In an article for the San Francisco Chronicle three weeks earlier, Sjoberg had spelled out those differences and why Howle’s calculations made little sense. Among Howle’s mistakes, she had chosen fiscal 2000 as the baseline year—a period covering nine months of calendar 1999, a time when public sector IT budgets had been almost wholly dedicated to preventing millennium bug meltdown. Ellison says, “The two audit reports came to diametrically opposite conclusions; therefore, one of the two auditors must have been wildly mistaken in their analysis and calculation of savings. Sjoberg had about twenty years’ experience as state auditor, while Howle was brand new to the job. It was obvious to anyone who bothered to read both reports which one was right.” Sjoberg’s report was casually and rather disgracefully dismissed by Florez (“After sitting through nine hours of testimony, my conclusion is that we stand by the state auditor’s report. You get the numbers you pay for, and Oracle got those numbers.”), and he was equally unimpressed by the evidence from Kevin Fitzgerald about Oracle’s minimal involvement in drawing up the contract. Other members of the committee, however, had the decency to look a little uncomfortable. Ellison observed, “When the facts finally came out, Florez found it hard to completely ignore them—but he tried.”
When Dean Florez finally called it a day on June 17 after two months, thirty witnesses, and more than 2,500 pages of documents, there was still no smoking gun. Most critically, nobody had been able to make any connection between the contract and “the buying of lawmaker influence,” as one headline had optimistically put it. Florez concluded, “I think we got taken to the cleaners. If we have accomplished anything, I hope it is to warn state officials that the next time a smelly deal crosses your desk at the eleventh hour, you’d better do what’s right instead of what’s politically convenient.” But he tacitly admitted failure when he decided to close the investigation without producing a report. John Borland of CNET, the San Francisco–based technology news Web site, wrote: “Contradictory testimony throughout the trial has failed to provide an accurate picture of how the deal was finalized or structured. It’s also unclear whether Oracle or Logicon is to blame for the convoluted contract—or whether state officials were too trusting of the promised cost savings and license estimates in the first place.” Or if Elaine Howle had simply gotten her sums wrong, he could also have added.
In the aftermath, there was a conspicuous attempt to mend fences. On July 23, Oracle’s offer to rescind the contract was gratefully accepted by the state’s attorney general, Bill Lockyer: “Today’s agreement serves California taxpayers and successfully meets the state’s objectives by rescinding a $95 million contract that had become controversial. Oracle and Northrop Grumman [the parent company of Logicon] were under no obligation to help California by rescinding this contract, and we appreciate their cooperation and willingness to rescind the ELA.”
Three weeks after the closure of his committee’s hearings, Dean Florez was stripped of his chairmanship. It was, of course, an act of vengeance by supporters of Governor Davis, but Florez had won few admirers by his relentlessly egotistical behavior. The Sacramento News and Review editorialized, “Let’s talk about what a legislator ought not to be: more about self-interest than public interest, arrogant, grandstanding, difficult for both friends and enemies to deal with. This describes Dean Florez to a T. There are few people in the Capitol with anything nice to say about the Democrat from Shafter, except maybe for Republicans who cheered his over-the-top investigation of the Oracle scandal, and even they grumbled about his autocratic tendencies during the hearings.”
• • •
How much real damage had the whole messy business done to Oracle? It wasn’t as if its “perfect storm” had come out of a cloudless sky. The company was only just beginning to recover from the damage caused by the birth pangs of the E-Business Suite and was continuing to suffer slack demand for enterprise software. Also, more or less coincident with the California imbroglio, the market research arm of Gartner delivered what looked like another hammer blow: survey numbers suggesting that IBM’s DB2 had pushed Oracle into second place in the high-end database market for the first time in years—a finding gleefully leapt upon by the tech press and a good many analysts who ought to have known better.
