Softwar

by Matthew Symonds

  • • •

  A week later, at Oracle’s OpenWorld event in Berlin’s huge and hideous ICC conference center, Ellison does a repeat of his 9i launch presentation. This time, however, the seven-thousand-strong audience is largely made up of enthusiastic Oracle partners and database administrators. The professionally skeptical analysts and reporters have their chance to question Ellison in a side room, but it’s a fairly tame affair despite the presence of one or two known Oracle scourges, including Gartner’s Betsy Burton.

  Burton has suggested in recent interviews that Oracle’s database is losing ground because of intense pricing pressure from IBM and Microsoft, even though the evidence, and even the market research data—itself of questionable quality—is ambiguous. While it’s taking flak over the early bugginess of 11i, the last thing Oracle needs is for the biggest technology research and consulting house to be suggesting that there are chinks in the armor of its core database product. In fact, in the course of Oracle’s fourth-quarter earnings call earlier that week, Ellison had to admit to a 5 percent falloff in database revenues. While the financial analysts took comfort from the fact that Oracle’s earnings had held firm, industry watchers such as Burton took it as further evidence that Oracle’s database business was coming under pressure.5

  The next day we’re in London at his favorite hotel, the Lanesborough. After yet another CEO roundtable, Ellison is going to meet with a group of industry analysts from different firms, among them the dreaded Betsy Burton and two or three of her colleagues from Gartner. Although Mark Jarvis has been trying to mend fences with Burton, Ellison is more inclined to be confrontational. “Like a lot of these analysts,” he says, “all she does is predict the past.”6

  About fifteen analysts from firms such as IDC, Forrester, AMR, and, of course, Gartner, are seated around a long table. Their demeanor is mostly earnest and respectful, although one, Paul Mason, a senior analyst from IDC, soon starts to rub Ellison the wrong way with a fatal mix of overfamiliarity and sarcasm. Ellison starts off by ramming home his message about data fragmentation: “It’s the issue. It’s that simple.” What’s unique about Oracle, he argues, is its combination of a datacentric information architecture and the ease with which the E-Business Suite can be extended because Oracle uses Java rather than the proprietary tools of its rival applications vendors. The key difference between Oracle and the others—its counterpoint—is information management. Ellison says that he doesn’t believe in separating data warehouses from transaction processing because too much time is lost. The combination of RAC and another feature on 9i that automatically maintains both old and new versions of the same data would make warehousing irrelevant. Burton volunteers that one of the benefits of this is that it “reduces conflicts between two separate instances.” “Absolutely,” purrs Ellison.

  Ellison feels strongly that the analysts haven’t yet given Oracle enough credit for RAC because they are hostile to real innovation. “It’s like Galileo and the pope. Every time he had a new idea, the pope threatened to kill him.” Instead of talking up IBM’s DB2, why not point out that with its shared-nothing architecture all it can do is share nothing really fast? “Drill down into this,” he demands. “Write this down. Ask IBM why its UNIX database takes a different architectural approach from its mainframe versions. Please write this down. Why won’t IBM demonstrate the speed of their hardware with their own database? You guys always say we’re more expensive than IBM. But it’s the cost of ownership that matters—the networking, the hardware, the labor. The software is always a tiny proportion of that. I’m almost done. This is a scam, and you guys have all fallen for it.”

  Having let off steam, Ellison answers questions on everything from peer-to-peer computing (which he dismisses as a “cute little technology that’s good for exchanging music” but not something for the enterprise) and software as a service. Ellison says that in a very few years, the bulk of Oracle’s business “will be on servers we manage.” But the conventional ASPs (application service providers) were all wrong. Where the computers were located was irrelevant; what mattered was who provided the skilled labor. “It doesn’t matter where the computer is. As long as it’s running in a standard certified configuration, you can put hardware wherever you want and we’ll run it for you. And another thing: everyone focuses on our new license revenue, but the real value in our business is recurring revenues from license updates and support.” IDC’s Mason suggests that this is typical Ellisonian arrogance. Everyone’s wrong except him. Ellison responds, “Sure. If you’re the first to say something or do something that conflicts with popular conventional wisdom, then what you’re saying is, in effect, ‘I’m right and everyone else is wrong.’ That’s what innovation means—being first. People never forgive you for being right too early.7

