5. LE writes: It’s not fair to blame Peter for the mistakes. Everyone in our after guard, including our tactician, strategist, and navigator, contributed. We made so many errors during the GBR race that it’s hard for me to remember them all: we called the wind shift wrong on the first two beats; we did a bear-away set when we should have done a jibe set at the second weather mark and got passed as a result; we set the wrong spinnaker on the second run; and so on and so on. It was a fiasco. Before the race I had told Melanie that the only way we could lose to GBR was if we sank. It would have been less embarrassing if we had.
6. LE writes: Alinghi won the first race because Russell Coutts won the start. The start is critical in match racing. Winning the start does not necessarily mean that they started in front of us; it means that they got the side of the race course that they wanted. Five minutes before the race begins, both boats decide which side of the race course has the better wind conditions: more wind and/or a favorable wind shift up the course. If both boats pick the same side of the course, they fight for it by executing a complex set of maneuvers with each boat trying to gain a right-of-way advantage against the other. This five-minute maneuvering period before the start is called the prestart. It’s the most exciting part of match racing. Peter’s good at starting, but not as good as Russell Coutts. Russell won the start ten of the twelve times we raced Alinghi.
7. LE writes: I spoke, but it’s a good thing that nobody was around to hear what I said.
8. LE writes: With the larger sail area, USA-76 was faster downwind in all wind conditions and faster upwind in wind speeds of less than ten knots. We had gone from being a heavy-air upwind boat to being a light-air downwind boat. If the weather cooperated and we sailed well, we had a good chance of beating Alinghi in the finals.
9. LE writes: The engineering changes worked perfectly. As expected, Coutts kept winning all the starts and getting to the weather mark ahead of us, but we had enough boat speed to catch up on the downwind leg. Unfortunately, we kept making mistake after mistake out on the water. A bad approach to the first leeward mark caused a cascading series of disasters that took us out of the second race. In the third race we actually crossed the finish line well ahead of Alinghi, but we lost because of a penalty. We had no one to blame but ourselves. We just sailed badly.
10. LE writes: When we tried that, we missed Chris as tactician. We needed a second Chris.
11. LE writes: The finality of the defeat is hard to describe. All the time and effort you spent wasn’t enough. You weren’t smart enough. You weren’t good enough. You lost. It’s over. Go home. It’s a vacant, numb feeling. You congratulate the winners and slip away as quickly and as quietly as you can. As you leave you can hear the cheering—but they’re not cheering for you. You just don’t matter anymore. Your race is over. I hate that feeling. I hate being beaten. Especially when I know it was because of mistakes I made.
12. LE writes: If we hadn’t lost Gavin Brady to Prada, we just might have won the Cup. It was that close. We were just one person away from giving Brad and Russell a real race for it.
27
THE BIGGEST WATER BOTTLE
In the weeks after his return to California from New Zealand, as well as reimmersing himself in Oracle’s affairs Ellison worked on his “dialogue” with this book—the commentary that runs through these pages. We had agreed to meet up in California in early April to review and complete the work; partly because of the emotional roller coaster of the America’s Cup, an earlier deadline had already been missed. For me, it would also be a chance to talk not only to Ellison, which I had continued to do regularly, but also to some of the other Oracle executives I had come to know during the two years or so I had been involved in the life of their company. As well as catching up, it would help to put all that had happened into some kind of rough perspective.
Neither of the two Ellison campaigns I had intended to chronicle had exactly gone smoothly. The E-Business Suite was still some way from realizing Ellison’s prediction that it would become the default choice for any firm wanting to use the Internet to change its way of doing business, while the America’s Cup was sitting proudly in the trophy room of the Société Nautique de Genève rather than that of the Golden Gate Yacht Club. In fact, there were similarities between the two. Ellison’s USA-76 was almost certainly the fastest boat in the regatta and could have won had it not been for some silly mistakes on the water. So too, the extraordinary technical ambition of the E-Business Suite had been temporarily set back by what, with hindsight, seemed frustratingly avoidable execution errors. The difference was that the failure to win the America’s Cup was brutal and definitive, while it is clear that the campaign to establish the dominance of the E-Business Suite is only now getting into its stride. The E-Business Suite challenges conventional software industry thinking on so many different levels that the arguments have to be won one at a time. It’s a process of attrition: a war, not a battle.