Morgan Stanley’s Chuck Phillips was pretty much a lone voice in putting the apparently sensational news into context. He wrote, “Gartner Dataquest’s numbers largely reflect whatever Microsoft and IBM provide to the market analysts as revenue for the past year’s performance. There is no way to verify the numbers since they aren’t audited and aren’t provided to financial analysts. IBM allocates database revenue out of deals that include services, hardware, and even mainframe maintenance. We don’t think IBM gained in the last year although they did in the previous year when Oracle was too focused on dotcoms and raised prices too aggressively but even then it was a small gain. We tend to look to actual usage since that’s a more realistic measure of what customers believe they are buying in; by that measure Oracle is still around 50% share, Microsoft 25%, and IBM 34%. [The numbers total more than 100 percent because some companies use more than one database.] This is a numbers and marketing game between Microsoft, IBM, and Oracle that has gone on for years. Oracle has no flexibility since its numbers must be reported externally and audited. It has two large competitors that can allocate from other product areas. The game helps Microsoft and IBM, which is why they keep playing it and there is no downside for them since journalists and market analysts pretty much print the numbers they are provided.”5
However, just as more analytical reporting of the “Oracle scandal” in California would have “spoiled the story,” so would listening to the punctilious and experienced Phillips. But for all the media’s pleasure in using these misfortunes to paint a picture of a company in deep trouble, apart from the flagging stock price, the actual harm to the business was probably far less than met the eye. Real-world customers were largely uninfluenced by dubious market-share figures—most of the growth in Oracle’s database business is founded on selling more licenses to its installed base. For the same reason, although some analysts speculated that the adverse publicity in California would hurt Oracle’s vital public-sector sales, there was subsequently no evidence of it. Ellison said, “Several large deals in the quarter got slowed down; they were delayed, but they didn’t disappear. In the end, if you need the software, you buy the software.” That would also apply to California: Ellison had no doubt at all that the state would end up paying more for less software because of canceling the ELA. In principle, he never wanted to do another deal like that one: “These big up-front all-you-can-eat deals are just not as profitable for us as selling software on an as-needed basis.”
The whole episode, however, made Ellison deeply miserable. For somebody not temperamentally given to depression, he came close to it. Above all, the sheer capriciousness of events combined with the mounting hostility toward big business had shaken him. When we met in London late in June, I was struck by how gloomy he seemed. He told me, “I’m not depressed, exactly. I’m just tired, and I feel vulnerable—like a zebra on the savanna at night. There are a lot of hungry animals out there hunting. They’d like to eat me and my family. I have to be alert, and I have to stay alert all night. And this is going to be a very, very long night. When the sun rises, a lot of zebras are going to be gone.”
• • •
The California brouhaha had upset Ellison deeply. Oracle had finished its fiscal year quite strongly, beating analysts’ expectations with fourth-quarter revenues of $2.774 billion combined with profits and margins exceeded in the software industry only by Microsoft. But the solid showing was rather undermined by Jeff Henley’s downbeat forecast. The market took fright at his view that first-quarter license revenue could be anything between 15 and 25 percent down, sending the company’s already sagging stock price to little more than $8. Even prompting Ellison to talk about his prospects in the America’s Cup—just three months away—didn’t seem
to lighten his mood.
Yet only a few weeks later Ellison gave what Carolyn Balkenhol, among others, thought was one of his feistiest and most compelling performances at Oracle’s Analyst Day. Nearly three hundred analysts and reporters had shown up at Oracle’s conference center expecting to put Ellison on the rack. Typically, he came out fighting with plenty of jokes and a major product announcement. The big news was the launch of what Oracle was calling its Collaboration Suite—a software package that combined e-mail, voice mail, calendaring, scheduling, file sharing, and an enterprisewide search engine—all accessible by wireless. But this wasn’t just any old product announcement. What really juiced it up was that the Collaboration Suite was targeted at a major Microsoft business. For all the rivalry between the world’s two biggest software firms over architectural issues (the Internet versus client/server and, more recently, .NET versus the J2EE Java standard) and Microsoft’s persistent harrying of the low end of Oracle’s business with its SQL Server database, this was the first time that Oracle had gone directly after a big chunk of Redmond’s revenues.
