Subject: Re: Successful FSBs
From: "Tim O'Reilly" <>
Date: Wed, 30 Oct 2002 08:30:40 -0800

On 10/29/02 10:53 PM, "Benjamin J. Tilly" <> wrote:

> Of course a crumb from a big market may be sizeable
> enough.  And the FSB opportunity may lead to opportunities
> in other markets.  But make no mistake, any FSB has to
> understand that OSS dynamics put consumers in a very
> strong bargaining position.  You need to work with that
> dynamic, not against it.

For that matter, OSS dynamics put competitors in a very strong position as
well, and competition keeps prices down. (That's another way of saying it
puts consumers in a strong position.)  You can build a huge business with
commodity software (viz. Uunet or Earthlink in the ISP market), but it's not
going to be anywhere near as profitable as a similarly sized proprietary
software business.

I tend to see F/OSS advocates wanting to have their cake and eat it too.
They imagine being as big and successful as Microsoft or Oracle with a
business and licensing model that is *designed* to keep companies from being
as big and financially successful as MS or Oracle.  I actually think that
F/OSS is as big and successful, if you look at impact, but not if you look
at dollars extracted from users.  And that is as it should be.

There's another important factor, which I picked up years ago from Bill
Davidow's book Building High Technology.  By definition, "dominating a
market" means being more than 50% of the market, which means that your
company has to be able to become half the size of the market as quickly as
the market gets there.

This means, for disruptive technologies that are also commodities:

* You are probably under the radar, so financing will be hard to come by
* Once you are no longer under the radar, competitors can easily jump in

In such circumstances, it is very hard to become dominant without
identifying *some* defensible proprietary component, or without pouring in
so much money that you can grow faster than the market.

FWIW, this is the thinking that led me to sell GNN to AOL when the web
started to take off.  I realized that without huge cash infusions, we
wouldn't have been able to keep up with the market, and with those cash
infusions I didn't think we'd be able to stay O'Reilly as we were, with the
values we had.  I tried (and failed) to have my cake and eat it too, by
selling GNN (which was the first web portal, before even Yahoo!) to AOL,
defining it as a distribution outlet for other web publishers (of whom we
would be one).  But of course, as often happens in an acquisition, the
strategy that everyone bought off on during the sale is quickly shelved once
the deal is done.  But that's water under the bridge.  And also a big
digression from the topic...

Back to an example from FSB.  In the case of ISPs (a business mainly built
using commodity TCP/IP stacks), Uunet came closest.  Their strategy was to
dominate the backbone -- to be the best connected site, with lock-in on on
key interchanges -- and it worked pretty well for them.  So they didn't lock
in on software but on network topology.  Meanwhile, AOL and Earthlink
followed the strategy of customer acquisition at all costs, pumping huge
sums of money into folks who would pay off in the future.  But even with all
these strategies, no one became dominant in the way that Intel, or
Microsoft, or Cisco became dominant with their respective, proprietary IP

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