct 2002 08:40:51 -0800

You're right.  I overstated the case.  But I stand by my point.  It's
extremely unlikely that IBM *could* have managed to become either Microsoft
or Dell, for the reasons you outline.  But to take these counterexamples to
mean that it's impossible for a company or individual who understands what
they've got to realize the fruits of their innovations is equally

Rick Adams *did* come up with the right business model for his free software
(B News), while Paul Vixie (Bind) did not.  IBM has integrated free software
into its business model, while rivals have not.  (I was recently talking to
an SAP executive, and it was really clear how envious and impressed they are
with IBM's strategy for its internet software stack.  While SAP contributes
1/3 of rival Oracle's revenue, the software that IBM doesn't own is
commodity free software that provides no advantage to competitors.)

So yes, it's unlikely that a large company can benefit from its own
disruptive innovations as well as more nimble rivals, but it's also true
that a company that understands the disruptive landscape can do a better job
than one that doesn't.  Even a marginally better deal (for IBM) between IBM
and Microsoft might have had very different outcomes for the computer

On 10/28/02 6:25 AM, "Benjamin J. Tilly" <> wrote:

> First of all that is like asking Linus Torvalds whether
> he wishes that he had kept Linux proprietary and made
> money from it.  In hindsight the project was a success.
> But it would never have succeeded if he had maintained
> control.  And likewise the commodity dynamics that made
> the PC what it is wouldn't have worked if IBM had tried
> to drive all sides of the bargain at once.

Right.  We see that fallacy played out over the past seven years with Sun
and Java.
> Secondly you are ignoring an important point that
> Christensen made, which is probably a standard piece of
> knowledge from some business classic that I don't know
> about but should.  (References appreciated, I have seen
> similar points elsewhere.)  And that point is that the
> structure of an organization defines both its abilities
> AND disabilities.  The kind of corporate culture and
> organization that drives Dell's commodity model also
> makes it incompetent at taking routing big long-term
> risks.  Conversely Microsoft can take that long-term
> view, but could never compete with Dell in a
> cost-concious cookie-cutter environment.  These need to
> be separate organizations with radically different
> structures to succeed.  It is therefore disingenuous to
> say that if only IBM knew better they could have taken
> advantage of every opportunity arising out of the PC
> when what you have to do to take advantage of one option
> necessarily makes you incompetent at taking advantage of
> the third.

Not necessarily.  Some companies recognize that different businesses have
different dynamics, and are able to adapt accordingly.  Heck, just look at
IBM's success at the service business as well as the hardware business vs.
the struggles of some of their competitors to adapt to that model.

And take a look at Microsoft's success in building online services (Expedia,
MSN, etc.)  The jury's still out, but I think that they are going to be long
term winners in a lot of the spaces they are competing in there, as well as
in their original core business.
> And last, but not least, for each winner that we know
> about there are many losers.  It is tempting to think
> that support from a parent company can load the dice.
> But for reasons that Christensen gave, attempts to load
> the dice tend to backfire.  (Bob Young used to have a
> great speech comparing how Red Hat and Caldera started
> which illustrated this for me.  And the latter third of
> _The Millionaire Next Door_ shows the same dynamic in a
> different context.)  If the market lends itself to
> having several winners, then the gamble is not so bad.
> If the market lends itself to winner take all, then that
> is a very dangerous and psychologically hard risk to
> take, even if it is justified.

Agreed.  There's a great book called The Roots of Coincidence by Arthur
Koestler, which explains why hindsight makes a lot of things look more
significant than they actually were.

And that makes me think of the time Jon Orwant was asked by the NY Times to
debunk the bestselling book The Bible Code (which claimed to find all kinds
of revelations hidden in the Bible.)  Jon wrote a perl program that would
taken any desired text, analyze the bible's text, and generate an algorithm
to find one in the other.  For the Times, he developed an algorithm to find
the cast of Melrose Place hidden in the Bible.

Tim O'Reilly @ O'Reilly & Associates, Inc.
1005 Gravenstein Highway North, Sebastopol, CA 95472