Since the previous summer, Ellison had been touting the advantages of running e-mail on top of Oracle’s database rather than employing the hundreds of Microsoft Exchange servers that were common in most big companies. Keep Microsoft’s Outlook as your desktop e-mail interface, he’d been saying, but replace Microsoft Exchange with Oracle’s e-mail server and get the virus protection, security, reliability, and ease of administration that come with our database. In addition to e-mail, Collaboration Suite included a new technology called the Database File System: an extensible file system for storing, sharing, and searching all your PC files regardless of format. In other words, Oracle’s Collaboration Suite wasn’t just a replacement for Microsoft’s Exchange servers; it was a replacement for Microsoft’s file servers as well.
A year earlier I’d been with Ellison in a meeting with Ned Johnson, the chairman of America’s biggest mutual fund, Fidelity Investments. Ellison had dubbed Microsoft’s e-mail server software the “Virus Exchange” and had asked Johnson if he knew how much he was paying for this truly terrible system. Johnson’s CIO had figured maybe as much as $300 million a year.6 Ellison had immediately offered to write Johnson out a personal check for that amount if he couldn’t save him at least 85 percent of the cost while replacing it with something much better. It was the same pitch he’d made a few months later to Tommy Thompson in Washington, D.C. It was a pretty good way of getting the customer’s attention.
Unfortunately, at that stage Oracle didn’t have a complete product. Although it had the e-mail and the file system, it lacked a calendar—critical for coordinating meetings, travel, and appointments. Oracle’s developers had discovered that although Microsoft’s Outlook e-mail interface supported e-mail servers using either the standard IMAP4 Internet messaging protocol or Microsoft’s proprietary MAPI protocol, the moment you put an IMAP4 e-mail server such as Oracle’s underneath Outlook, Microsoft’s Calendar server automatically turns off. The result was that Oracle’s e-mail server and Microsoft’s calendar server wouldn’t work together. Ellison observed, “Microsoft is incapable of competing fairly. First they bundle their Outlook e-mail interface with Windows so everyone gets Outlook for ‘free.’ Then, if you use the ‘free’ Outlook as your e-mail interface, Microsoft makes it as difficult as possible to use any e-mail server other than theirs. Microsoft may claim that Outlook works with industry-standard IMAP4 e-mail servers like Oracle, but if actually you try to use one, everything else breaks. This isn’t bundling 101; this is graduate-level bundling. They’re the world’s experts. They should host a symposium on bundling theory. They have loads of land mines buried in their software to blow up the competition. It’s illegal, but they’d gotten away with it for years, so we had to figure out a way around the problem.” The solution, in June 2002, was to buy a little Canadian company named Steltor that claimed to make “the best open-system time management communication solution on the market” and had a formidable customer base that included BMW, Verizon, Texas Instruments, and Kellogg. Best of all, Steltor’s software fully supported every one of Outlook’s features. It was Oracle’s key to Microsoft’s door. Suddenly, all the pieces worked together.
The Collaboration Suite wasn’t just a welcome stick to beat Microsoft with. For years, Ellison had been arguing that Microsoft’s low-price, high-volume model wasn’t all it seemed because it brought with it so much complexity that had to be expensively managed. But now he was explicitly trying to beat Microsoft at its own game, and he was determined to enjoy himself in front of the assembled analysts. “Okay, here’s where we have some fun. Just for the heck of it, we decided to price Collaboration Suite, with all its features, at one third of what Microsoft charges just for its e-mail server alone. Bill says that Microsoft’s the place to go for low-priced software. Really? Not anymore. Collaboration Suite includes a secure, reliable database e-mail server, plus the best calendar/scheduler in the world, integrated voice mail, and a powerful database file-sharing system. Microsoft Exchange includes an e-mail server, plus a never-ending plague of viruses. Keep using Outlook; say good-bye to viruses with Oracle Collaboration Suite.” To cap even that, Ellison had a further offer: for $10 a user per month, Oracle would supply a completely outsourced service.
