The 'other' billionaire
Oracle's Larry Ellison revealed
Mike Wilson's The Difference Between God and Larry Ellison: Inside Oracle Corporation provides an in-depth, entertaining look at software's "other" billionaire -- his flamboyant, complex personality; his scrapes with women, yachts, and the law; and his success in building one of the most important software franchises in the world. (2,400 words)
Q: What's the difference between God and Larry Ellison?
A: God doesn't think He's Larry Ellison.
Mike Wilson makes the mistake of giving away the punch line on the book's cover. To my knowledge, this is the only Oracle joke that has reached the level of common parlance in the high-tech industry -- compared with dozens or even hundreds of jokes about Ellison's archrival, Bill Gates. As readers of The Difference Between God and Larry Ellison: Inside Oracle Corporation will gather, the joke's shortfall must cause Ellison considerable pain: Ellison thrives on public exposure and is intensely jealous of Gates's higher profile. Therefore, even though this book is far from entirely complimentary toward its subject, Ellison must be glad it made it to print. That he agreed to let the publisher use his photo on the cover is further evidence that he prefers this kind of exposure to no exposure at all.
As this book shows, there are some superficial similarities between Gates and Ellison, although they are two very different sorts of people. They are, respectively, the number one and number two software billionaires in the world right now. They both got their big breaks as a result of IBM's inability to capitalize on a certain huge opportunity in the marketplace. For Gates, it was Big Blue's decision to award Microsoft a contract to license MS-DOS as the operating system for the original IBM Personal Computer. For Ellison, IBM's missed opportunity was the now-standard relational database model and its associated programming language, SQL.
Bringing SQL to market
Edward (Ted) Codd, an IBM scientist, invented the relational database concept as an alternative to the two existing database models IBM had implemented for its mainframes: the hierarchical and network models. As every database programmer knows, the relational model is a marvel of elegant simplicity. By grouping data into tables, it lets users find answers to complex questions using simple commands rather than long, complex programs written in COBOL or some other language. Codd's intent with the relational model was to define a database programming language that even nontechnical end users could use. He didn't quite achieve that goal, but SQL endures today as one of the most important technologies in the software field.
One especially important attribute of the relational model and the SQL programming language is that it was intended to be portable -- to run on any type of machine, not just mainframes derived from IBM's System 360. Larry Ellison and his partner Bob Miner decided that they could implement the relational model for the various minicomputers, such as Digital Equipment Corp.'s VAX, that were becoming popular in the late 1970s. They founded the company Relational Systems Inc. (RSI) to do this in connection with some consulting work they were doing for Ampex, the recording tape company. They used information from Codd's published papers to build their database. They called it Oracle, and they beat IBM to market with a finished product.
Ellison and Miner weren't the only ones to upstage IBM with a relational database. Michael Stonebraker, a professor at the University of California at Berkeley, had implemented the relational model for the Unix operating system (which was only popular in research labs at the time). He started the company Relational Technologies Inc. (RTI) to take his database into the commercial realm; he called his product Ingres. Stonebraker and Ellison became rivals years before Bill Gates and Microsoft entered the picture. They both changed their company names to those of their products to avoid confusion in the marketplace.
Ingres is now a footnote in computing history. After falling behind not only Oracle but also Sybase and Informix in the relational database market, the company was bought by Ask Group, and subsequently by that carrion feeder of the software industry, Computer Associates. As this book tells it, the reason Ingres failed where Oracle succeeded was that whereas Stonebraker concentrated on excellence in engineering -- most database experts agree that Ingres was the superior product -- Ellison prevailed through flamboyant, aggressive salesmanship and marketing. Stonebraker went on to develop Postgres, an enhanced relational database model that could be used to implement object-oriented databases. He implemented Postgres in a commercial database called Montage, which was renamed Illustra, bought by Informix, and re-renamed Informix Universal Server. Oracle then shamelessly copied Universal Server's key technology and proceeded to overshadow Informix. Informix is now struggling in the marketplace.
Stonebraker was perhaps as colorful a character as Ellison, but his notoriety never traveled much beyond the world of database geeks. Ellison, in contrast, craved exposure in the world at large.
In fact, the most important theme of this book is Ellison's (and Oracle's) grandiosity. There is a quality about Larry Ellison that, for lack of a better term, I will call the art of constructive fuzziness. It is a quality that has allowed him to be outrageously successful in the fast-moving information technology business -- and one that would lead to failure in most other fields.
The computer field is a curious dichotomy of solid and fuzzy. The solid part has been there from the beginning: it consists of things like engineering, materials, precision, processes, skill, science, and facts. But at some point, fuzziness began to creep in -- things like gut instinct, creativity, elusiveness, hype, rumor, FUD (fear, uncertainty, and doubt), and wishful thinking. Larry Ellison was one of the first people in the computer business to give fuzziness the prominent role in the way his company worked and how it dealt with its customers. Thus, he changed the industry forever. Ellison knew instinctively that in a world where everything changes constantly and there's far too much information to assimilate, much less validate as fact, the right kind of fuzziness is a serious competitive advantage.
a world where everything changes
constantly, the right kind of fuzziness
is a serious competitive advantage.
After Oracle's initial release, Ellison developed fuzzy tactics that are now standard elements of the computing scene but certainly weren't then: releasing software that wasn't fully debugged, preannouncing nonimplemented features, claiming software portability that didn't really exist, and so on. Bill Gates does this all the time at Microsoft (as do many other vendors nowadays), but somehow it feels different. When Gates does it, you get the feeling that he knows exactly what pronouncements to make to manipulate the market; he seems like the omniscient spider at the center of a great web, like Professor Moriarty in the Sherlock Holmes stories. On the other hand, when Larry Ellison promises a customer that a certain feature will be available in the next release, it's more off-the-cuff and opportunistic, like a used-car salesman saying, "Yeah, yeah, don't worry, we'll take care of that engine gurgle by the time you come in with a check."