  Ellison had been forceful, funny, and brilliantly controversial, but I didn’t think he had necessarily won them over. He was more than capable of charming them; but it was as if he was too fastidious to try. Eight weeks later, Betsy Burton produced a report for Gartner entitled “Oracle Under Fire,” which began with this broadside: “During the past 12 months, Oracle has been under increasing scrutiny from many of its customers, the business and technology press, investment firms, and Gartner. Some of the press attention is because of Oracle’s sheer size, its impact on U.S. financial market indicators, and its 60 percent and larger stock price slippage in the past twelve months. Some is because of Oracle’s flamboyant and contrarian CEO, Larry Ellison. Much of the attention, however, is because of real product, marketing and business issues.”

  The report itself was a mixture of the obvious and the out of date. Oracle, Burton said, would have to: provide “proof points” for 9i’s “clustering scalability” and ability to “support very large databases”; regain “customer trust” after a U-turn on power unit pricing; “continue a sustained level of commitment to its application server”; resolve the quality issues with some of its ERP components, such as order management; deliver CRM product parity with Siebel; make sure that Oracle consulting and Oracle.com (its ASP offering) complemented its other channel relationships. Burton pointed out that she was far from suggesting that “Oracle, as a company was failing—quite the opposite,” but the overall impression the report gave was highly negative, thanks in part to its last line: “We believe that recent actions and events will leave an indelible scar on customers’ views of Oracle.”8

  Around the time that Burton’s report came out, Oracle hired one of her Gartner colleagues, Peggy O’Neill, to take charge of analyst relations. I asked O’Neill what Oracle needed to do to get on better with such an influential constituency. She was in no doubt: “We fundamentally need to change the relationship with the analysts. We must talk to them more. . . . Don’t broadcast to them. Listen to their advice and, if at all possible, incorporate it in your planning, marketing, and development. Historically, Oracle has never had that kind of interaction.” I told her that I found it hard to imagine Ellison, whom she had yet to meet, taking that advice to heart. “The analysts are prepared to give Larry a bit more room because he’s Larry,” she said. “But the folks underneath Larry who copy that style won’t get away with it. I’ve seen them with customers: they know how to be charming. They should try treating analysts like customers.”

  O’Neill was almost certainly right. But I found it hard to imagine Oracle schmoozing analysts and journalists with the polish and attention of, say, an IBM or a Cisco Systems. Technology companies that are run by engineers, such as Sun Microsystems, Microsoft, SAP, and Oracle are characterized by a particular kind of intellectual competitiveness that is highly unreceptive to the opinions of outsiders, who they think (usually correctly) are not as smart as they are. As long as they execute flawlessly, it’s not a problem. But when something goes wrong, they can’t expect any slack. O’Neill made another point: analysts, she said, “are a jaded lot.” She might as well have said that analysts and technology journalists are professional cynics and that someone l
ike Larry Ellison, who is relentlessly optimistic, calculatedly contrarian, and shamelessly devoted to hyperbole, brings out the absolute worst in them.

  * * *

  1. LE writes: Simply stated, shared-nothing clusters get less reliable as you add computers to the cluster. An eight-machine shared-nothing cluster is one fourth as reliable as a two-machine shared-nothing cluster. Hard to believe, but true.

  2. LE writes: Right. We should have held the pricing change until the following week. It was really dumb to mix it with the RAC announcement.

  3. LE writes: San Jose Airport had a bizarre noise abatement program based on the weight of an airplane rather than the noise it made. My quiet plane was banned from landing after 11 P.M. because it was too “heavy.” Noisier planes were allowed to land anytime because they were “lighter” than my airplane. Call me crazy, but I thought that noise ordinances had to do with restricting noisy airplanes, not heavy airplanes. The judge agreed. But in the meantime the press wrote lots of nasty articles about my not obeying the rules. It was quite the little local controversy.