Launching the Collaboration Suite had allowed Ellison to do one of the things that in front of an audience he was great at: making jokes about Gates. But it also helped him make a much more serious point: Oracle might think that it didn’t need lessons from Microsoft about building software, but it could learn a great deal from the business strategies that had made Gates’s firm one of the world’s most fearsome competitors. From now on, Ellison was saying, Microsoft would be the model for the way Oracle would do business. It had started with the E-Business Suite, but now in everything that Oracle did it would emphasize the same tight integration that Microsoft understood so well, the same determination to throw in a bewildering array of features that smaller rivals had to charge for, the same aggressive pricing and ruthless pursuit of volume.
Ellison told the analysts, “Our strategy in enterprise software is integrated suites with a low total cost of ownership. The only software company we worry about, watch, and try to learn from is Microsoft. We think their basic model—integrated suites, low price, high volume—is the winner. Nothing else works, and that’s where we’re going. In the end, there will be just three software infrastructure companies: Microsoft and their proprietary Windows and .NET stuff; and IBM and Oracle supporting Linux, Java, and open systems. That’s it. Nobody else.”
If anyone in the audience disagreed, they weren’t saying so. The toughest question Ellison faced was about the impact on Oracle of competing in the America’s Cup. After “getting a pass” on winning four world maxi-yacht titles in recent years, Ellison said he’d been hoping for the same on the America’s Cup, “but it’s just much more visible.” However, he didn’t think it would be that bad: “eight weeks of driving and racing max over this year and next year—and that’s only if we win.” Most years, Ellison said, he took five weeks out of the office on vacation. As he’d be working through August, he didn’t think he’d have more time away from the office than usual. However, if Oracle BMW Racing went all the way, Ellison would be in Auckland for around eight weeks in 2002 and another four weeks in 2003. Luckily, none of the analysts seemed to have consulted the official race calendar.
And the dog that didn’t bark? Throughout the entire day there was just one question about the effect on Oracle’s business of the state of California contract “scandal”: Would Oracle find itself in the “penalty box” with some states and agencies? Not at all, replied Kevin Fitzgerald smoothly. Even in California itself, he expected to see new revenues in the current quarter. As for other public-sector customers, Fitzgerald said, “Most agencies and government entities recognize it for what it was. They ask you what lessons you learned. You say, don’t get in the way of an ambitious politicia
n, and don’t become an issue in a gubernatorial election year.”
* * *
1. LE writes: It never occurred to me that Oracle had done anything wrong, for the simple reason that Oracle hadn’t done anything wrong. Unfortunately, it took a while for that fact to come out. In the meantime, the media simply presumed us guilty of something—without ever saying exactly what.
2. LE writes: Logicon, not Oracle, recommended the deal to the state; Logicon, not Oracle, negotiated the terms of the deal with the state; Oracle, not Logicon, got all the headlines.
3. LE writes: We naively believed that because we hadn’t done anything wrong we had nothing to worry about.
4. LE writes: The state was saying it was a very bad deal, and at the same time, the state was refusing to undo the deal. It was a bit hard to understand—like the conversation between Alice and the Red Queen.
5. LE writes: To this day, Gartner calculates market share numbers using un-audited sales figures provided by IBM and Microsoft, rather than taking the trouble to survey customers.
6. LE writes: Microsoft’s e-mail and file servers are astonishingly expensive to run. Most companies can’t even calculate their total cost of ownership, which includes hardware, software, network, and labor costs. Then there are the huge additional costs related to e-mail viruses and lost files. Microsoft Exchange has probably lost business more money than any product in the history of the software industry.
26
“THE FIRST LOSER”
Hauraki Gulf, Auckland
Softwar Page 61