One episode in The Difference Between God and Larry Ellison neatly encapsulates the differences between Gates and Ellison: when the two of them had their first face-to-face meeting in Ellison's sumptuous Japanese-style home, Ellison mostly talked, while Gates mostly listened. You figure out who got the better of the conversation. Now guess who claimed to have gotten the better of the conversation.
One of Ellison's most important talents is that he's smart enough to know just how much fuzziness the public will swallow. The Difference Between God and Larry Ellison is full of examples of his ability to manipulate facts and stretch truth. One of my favorites is an anecdote about Ellison as a college-age kid in Chicago, driving around with some friends and being stopped for speeding. When the policeman pulled him over, he fabricated a story about being a medical student rushing to the hospital to witness an operation. By the time he was done, the cop was offering to give him an escort.
The combination of Ellison's fuzziness and his engineers' more solid skills enabled Oracle to reach, in the mid-'80s, what technology marketing guru Geoffrey Moore calls the "tornado" stage of the market. Relational databases became a gotta-have-it technology, and growth exploded for all of the leading vendors: Oracle, Sybase, Informix, and others. Oracle emerged as the market leader, largely because it had versions that ran on all types of computers, including mainframes and PCs. So what if the product was buggy, slow, or didn't quite run on one particular type of machine. Larry was always there to say it would be fixed in the next release, and Oracle's growth continued unabated.
Pushing the sales force
Another Ellison innovation that became widely imitated across the high-tech industry was his heavy emphasis on the company's salesforce. Ellison's sales managers drove their people hard to achieve dizzily escalating sales goals, yet gave them lots of latitude to do whatever it took to bring in deals, and compensated them highly for their efforts. As a result, the salesforce essentially ran the company. You can see the same strategy at work nowadays in fast-growing upstarts like storage vendor EMC and salesforce-automation software maker Siebel Systems (whose CEO, Tom Siebel, is a former Oracle sales executive), as well as large, established companies like Sun.
people hard to achieve dizzily escalating
sales goals, yet gave them lots of latitude
to do whatever it took to bring in deals.
For a while, Oracle's aggressive sales efforts achieved incredible annual growth rates, often above 100 percent per year. But when growth finally began to slow in the early '90s, the strategy began to backfire. Oracle salespeople, who were getting more and more desperate to claim revenue, started relying on more and more dubious techniques -- like drawing up sales contracts with such dubious customer parameters as "I agree to buy Oracle licenses, sometime in the future, if I feel like it." Oracle's financial chiefs tried to stop this spurious practice, but Ellison wouldn't let them. Finally, in 1993, the bubble burst as the Securities and Exchange Commission came down hard on the company's accounting irregularities. Oracle had to shift gears, adjusting to more "normal" growth rates, becoming a fiscally responsible company, and taking steps toward figuring out where its next big growth was going to come from.
Figuring out the next growth phase was difficult. Oracle's core database engine is by far its most important asset, and every other product Oracle has produced has been ancillary to it. In other words (and apart from its successful line of financial applications), Oracle is essentially a one-product company. For several years, Oracle has faced the risk of its entire business going down the drain once the market for relational databases begins its decline.
To combat this risk, Ellison tried venturing out into territory beyond relational databases. As this book tells it, he ended up with a string of failures. The first was his purchase of NCube, a maker of massively parallel computers. Next was his attempt to get into interactive television by partnering with Apple on set-top boxes and repositioning NCube as a digital video server platform. Another was an attempt to evolve Oracle into an object-oriented database; this was during a time when many people thought that object-oriented databases like Versant and ObjectStore were going to displace relational systems. Finally, Larry Ellison decided that the Internet was going to be his next growth phase. As the book closes, he's declaring all-out war on Bill Gates and ramping up his Network Computer (NC) business. We now know that NCs haven't been big success either.
We also know that Larry finally did find Internet success by returning to his core competency, relational databases. With Oracle 8, Oracle introduced the Data Cartridge concept, a blatant rip-off of Informix/Illustra's Data Blade object-oriented API technology that positioned Oracle well for multimedia development. Then Ellison introduced Oracle 8i, also called InterMedia, an enhancement of Oracle 8 repositioned to be an Internet and multimedia server software engine, and announced a database/server partnership with Sun that eschews the need for a traditional operating system. Oracle's salespeople started flogging 8i hard, and its developer relations people have been equally aggressive in recruiting software vendors to write applications for it. As a result, Oracle 8i is flying off the shelves, and it appears that relational database technology remains uncontested as the foundation for complex Web sites.
The Difference Between God and Larry Ellison is full of information about life inside Oracle, including Bob Miner, Ellison's publicity-shunning technologist and partner, various whip-cracking sales managers, beleaguered finance people, and Ellison's myriad girlfriends -- many of whom were also Oracle employees. It is rife with tales of his romances, extravagant homes and cars, yachting exploits, lawsuits, and other bits of derring-do -- all of which Ellison is able to rationalize with his characteristically glib fuzziness. The book offers deep insight into a complex character, and it does so through clear writing and good storytelling.
Title: The Difference between God and Larry Ellison: Inside Oracle Corporation
Author: Mike Wilson
Publisher: William Morrow & Company
List price: $25.00
About the author
Bill Rosenblatt is vice president of technology and new media for publication services at The McGraw-Hill Companies.
If you have technical problems with this magazine, contact firstname.lastname@example.org