Oracle as an investment

On June 24, 2002 I bought 250 shares of Oracle Corp. (ORCL) at $8 a share.

The case against Oracle

Oracle's main product is the Oracle database. It's a very good product and it is incredibly profitable. But it is a platform. People don't buy Oracle because they want to store and retrieve data. They buy Oracle so they can sell things over the internet or track millions of packages or automate accounting procedures or whatever. These applications, which are built on top of the database, are the real reason most people buy Oracle.

To make an analogy, Oracle sells T-shirts and other companies put designs on the shirts and sell them to the end customers. (In this analogy, OS vendors would sell the cloth and hardware vendors would sell the thread. Or something.) Most people won't buy a blank T-shirt, but they will buy a T-shirt with the Lakers logo on it. So a clothing manufacturer might produce a shirt for $2 and sell it to the Staples Center for $5. The Staples Center pays the Lakers $2 to license the logo, spends $1 to put the logo on the shirt, and sells it for $16. Everyone's happy.

But what happens if another clothing manufacturer sells T-shirts for $4? If the $5 shirts are significantly better than the $4 shirts (thicker fabric, stronger seems, better fit, etc.), nothing happens. The Staples Center knows that they'd lose sales if people noticed they were paying $16 for poorly made T-shirts. But if the shirts are nearly the same, watch out. Pretty soon, both shirts will sell for $2.50. The only real winner is the Staples Center.

This is exactly what's happening to Oracle's database business. Microsoft and IBM are catching up, but so is a free database, PostgreSQL. So Oracle has had to drop the price of their low-end "Standard Edition" and add features to their high-end "Enterprise Edition". For technical reasons, only PostgreSQL is ever likely to match the scalability of Oracle, but as computer hardware gets faster and other databases find ways to work around their limitations, this will matter less and less. Eventually customers will decide that cheaper software is "good enough".

The case for Oracle


  1. The main advantage is Oracle's multi-version concurrency model. For a good explanation, see Expert One on One Oracle by Thomas Kyte.

Jon Ericson
Last modified: Tue May 20 14:11:30 PDT 2003