The Difference Between God and Larry Ellison

Mike Wilson · 1997

Thesis

Oracle became one of the largest software companies in the world less through technical superiority than through the will, salesmanship and self-mythologizing of its founder, Larry Ellison, whose habit of selling what did not yet exist drove the company's growth and nearly destroyed it.

The main ideas

Ellison built Oracle by seeing the commercial value of a database idea that IBM researchers had published and IBM was slow to sell.
Oracle's early success came from aggressive selling, often of features not yet built.
The same sales culture produced the 1990 accounting crisis, in which revenue was booked before it was earned.
The company survived by bringing in managers who imposed financial discipline that Ellison had not.
Ellison's public story about himself was partly invented, and a founder's myth can shape a company as much as its products.
A company built around one personality takes on that personality's strengths and flaws.

The joke

The book's title comes from an old Silicon Valley joke. What is the difference between God and Larry Ellison? God doesn't think he's Larry Ellison. A later paperback edition put the punchline on the cover.

Mike Wilson was a reporter at the St. Petersburg Times when he wrote the book, published in 1997. He had earlier worked at the Miami Herald, and later became managing editor of the Tampa Bay Times and editor of The Dallas Morning News. Ellison agreed to be interviewed, and Wilson spoke to more than a hundred of Ellison's friends, enemies and former employees. The book is not authorized; Ellison had no control over its contents.

Publishers Weekly found that Wilson did a solid job of puncturing the myths surrounding Ellison, who was known for embellishing his own story.

Implication: when a founder's personal legend is central to a company's image, check it against the accounts of people who worked with them.

Founding Oracle

Ellison was born in New York in 1944 to an unmarried teenage mother and adopted as an infant by relatives in Chicago. He attended college without graduating and worked as a programmer, including at Ampex, where he met Bob Miner and Ed Oates.

In 1977 the three founded Software Development Laboratories. Their first major client was the Central Intelligence Agency, on a project code-named Oracle, which gave the product its name.

The key idea was not theirs. IBM researchers, building on E. F. Codd's relational model of data, had published work on a relational database and a query language, SQL. IBM was slow to turn the research into a product. Ellison and his partners built a commercial version, released in 1979 and usually credited as the first commercial SQL relational database. They called it Version 2, because, the story goes, customers distrusted first versions.

The mechanism is that a large company's research can be commercialized faster by a small, hungry one, which then sells it back to the market the large company should have owned.

Implication: watch for valuable ideas that large institutions publish but are slow to sell, and remember that being first to market can matter more than inventing.

Selling the future

Oracle grew very fast through the 1980s. Its sales force was famously aggressive, and the company often sold features before they existed, a practice critics called vaporware. In Wilson's account, the company's growth owed more to marketing and sales than to the quality of its software.

Wilson presents Ellison as a brilliant salesman and marketer, competitive to an extreme, who treated business as war and his rivals as enemies. He ran attack advertisements against competitors and drove his staff hard.

The mechanism is that in a new market, the company that persuades customers first can win even with a weaker product, because customers commit to a platform and rarely switch.

Implication: in a new market, expect the most aggressive seller to win early, but judge it by whether its product eventually catches up with its promises.

The 1990 crisis

The sales culture caught up with Oracle in 1990. Salespeople, paid on bookings, had been recording future license revenue in the current quarter to meet targets and raise their bonuses. The company had to admit that its revenue and earnings were overstated and restate its results. Its stock collapsed and shareholders sued; the class action was settled for $23.25 million.

Oracle recovered by changing how it was run. Jeff Henley became chief financial officer in 1991 and imposed financial controls, and professional managers took on more of the operating work. Within a few years the company was growing again.

The mechanism is that incentives that reward bookings over real revenue invite people to stretch the rules, and a culture that celebrates winning at any cost will not stop them.

Implication: when a company's growth depends on aggressive sales targets, look closely at how and when it recognizes revenue.

Limits

It ends in 1997. The book predates the dot-com boom and bust and everything Oracle and Ellison did afterward, so it describes the company's adolescence rather than its mature form.
Its sources lean against Ellison. It relies heavily on former employees and rivals, and its skeptical tone may undersell Oracle's genuine technical achievements.
It is a personality study. The title's joke sets the frame, and readers looking for a business or technical history of the database industry will find less of one.

The Hard Thing About Hard Things — another account of a software company nearly destroyed and rebuilt.
Zero to One — the case for building a company that dominates its market.
Dot.con — the boom that followed, in which promising more than you could deliver became an industry.