  4. LE writes: The Linux operating system running on Intel servers is rapidly on its way to becoming the most popular hardware/software combination in large data centers. This presents a great opportunity for Oracle because our unique clustering technology lets us deliver our high-performance, fault-tolerant database on low-cost Intel servers running Linux. We can’t continue to let Microsoft be the only software company to exploit low-cost Intel hardware.

  5. LE writes: Every survey that asked customers what database they were using or planning on using showed Oracle gaining share against IBM. A Morgan Stanley survey had Oracle six times more popular than IBM’s DB2 under enterprise applications like SAP and Siebel.

  6. LE writes: The most respected analyst in our industry is Chuck Phillips. When Chuck publishes his survey results, he includes all of his supporting data and his methodology. Anyone can independently verify the data and double-check his calculations and conclusions. Chuck is completely devoted to the scientific method, invented by Sir Francis Bacon more than four hundred years ago. Betsy Burton uses a more modern approach: she publishes her results without any supporting data or details about her methodology. You can’t check a damn thing. You just have to take her word for it.

  7. LE writes: If you speak out in support of small, unimportant innovations that fly in the face of widely held beliefs—I do it all the time—you are likely to be dismissed as stupid or arrogant, and that’s pretty much the end of it. However, if you defend a really big idea that challenges widely held beliefs, you’re likely to generate a mass of hatred, and you just might pay for it with your life. When Galileo defended Copernicus, he was ridiculed, imprisoned, and then threatened with death unless he recanted. Charles Darwin cautiously postponed publishing On the Origin of Species and The Descent of Man for more than twenty years, but that judicious delay did not save him from vicious personal attacks coming from all ranks of contemporary society, from the dons at Oxford to the man in the street. In pleasant contrast, big discoveries that don’t come into conflict with popularly held beliefs are quickly embraced and the discoverers are rewarded with love and fame. Such was the good fortune of Jonas Salk.

  8. LE writes: What most analysts say about your company has more to do with the financial results of your last couple of quarters than anything else. If you’re growing revenue and earnings, they usually say nice things about your products and management. As Tevye says in Fiddler on the Roof, “When you got the money, they think you know.” But our previous two quarters had not been good, so we got hammered.

  15

  ENEMIES

  A subject that’s close to Ellison’s heart is Oracle’s enemies. He strongly believes that Oracle is always at its best when it has an identifiable enemy to go after: “We pick our enemies very carefully. It helps us focus. We can’t explain what we do unless we compare it to someone else who does it differently. We don’t know if we’re gaining or losing unless we constantly compare ourselves to the competition.” When Oracle was fighting its relational database rivals for market supremacy in the late 1980s and early 1990s, it was famous for the in-your-face aggression of its “attack” advertising.

  The aggression hasn’t altered, but the size and power of the firms with which Ellison wants Oracle to be compared have. These days, none other than IBM finds itself the regular target of Oracle’s ads and Ellison’s combative speeches, while not a little of Ellison’s own fame has come directly from his highly public assaults on Microsoft and his obsession with one day overtaking the colossus of Redmond to make Oracle the number one software company in the world. When people think of Ellison, it’s all too often as a kind of alter ego to Bill Gates, software’s other billionaire.1

  An odd effect of this was to diminish Oracle’s own extraordinary success. Surely it was better to be known as the world’s biggest enterprise software firm than to be seen as Microsoft’s perennial challenger. As for Ellison’s attacks on Gates, they could make him seem “chippy” and resentful, both of which were far from the truth. What made it even stranger was the fact that Microsoft and Oracle compete only at the margins. To be sure, Microsoft has a database product, originally licensed from Oracle’s old rival, Sybase, more than a decade ago. But despite attempts by Microsoft to present the latest versions of its SQL Server as being sufficiently capable for data center duty, its deployment is still mostly departmental. Although Microsoft would like nothing better than to destroy Oracle’s profitability by commoditizing the database business, the demands of Internet computing have so far thwarted that ambition. SQL Server remains essentially a “good enough” database that’s bundled into Microsoft’s BackOffice server suite to put some price pressure on Oracle at the low end of the market. As for applications, Microsoft has remained content to dominate the desktop with Office, preferring to partner with Oracle’s competitors, such as SAP, than to compete with Oracle head-on.