The inescapable background to the writing of this book has been the aftermath of the “tech smash.” Oracle’s stock had defied gravity longer than just about anybody else’s save those of Microsoft and IBM, neither of which had enjoyed the same run-up during the years of Internet fever. But coincidentally, the crack had come just weeks before I started work, when Jeff Henley issued his first earnings warning since 1997. Like other big Internet infrastructure players, such as Cisco and Sun, Oracle had been unable to escape the fallout from disappearing dot coms and the imploding telecom sector. The vision of Oracle’s coming world domination (at least in terms of the software business) that Ellison had sketched out in his garden at Atherton with such certainty only a few months earlier suddenly seemed a long way off.
Some of the damage done to Oracle’s prospects was undeniably self-inflicted. As Ellison subsequently acknowledged, the supposedly all-conquering E-Business Suite had been hit by lapses in quality control, an overaggressive rollout, and a marketing message that rivals and analysts found all too easy to hijack and distort. Oracle had also managed to make enemies unnecessarily: by initially attempting to impose pricing models that only confused the market; by allowing a minor spat with the application users’ group to escalate into full-scale hostilities; and by appearing to declare war on the systems integrators. But while a little application market share was temporarily lost as the bugs infecting 11i’s stability were being hunted down, the overwhelming cause of Oracle’s slowing license revenues was the collapse in IT spending by companies that had gorged themselves on new technology during the Internet boom years and were now sitting on more hardware and software than they knew what to do with. For much of Oracle’s installed base, waiting a year or two for the E-Business Suite to overcome its well-publicized teething problems was just common sense.
That said, the extent of Oracle’s difficulties during the IT spending slump needs some context. Unlike both Cisco and Sun, far from plunging into big losses and write-offs, Oracle has remained steadfastly profitable, with its operating margins staying in the mid-thirties—a level barely contemplated, let alone seen, in the previous decade. As it has kept up the momentum of its internal e-business revolution, it has shed small pockets of labor, but Oracle has completely avoided the large-scale layoffs that have scarred the rest of Silicon Valley. And while others have pared research and development to the bone, Oracle continues to spend more than a billion dollars a year. Meanwhile, its cash mountain of nearly $6 billion has remained untouched and unchanged.1
Although Oracle’s market cap was “only” about $60 billion in April 2003 thanks to one of the lowest multiples in its history, it is still worth nearly two and a half times as much as the third biggest software firm in the world, SAP. Microsoft, it is true, has weathered the storm still better. But Microsoft is not as dependent on corporate spending and, crucially, has been able to exploit its monopoly by raising its prices as sales have slowed. Ellison says, “What do you do when there’s a recession? Raise prices, of course. Bill’s a genius. Why didn’t I think
of that?”
As for the specialized business application vendors that achieved spectacular valuations during the B2B phase of Internet hysteria, e-procurement specialists Ariba and Commerce One have almost disappeared beneath the waves, while the onetime supply-chain leader, i2, whose shares once reached $110, is now almost worthless, its stock having been delisted from the Nasdaq in May 2003 because of “accounting irregularities.”