  However, from Ellison’s perspective, the assault on Microsoft and all its works, which he initiated in September 1995 when deriding the PC as “a ridiculous device,” had not only evolved into the much broader war on complexity, but created an awareness both of Oracle and of its vision of computing that nothing else could have achieved. Since then, although Microsoft’s wealth had grown almost exponentially thanks to its near-monopoly profits from Windows and Office, it no longer had quite the aura of invincibility it had previously enjoyed. Thanks to the Internet, computing had moved decisively toward a model that played much more to Oracle’s strengths than to Microsoft’s. As for the antitrust case against Microsoft that had arisen from its brutal suppression of Netscape, it had not only hugely distracted its senior management but done great damage to the company’s reputation. A few days after, I discussed these issues with Ellison. The Washington, D.C., Court of Appeals found Microsoft guilty of serially abusing its monopoly power while rejecting the controversial remedy of District Court Judge Thomas Penfield Jackson that it should be broken into two companies.

  Ellison’s antipathy toward Microsoft seemed to go much further than simply seeing it as a dangerous business adversary. He once said to me that what he really didn’t like about Microsoft was that it didn’t have any taste. What did he mean? “Well, actually I was quoting Steve Jobs. He said that the thing that really bothers him most about Microsoft is not how successful they are or how much money they have; it’s the tasteless mediocrity of their software.2

  “I totally agree with Steve, Microsoft’s software is rarely first rate. They never, ever innovate, but they’re pretty good copiers. All those bright people up in Redmond remind me of the guys you see sitting in museums making beautiful copies of great art. Their pictures are beautiful, but they’re copies—forgeries. Ever since the guilty verdict in the antitrust case, Bill began chanting Microsoft’s new mantra: ‘Please, please don’t take away our right to innovate.’ Microsoft innovate! Give me a fucking break. The innovation for the Internet came from Netscape. All
Microsoft did was copy Netscape’s browser and bundle a ‘free’ copy of the browser with Windows. But it wasn’t really ‘free’ at all. Microsoft got paid for its ‘free’ browser by raising the price of Windows. But Netscape couldn’t charge for its browser because Microsoft included a ‘free’ browser as a part of Windows. It was a brilliant business strategy. It killed Netscape. But it’s illegal. Now Microsoft is trying to do the exact same thing to RealNetworks [the innovators of Internet streamed video and audio] by bundling a ‘free’ media player with Windows. They’ll just keep doing it over and over again until somebody penalizes them for doing it. You’re got to give them credit for balls, but not for innovation. Even Bill’s business strategy is just a copy of Standard Oil’s strategy back in the 1870s. But when Rockefeller used his monopoly to crush his competitors, it wasn’t illegal. There were no antitrust laws back then.

  “So what’s Microsoft’s single greatest innovation? Take your time. It’s a trick question. There aren’t any. All that money Microsoft spends on research; what have they got to show for it? Nothing! Compare that to IBM’s research results. IBM invented the disk drive, they invented core memory, they invented the scanning tunneling microscope, they invented fractal geometry. The list of IBM inventions goes on and on. IBM researchers have won Nobel Prizes. IBM, at the height of their greatness, was a national treasure, an institution that anyone would be proud to be a part of. They don’t do software very well anymore, but their old mainframe stuff was great. I make fun of a lot of other databases—all other databases, in fact, except the mainframe version of DB2. It’s a first-rate piece of technology. Microsoft may have more money than IBM ever did, but they don’t have more ideas. It’s the difference between a great fashion designer in Paris and someone who just does knockoffs in Brooklyn. Except that Microsoft would bundle the dresses with Windows and give them away for ‘free.’ ”3