In a way, even more dramatic is Siebel’s fall from grace. Two years ago, Siebel, the sales automation pioneer, had a claim to being the hottest software company on the planet. But in 2002 it plunged into loss as software sales evaporated, and a huge restructuring charge took an additional toll. Siebel was further tarnished as a result of doing too many of the now-discredited swap transactions that created such accounting havoc in the telecom sector. Siebel earned $30.7 million in the three months before June 2002 from deals in which the company had purchased goods or services from vendors within six months of having licensed software to them. Strip out Siebel’s cash pile of $1.6 billion, and the firm that Tom built is these days worth only a bit more than Oracle will spend on R and D this year. Unlike most of the others, it’s too early to write Siebel out of the game. But analysts for whom Siebel could previously do no wrong are predicting that it will find it hard to survive as an independent entity—SAP, Oracle, and even little PeopleSoft, once worth only a fraction of Siebel’s towering market cap, have all been touted as potential predators/suitors. Ellison, who bitterly regrets not buying PeopleSoft a few years ago, admits to having been tempted—not because he wants Siebel’s software but because he wouldn’t mind buying its installed base.
But Siebel’s problems go much deeper than some sloppy accounting and a nasty downturn in the economy. As Ron Wohl puts it, “Siebel had a near monopoly in CRM, but it turned out that Tom’s business model was flawed. It is now clear to nearly everyone that CRM is not the integration point. In a sense, there is no such thing as CRM, only sales and service automation that has to integrate with all the other ERP modules.” Many of Siebel’s customers had found, to their cost, that implementing the software they had bought was expensive and difficult. Nucleus Research, an East Coast consultancy, issued a damning report on Siebel during the summer of 2002. Based on interviews with sixty-six reference customers that Siebel had profiled on its Web site, Nucleus found that 61 percent believed they had achieved no positive return on investment (ROI) from their deployment and that nearly 80 percent claimed a major reason was because the software was much too difficult to use. Ellison’s charge a year earlier that Siebel often ended up as “shelfware” seemed proven.
But the main threat to Siebel is the growing credibility of the main suite vendors’ CRM offerings. It had taken Oracle and SAP a long time even to create “good-enough” CRM. And Ellison still wasn’t convinced that the CRM components in the E-Business Suite were capable of the sophisticated sales “choreography” that he was looking for.2 But in a “down” market that was risk-averse toward both complex implementations and software companies that might not be around long enough to maintain and upgrade their products, integrated suites from mature vendors with both a history and a future had become mighty attractive. Ellison says: “Best-of-breed products are just transitional technologies; suites always win in the end. Every six months Oracle and SAP come out with improved CRM products, and we eat another slice of Tom’s pie. I don’t see how he can survive. Siebel is the largest best-of-breed vendor and the last one standing. The second largest, i2, is dead and buried. Best-of-breed is too complex, too labor-intensive, too expensive to survive the stress of a long-lasting technology recession.” Ellison is more convinced than ever about the pace of consolidation in the application business: PeopleSoft had a chance of hanging in as a viable midmarket competitor in the United States, but the future belonged to the big battalions from SAP and Oracle.
Before he was replaced as head of North American sales by Keith Block, George Roberts had said to me, “There are three street rules if you’re going to survive in technology. First of all, don’t screw up so bad you knock yourself out of the market. Second, it’s a marathon, not a sprint. Lastly, when you get up in the morning you better be running. This is like the world championship of poker. Every year you’ve got to ante up more and more just to stay in the game. I don’t see any way that these smaller software companies can continue the level of investment they need, based on their profitability level versus the profitability and assets of companies like Oracle. I think you’re going to see them one by one all drop by the wayside. We’re all in the middle of the desert. We’ll see who gets out the other side.” It was similar to a metaphor coined by Mark Jarvis: in the world of business software, Oracle had the largest water bottle.
As for screwups, Oracle had certainly made a few with the E-Business Suite, but none had been bad enough to take it out of the race, a race made longer still by the downturn. It seemed to me this was Ellison’s single biggest mistake. Despite his almost matchless experience of the software business, he had expected success for the E-Business Suite to come much more quickly and more easily than it could ever realistically have done—even without the screwups.3
But although Ellison had been hopelessly wrong about timing and crazily overconfident about the true state of the E-Business Suite in the months immediately after its launch, about the really big things he’d been vindicated: the superiority of suites over best-of-breed; the importance of reengineering business practices to fit the software instead of rewriting code to produce customized systems that were unreliable and impossible to upgrade; the need for the automation of whole business flows to be complete and seamless; the understanding that the key to getting useful information out of complex computer systems was for applications to share consistent database schemata.
But despite the rightness of Ellison’s architectural vision, there were still many doubters. In April 2003, Ron Wohl said to me, “At the outset, people said that it was impossible for the E-Business Suite to do just about everything. Then they said, maybe it was possible after all, but you messed up the execution. Next they said, okay, it works, but it’s not for us. How do we persuade people finally to accept the superiority of the E-Business Suite? References and proven cost reduction. If you save time and dollars, nobody can debate it.”4
Within Oracle, there continued to be worries, as Mark Jarvis put it, that “Larry was too much the visionary purist—to some people he seems to be describing a perfect enterprise that in reality doesn’t exist.” Jeff Henley adds, “Larry still hurts himself by being too pure. Large companies get confused by the message—we have to do more to show that we live in a complicated world.” Keith Block had a slightly different perspective. He says, “CEOs like the architectural sell, and they really get the idea of the single data model. But the truth is that procurement tends to be driven further down the organization. If a CIO has invested $100 million in an SAP implementation, he’s not going to want to admit that he’s made a mistake.”
Although Oracle had taken major steps nearly two years earlier to ensure that integrating the E-Business Suite with legacy systems was perfectly feasible, you would not have gotten that impression from the firm’s advertising. Whereas SAP was forever banging the drum for its NetWeaver integration platform and talking up how its Web services and messaging software could make all kinds of applications work together without the pain of customization, Ellison was steadfastly dismissive of that approach: “Web services makes it easy for one application program to connect and pass data to another application—even if the applications are running on different computers. That’s great. But the data passed by one application has to be structured in a way that’s understandable by the other applications. Web services are analogous to the global telephone network, which makes it easy to call and connect to any telephone in the world. But if the guy you call only speaks French and you only speak English, you still can’t communicate even though you’re connected. Web services give us easy c
onnectivity but not a common language. Web services on their own don’t enable Siebel applications to talk to SAP applications.” Both SAP and Ellison were exaggerating. Web services were by no means the silver bullet that the German firm claimed, but they could help deliver some functionality across a heterodox collection of applications, more anyway than Ellison was prepared to admit.5
The irony of this was twofold: Oracle was itself a big player in Web services, and the company’s entire tradition was based on working with open standards. Nothing had changed, but it sounded as if Oracle had gone “proprietary.” While SAP was delivering a message of comfort for customers who feared being locked into a single vendor, Oracle seemed to be saying, “it’s all or nothing.” It was a real dilemma for Oracle’s marketing. On the one hand, Ellison didn’t want to muddy the differences between what he believed was the hopelessly compromised SAP philosophy and the much more elegant Oracle way of doing things. On the other, Mark Jarvis had a point in saying, “Oracle needs to bend a bit with the market. We have all the interfaces to link to other systems, but what we don’t have is a strong integration marketing message.”
The “purity” of Ellison’s vision was, however, unambiguously an advantage when it came to delivering what he and others were increasingly calling the “utility model” of delivering IT services. In Ellison’s view, the “information utility” was something much more than a fancy new name for the old-style outsourcing of data center management, practiced very profitably by the likes of IBM and EDS. If the utility concept meant anything, it had to deliver a quality service at very low cost and with complete reliability. That meant using only software that was preintegrated and designed to work out of the box on certified hardware. Ellison’s point was that computer systems didn’t need to be different. The “Oracle guarantee,” in which Oracle promised to fulfill all a firm’s computing needs for a guaranteed, fixed price—something quite revolutionary in an industry that has trained its customers to write blank checks—was rapidly winning converts, and not just in the midmarket, as had been expected. Big firms, such as the networking giant JDS Uniphase, were going down that road, and even the federal government was showing signs of interest.6
Softwar Page 64